Discussion Pysch - Paper Answers

Question:

Your textbook presents a figure (1.2) showing where I/O psychologists are employed in the U.S. Discuss which of these locations you would find most, and least, interesting and why? You need to have a good strong discussion and at least 250 words or more to make sure you understand the material.

The purpose is to engage with the assigned readings and integrate them with other ideas and into your daily (or future) practice. Please do use terminology from the assigned readings, cite the textbook with at least an in-text citation (Dipboye, 2016), and bring in current events (provide a URL) to add to the conversation and to support your points. These discussions are the main learning and assessment tool in this course. *Textbook attached*

EXPLORING INDUSTRIAL &
ORGANIZATIONAL
PSYCHOLOGY: Work &
Organizational Behavior
Robert L. Dipboye
First Edition

The first edition published 2016.
Text copyright. Robert L. Dipboye.
All rights reserved

ABOUT THE AUTHOR
Dr. Robert L. Dipboye received his undergraduate degree from Baylor University and
his Ph. D. in Industrial and Organizational Psychology from Purdue University. He was
Associate Professor in the Stokely School of Management at University of Tennessee
and Professor of Psychology and Management at Rice University. In his 26 years at
Rice he held the Herbert S. Autrey Chair in Social Sciences, was chair of the
department of psychology, and directed the program in Industrial and Organizational
Psychology. Most recently he was professor of psychology and department chair at the
University of Central Florida, where he is currently Professor Emeritus. Dr. Dipboye
was on the editorial boards of several professional journals including the Academy of
Management Review, Journal of Organizational Behavior, the Journal of Applied
Psychology, and the Frontiers Series of the Society of Industrial and Organizational
Psychology (SIOP). He also served as as an Associate Editor of the Journal of Applied
Psychology and as a Consulting Editor of the Journal of Organizational Behavior. He is
a fellow of the Society for Industrial and Organizational Psychology, the American
Psychological Association, and the Association for Psychology Science. Dr. Dipboye
has published several books and over 100 articles in the field of Industrial and
Organizational Psychology. He has also consulted with major corporations on problems
of Human Resources Management and Organizational Behavior.

TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 4
TABLE OF CONTENTS
PREFACE ………………………………………………………………………………………………………………………………. 20
CHAPTER 1: WHAT IS I/O PSYCHOLOGY? …………………………………………………………………………………… 27
INTRODUCTION ………………………………………………………………………………………………………………………….. 28
I/O AS A SPECIALTY WITHIN PSYCHOLOGY ……………………………………………………………………………………………. 28
WHAT DO I/O PSYCHOLOGISTS DO? …………………………………………………………………………………………………. 29
WHERE DO I/O PSYCHOLOGISTS WORK? ……………………………………………………………………………………………. 31
HOW MUCH DO I/O PSYCHOLOGISTS EARN? ………………………………………………………………………………………. 32
THE SCIENTIST-PRACTITIONER MODEL ……………………………………………………………………………………………….. 34
I/O IS A PROFESSION AS WELL AS A JOB ……………………………………………………………………………………………… 35
A body of specialized knowledge: expertise ……………………………………………………………………………. 36
A normative orientation to the service of others: ethics …………………………………………………………… 36
Technical autonomy: self-governance in professional associations……………………………………………. 40
BECOMING AN I/O PSYCHOLOGIST ……………………………………………………………………………………………………. 42
THE FUTURE LOOKS BRIGHT FOR I/O PSYCHOLOGY ………………………………………………………………………………… 44
SOME POINTS TO PONDER ……………………………………………………………………………………………………………… 46
CONCLUSIONS ……………………………………………………………………………………………………………………………. 46
CHAPTER 2: A HISTORY OF I/O PSYCHOLOGY …………………………………………………………………………….. 48
INTRODUCTION ………………………………………………………………………………………………………………………….. 49
THE EARLY YEARS (1880 – 1920) ……………………………………………………………………………………………………. 51
What were the major forces shaping work during the early years? …………………………………………… 51
Industrialization and mass production. ………………………………………………………………………………….. 51
Massive waves of immigration …………………………………………………………………………………………….. 53
Rise of the corporation. ……………………………………………………………………………………………………….. 53
What were the dominant theories of how to organize and manage work? ………………………………… 55
The emergence of psychology as a separate discipline …………………………………………………………….. 60
The early applications of psychology to work …………………………………………………………………………. 63
THE MIDDLE YEARS (1920 – 1945) …………………………………………………………………………………………………. 67
Economic boom and bust …………………………………………………………………………………………………….. 68
The Human Relations movement and neo-classical theories of management …………………………….. 74
The neoclassical and other notable organizational theorists…………………………………………………….. 80
THE POST WWII ERA (1945 – 1965) ………………………………………………………………………………………………. 83
Rapidly accelerating change in a period of progress and turmoil ………………………………………………. 83
The rise of organizational psychology ……………………………………………………………………………………. 87
THE MODERN ERA (1990 – NOW) …………………………………………………………………………………………………. 100
Global competition ……………………………………………………………………………………………………………. 101
Computer technology and the internet ………………………………………………………………………………… 101
Terrorism …………………………………………………………………………………………………………………………. 103
The population explosion, diminishing resources, and climate change …………………………………….. 104
CONCLUSIONS ………………………………………………………………………………………………………………………….. 105
CHAPTER 3: I/O PSYCHOLOGY AS A SCIENCE ……………………………………………………………………………. 107
INTRODUCTION ………………………………………………………………………………………………………………………… 108
DISTINGUISHING SCIENCE FROM NON-SCIENCE …………………………………………………………………………………… 108
Characteristics of non-scientific knowledge ………………………………………………………………………….. 109
Characteristics of scientific knowledge ………………………………………………………………………………… 110

THE SCIENTIFIC METHOD …………………………………………………………………………………………………………….. 114
The goals of scientific research …………………………………………………………………………………………… 116
Some basic terms ……………………………………………………………………………………………………………… 116
STATISTICS………………………………………………………………………………………………………………………………. 117
Distributions. ……………………………………………………………………………………………………………………. 117
Descriptive statistics………………………………………………………………………………………………………….. 119
Inferential statistics. ………………………………………………………………………………………………………….. 130
MEASUREMENT OF VARIABLES ………………………………………………………………………………………………………. 134
Levels of measurement. …………………………………………………………………………………………………….. 134
Reliability of measurement. ……………………………………………………………………………………………….. 135
Validity of measures. …………………………………………………………………………………………………………. 136
META-ANALYSIS ……………………………………………………………………………………………………………………….. 147
Basic steps involved in conducting a meta-analysis ……………………………………………………………….. 147
Corrections for statistical artifacts in meta-analysis ………………………………………………………………. 148
DESIGNING RESEARCH FOR EXPLANATION ………………………………………………………………………………………….. 151
Hold constant or eliminate variables …………………………………………………………………………………… 151
Manipulate the variable …………………………………………………………………………………………………….. 152
Measure the variable ………………………………………………………………………………………………………… 152
Randomly assign research participants to conditions …………………………………………………………….. 152
Match on the variable ……………………………………………………………………………………………………….. 154
DIFFERENT TYPES OF RESEARCH ……………………………………………………………………………………………………… 154
Experimental vs. nonexperimental ………………………………………………………………………………………. 154
Laboratory vs. field research ………………………………………………………………………………………………. 157
Obtrusive vs. unobtrusive research ……………………………………………………………………………………… 157
FACTORS INFLUENCING THE VALIDITY OF EXPLANATIONS ………………………………………………………………………… 158
Threats to the internal validity of a research design ………………………………………………………………. 159
Threats to the construct validity of a research design ……………………………………………………………. 161
Threats to external validity of a research design …………………………………………………………………… 162
ETHICS IN I/O PSYCHOLOGY RESEARCH …………………………………………………………………………………………….. 166
ACADEMIC RESEARCH VS. PRACTICAL REALITIES …………………………………………………………………………………… 168
Points to ponder ……………………………………………………………………………………………………………….. 169
CONCLUSIONS ………………………………………………………………………………………………………………………….. 170
CHAPTER 4: WORK MOTIVATON ……………………………………………………………………………………………. 171
INTRODUCTION ………………………………………………………………………………………………………………………… 172
Key behavioral indicators of motivation ………………………………………………………………………………. 173
Performance vs. motivation ……………………………………………………………………………………………….. 174
The internal focus of motivation theory ……………………………………………………………………………….. 174
The interaction of the environment with internal forces ………………………………………………………… 174
Seven practical questions …………………………………………………………………………………………………… 175
WHAT ARE EMPLOYEES’ GOALS? ……………………………………………………………………………………………………. 176
Why goals improve performance: Mediators of goal effects …………………………………………………… 176
What goal characteristics influence motivation and performance? …………………………………………. 177
What moderates goal effects? ……………………………………………………………………………………………. 180
Points to ponder ……………………………………………………………………………………………………………….. 184
WHAT ARE THE CONSEQUENCES OF EMPLOYEE BEHAVIOR? …………………………………………………………………….. 184
Types of conditioning ………………………………………………………………………………………………………… 185
Schedules of reinforcement ………………………………………………………………………………………………… 187
Application of BMod in the workplace …………………………………………………………………………………. 190

TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 6
Implications and current trends ………………………………………………………………………………………….. 192
Points to ponder: ………………………………………………………………………………………………………………. 193
WHAT ARE EMPLOYEES’ EXPECTATIONS? ………………………………………………………………………………………….. 193
Basic components and predictions of VIE theory …………………………………………………………………… 194
Research testing VIE theory ……………………………………………………………………………………………….. 200
Implications ……………………………………………………………………………………………………………………… 202
Points to ponder: ………………………………………………………………………………………………………………. 203
WHAT DO EMPLOYEES NEED? ……………………………………………………………………………………………………….. 203
Murray’s taxonomy of needs ……………………………………………………………………………………………… 205
Maslow’s need hierarchy …………………………………………………………………………………………………… 208
Alderfer’s existence, relatedness, growth (ERG) theory ………………………………………………………….. 213
McClelland’s three need theory ………………………………………………………………………………………….. 213
Practical implications of need theories ………………………………………………………………………………… 218
Points to ponder: ………………………………………………………………………………………………………………. 219
WHAT IS FAIR? ………………………………………………………………………………………………………………………… 219
Distributive fairness ………………………………………………………………………………………………………….. 221
Procedural fairness …………………………………………………………………………………………………………… 231
Interactional fairness ………………………………………………………………………………………………………… 234
Informational fairness ……………………………………………………………………………………………………….. 235
Inter-relationships among types of justice ……………………………………………………………………………. 237
Points to ponder: ………………………………………………………………………………………………………………. 237
DOES WORK ENGAGE THE SELF? ……………………………………………………………………………………………………. 238
Finding meaning through self-conception and self-evaluation ………………………………………………… 238
Theories of self-enhancement and self-maintenance …………………………………………………………….. 239
Self-related constructs ………………………………………………………………………………………………………. 240
Limitations on self-enhancement ………………………………………………………………………………………… 245
ARE EMPLOYEES INTRINSICALLY MOTIVATED? …………………………………………………………………………………….. 246
Herzberg’s two-factor theory ……………………………………………………………………………………………… 247
Self-determination theory ………………………………………………………………………………………………….. 249
Job enrichment as a strategy for increasing intrinsic motivation …………………………………………….. 251
Points to ponder: ………………………………………………………………………………………………………………. 256
INTEGRATIVE APPROACHES TO WORK MOTIVATION ……………………………………………………………………………… 256
The motivational theories are reconcilable …………………………………………………………………………… 256
Integrating with VIE theory ………………………………………………………………………………………………… 258
Work motivation as an unfolding process…………………………………………………………………………….. 259
Control theory as an integrative framework …………………………………………………………………………. 261
Points to ponder: ………………………………………………………………………………………………………………. 262
CONCLUSIONS ………………………………………………………………………………………………………………………….. 263
CHAPTER 5: ATTITUDES TOWARD WORK AND ORGANIZATIONS…………………………………………………. 267
INTRODUCTION ………………………………………………………………………………………………………………………… 268
JOB SATISFACTION …………………………………………………………………………………………………………………….. 270
How is job satisfaction measured? ……………………………………………………………………………………… 270
Points to Ponder: ………………………………………………………………………………………………………………………… 278
Work environment characteristics associated with job satisfaction …………………………………………. 279
Personal characteristics associated with job satisfaction ……………………………………………………….. 281
Person-environment fit as an antecedent to job satisfaction ………………………………………………….. 286
Points to ponder …………………………………………………………………………………………………………………………. 287
How does job satisfaction form? …………………………………………………………………………………………. 287

TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 7
Points to ponder …………………………………………………………………………………………………………………………. 295
Outcomes of job satisfaction/dissatisfaction ………………………………………………………………………… 296
Points to ponder: ………………………………………………………………………………………………………………………… 307
JOB INVOLVEMENT …………………………………………………………………………………………………………………….. 307
Measurement of job involvement ……………………………………………………………………………………….. 307
Correlates of job involvement …………………………………………………………………………………………….. 308
Points to ponder: ………………………………………………………………………………………………………………. 309
ORGANIZATIONAL COMMITMENT …………………………………………………………………………………………………… 309
Three components of organizational commitment ………………………………………………………………… 309
Correlates of organizational commitment ……………………………………………………………………………. 310
Commitment as an exchange process ………………………………………………………………………………….. 312
Points to ponder: ………………………………………………………………………………………………………………. 312
SATISFACTION, INVOLVEMENT AND COMMITMENT: SAME OR DIFFERENT CONSTRUCTS? …………………………………… 312
EMPLOYEE ENGAGEMENT: AN AGGREGATE ATTITUDINAL CONSTRUCT ………………………………………………………… 313
CONCLUSIONS ………………………………………………………………………………………………………………………….. 315
CHAPTER 6: WORK STRESS ……………………………………………………………………………………………………. 317
INTRODUCTION ………………………………………………………………………………………………………………………… 318
Myths about stress ……………………………………………………………………………………………………………. 319
Definition of stress. …………………………………………………………………………………………………………… 320
A TRANSACTIONAL PROCESS MODEL OF STRESS ………………………………………………………………………………….. 322
The physiology of stress …………………………………………………………………………………………………….. 323
Fight or flight. ……………………………………………………………………………………………………………………………… 323
General adaptation syndrome. ……………………………………………………………………………………………………… 323
The psychology of stress ……………………………………………………………………………………………………. 328
Cognitive appraisal. ……………………………………………………………………………………………………………………… 329
Person-Environment (PE) fit. …………………………………………………………………………………………………………. 330
Conservation of resources. …………………………………………………………………………………………………………… 330
Job Demands-Control-Support. ……………………………………………………………………………………………………… 331
Points to ponder ……………………………………………………………………………………………………………….. 331
WHAT ARE POTENTIAL STRESSORS FOR WORKERS? ………………………………………………………………………………. 332
What occupations are most stressful? …………………………………………………………………………………. 332
Lack of support …………………………………………………………………………………………………………………. 333
Work overload and demands ……………………………………………………………………………………………… 335
Lack of control and autonomy ……………………………………………………………………………………………. 337
Work underload: monotony and lack of challenge ………………………………………………………………… 340
Aversive physical work environment ……………………………………………………………………………………. 342
Shiftwork …………………………………………………………………………………………………………………………. 343
Role stress ……………………………………………………………………………………………………………………….. 344
Interpersonal conflict ………………………………………………………………………………………………………… 346
Organizational politics ………………………………………………………………………………………………………. 348
Organizational injustice …………………………………………………………………………………………………….. 349
Unfair discrimination ………………………………………………………………………………………………………… 351
Downsizing and job insecurity …………………………………………………………………………………………….. 352
Nonwork stressors …………………………………………………………………………………………………………….. 356
Points to ponder ……………………………………………………………………………………………………………….. 360
STRAINS: WORK-RELATED CONSEQUENCES OF SUSTAINED STRESS……………………………………………………………… 360
Effects of stress on performance …………………………………………………………………………………………. 361
Effects of stress on unsafe work behaviors and accidents ………………………………………………………. 363

TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 8
Effects of stress on work-related attitudes and OCBs …………………………………………………………….. 364
Effects of stress on burnout ……………………………………………………………………………………………….. 365
Effects of stress on work place violence ……………………………………………………………………………….. 367
Effects of stress on withdrawal behaviors…………………………………………………………………………….. 369
Points to ponder ……………………………………………………………………………………………………………….. 369
WHO IS MOST LIKELY TO SUFFER FROM STRESS? …………………………………………………………………………………. 370
Demographic variables ……………………………………………………………………………………………………… 371
Sex. ……………………………………………………………………………………………………………………………………………. 371
Ethnicity. ……………………………………………………………………………………………………………………………………. 372
Age. …………………………………………………………………………………………………………………………………………… 373
Discrimination against demographic groups. …………………………………………………………………………………… 373
Personality ………………………………………………………………………………………………………………………. 375
The big five personality traits. ……………………………………………………………………………………………………….. 375
Locus of control. ………………………………………………………………………………………………………………………….. 375
Type A-B personality. …………………………………………………………………………………………………………………… 376
Hardiness. ………………………………………………………………………………………………………………………………….. 377
Negative affectivity (NA). ……………………………………………………………………………………………………………… 378
Conceptual approaches to the role of personal characteristics in stress …………………………………… 378
Person-Environment fit. ……………………………………………………………………………………………………………….. 378
The conservation of resources model (COR). …………………………………………………………………………………… 379
HOW CAN EMPLOYEES MANAGE STRESS? …………………………………………………………………………………………. 380
Physical exercise ……………………………………………………………………………………………………………….. 380
Relaxation/meditation techniques ………………………………………………………………………………………. 382
Biofeedback ……………………………………………………………………………………………………………………… 383
Psychotherapy ………………………………………………………………………………………………………………….. 384
Personal coping styles ……………………………………………………………………………………………………….. 384
Do stress management interventions work? ………………………………………………………………………… 386
Organizational interventions ……………………………………………………………………………………………… 387
Points to ponder ……………………………………………………………………………………………………………….. 388
CONCLUSIONS ………………………………………………………………………………………………………………………….. 388
CHAPTER 7: SOCIAL PROCESSES IN ORGANIZATIONS ………………………………………………………………… 392
INTRODUCTION ………………………………………………………………………………………………………………………… 393
A working model of emergent social process ………………………………………………………………………… 394
Contextual factors influencing social process ……………………………………………………………………….. 395
Interpersonal factors …………………………………………………………………………………………………………. 398
Personal factors ……………………………………………………………………………………………………………….. 399
Formal vs. informal social processes ……………………………………………………………………………………. 401
Points to ponder ……………………………………………………………………………………………………………….. 402
COMMUNICATION PROCESSES ……………………………………………………………………………………………………….. 402
A model of organizational communication …………………………………………………………………………… 403
Points of vulnerability in the communication process …………………………………………………………….. 403
Picking a medium appropriate to the message …………………………………………………………………….. 414
Gossip and rumor ……………………………………………………………………………………………………………… 415
Points to ponder ……………………………………………………………………………………………………………….. 417
SOCIAL EXCHANGE PROCESSES ………………………………………………………………………………………………………. 418
Points to ponder: ………………………………………………………………………………………………………………. 422
COOPERATION, COMPETITION, AND CONFLICT PROCESSES ………………………………………………………………………. 422
Contextual factors affecting cooperation, competition, and conflict ……………………………………….. 423
Personal causes of cooperation, competition, and conflict……………………………………………………… 427

TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 9
The development of conflict ……………………………………………………………………………………………….. 428
Points to ponder: ………………………………………………………………………………………………………………. 430
ORGANIZATIONAL CITIZENSHIP BEHAVIOR …………………………………………………………………………………………. 431
COUNTER-PRODUCTIVE WORK BEHAVIOR (CWB) ……………………………………………………………………………….. 434
Points to ponder: ………………………………………………………………………………………………………………. 437
SOCIAL POWER AND INFLUENCE PROCESSES ……………………………………………………………………………………….. 437
The French and Raven model of social power ……………………………………………………………………….. 437
Social influence tactics ………………………………………………………………………………………………………. 439
Responses to influence attempts ………………………………………………………………………………………… 442
Points to ponder: ………………………………………………………………………………………………………………. 443
POLITICAL BEHAVIOR ………………………………………………………………………………………………………………….. 443
Political tactics …………………………………………………………………………………………………………………. 443
Situational antecedents of political behavior………………………………………………………………………… 445
Perceptions of politics (POPS) …………………………………………………………………………………………….. 447
Personality traits associated with perceptions of politics ……………………………………………………….. 449
Culture as a moderator of POPS………………………………………………………………………………………….. 450
Political skill and will …………………………………………………………………………………………………………. 451
Points to ponder: ………………………………………………………………………………………………………………. 452
CONCLUSIONS ………………………………………………………………………………………………………………………….. 452
CHAPTER 8: SOCIAL STRUCTURES IN ORGANIZATIONS ………………………………………………………………. 455
A model of imposed and emergent social structure ………………………………………………………………. 456
Formal vs. informal social structures …………………………………………………………………………………… 456
DIVISION OF LABOR ……………………………………………………………………………………………………………………. 458
Formal horizontal and vertical differentiation of labor…………………………………………………………… 458
Social roles ………………………………………………………………………………………………………………………. 459
Points to ponder ……………………………………………………………………………………………………………….. 461
DISTRIBUTION OF POWER AND STATUS …………………………………………………………………………………………….. 462
Alternative power distributions…………………………………………………………………………………………… 463
The universality of the pyramid ………………………………………………………………………………………….. 464
Why is the traditional pyramid so common? ………………………………………………………………………… 465
Emergent distributions of power and status …………………………………………………………………………. 466
Why do informal power distributions emerge? …………………………………………………………………….. 466
Points to ponder: ………………………………………………………………………………………………………………. 470
RULES……………………………………………………………………………………………………………………………………. 470
Formal rules …………………………………………………………………………………………………………………….. 470
Social norms …………………………………………………………………………………………………………………….. 471
Points to ponder ……………………………………………………………………………………………………………….. 475
CORE VALUES AS SOCIAL STRUCTURES ……………………………………………………………………………………………… 476
Formal communication of core values …………………………………………………………………………………. 476
Informal communication of core values via climate and culture ……………………………………………… 477
Points to ponder: ………………………………………………………………………………………………………………. 481
SHARED SOCIAL IDENTITY AS A SOCIAL STRUCTURE ……………………………………………………………………………….. 481
Formal attempts to create a shared identity ………………………………………………………………………… 482
Emergent identities in the organization ……………………………………………………………………………….. 483
The emergence of a shared social identity ……………………………………………………………………………. 485
Points to ponder: ………………………………………………………………………………………………………………. 490
SOCIAL NETWORKS ……………………………………………………………………………………………………………………. 491
The method of social network analysis ………………………………………………………………………………… 492

Points to ponder: ………………………………………………………………………………………………………………. 493
Dynamics influencing the emergence and maintenance of social networks ……………………………… 494
Points to ponder: ………………………………………………………………………………………………………………. 502
The social network as a source of social capital ……………………………………………………………………. 502
Points to ponder: ………………………………………………………………………………………………………………. 513
SOCIALIZATION …………………………………………………………………………………………………………………………. 513
Outcomes of socialization ………………………………………………………………………………………………….. 514
Anticipatory socialization …………………………………………………………………………………………………… 514
The initial encounter …………………………………………………………………………………………………………. 515
Socialization tactics …………………………………………………………………………………………………………… 515
Meta-analyses of the effects of socialization tactics ……………………………………………………………… 518
Conformity, rebellion, creative individualism, and cooptation ………………………………………………… 518
THE INTERPLAY OF IMPOSED FORMAL STRUCTURES AND EMERGENT INFORMAL STRUCTURES ……………………………… 519
CONCLUSIONS ………………………………………………………………………………………………………………………….. 521
CHAPTER 9: GROUPS AND TEAMS IN ORGANIZATIONS ……………………………………………………………… 524
INTRODUCTION ………………………………………………………………………………………………………………………… 525
WHAT IS A GROUP? …………………………………………………………………………………………………………………… 526
Formal work groups ………………………………………………………………………………………………………….. 526
How does a team differ from a group? ………………………………………………………………………………… 528
HOW DO GROUPS CHANGE OVER TIME? ………………………………………………………………………………………….. 529
WHY ARE SOME GROUPS MORE EFFECTIVE THAN OTHERS? ……………………………………………………………………. 531
Group inputs…………………………………………………………………………………………………………………….. 531
Group interaction processes ………………………………………………………………………………………………. 531
Emergent group structures ………………………………………………………………………………………………… 533
Proximal influences on group effectiveness ………………………………………………………………………….. 533
Group effectiveness …………………………………………………………………………………………………………… 533
Critical task contingencies ………………………………………………………………………………………………….. 533
Points to ponder: ………………………………………………………………………………………………………………. 534
GROUP INPUTS …………………………………………………………………………………………………………………………. 535
How many members are needed? ………………………………………………………………………………………. 535
Who are chosen as the members of the group? ……………………………………………………………………. 536
The knowledge, skills, and abilities of group members. ……………………………………………………………………. 536
Team KSAs. …………………………………………………………………………………………………………………………………. 537
Personalities of group members. …………………………………………………………………………………………………… 538
Attitudes about working in groups. ………………………………………………………………………………………………… 539
Diversity. ……………………………………………………………………………………………………………………………………. 540
1. Faultlines. …………………………………………………………………………………………………………………………… 542
2. Moderators of diversity and faultlines. …………………………………………………………………………………… 543
Points to ponder: ………………………………………………………………………………………………………………. 547
Group tasks ……………………………………………………………………………………………………………………… 548
Conjunctive, disjunctive and additive tasks. ……………………………………………………………………………………. 548
McGrath’s (1984) typology. ………………………………………………………………………………………………………….. 549
Points to ponder: ………………………………………………………………………………………………………………. 552
GROUP INTERPERSONAL PROCESSES ………………………………………………………………………………………………… 552
Who influences whom and how? ………………………………………………………………………………………… 552
Do members share leadership? …………………………………………………………………………………………… 555
Is there conflict among members and why? …………………………………………………………………………. 557
Do members share information with each other? …………………………………………………………………. 559
Points to ponder: ………………………………………………………………………………………………………………………… 560

Do members coordinate and synchronize their activities? ……………………………………………………… 561
Do members engage in prosocial behavior? …………………………………………………………………………. 562
Points to ponder: ………………………………………………………………………………………………………………………… 565
What is the affective tone in the group? ………………………………………………………………………………. 565
How motivated are members to work with their groups? ………………………………………………………. 566
Do members reflect, learn, and adapt? ……………………………………………………………………………….. 571
Do members engage in transactive memory? ………………………………………………………………………. 573
Higher order process dimensions ………………………………………………………………………………………… 574
Points to ponder: ………………………………………………………………………………………………………………………… 574
EMERGENT GROUP STRUCTURES ……………………………………………………………………………………………………. 574
What social roles emerge? …………………………………………………………………………………………………. 575
What social norms emerge? ………………………………………………………………………………………………. 577
What interpersonal relationships emerge? …………………………………………………………………………… 581
Points to ponder: ………………………………………………………………………………………………………………………… 583
How cohesive is the group?………………………………………………………………………………………………… 583
Effects of cohesion on group effectiveness……………………………………………………………………………………… 585
Downsides to cohesion. ……………………………………………………………………………………………………………….. 587
Points to ponder: ………………………………………………………………………………………………………………………… 589
Do members identify with the group? …………………………………………………………………………………. 590
Points to ponder: ………………………………………………………………………………………………………………………… 593
INTERVENTIONS TO IMPROVE GROUP PERFORMANCE ……………………………………………………………………………. 593
INTERVENTIONS THAT TARGET GROUP PROCESSES …………………………………………………………………………………. 593
Team training. ………………………………………………………………………………………………………………….. 593
Guided reflexivity. …………………………………………………………………………………………………………….. 595
Team building. …………………………………………………………………………………………………………………. 596
Process consultation. ………………………………………………………………………………………………………… 597
INTERVENTIONS THAT TARGET THE GROUP CONTEXT. …………………………………………………………………………….. 598
Group goal setting. …………………………………………………………………………………………………………… 598
Group feedback. ……………………………………………………………………………………………………………….. 599
Group incentives. ……………………………………………………………………………………………………………… 601
Group involvement in decision making. ……………………………………………………………………………….. 602
Programs that combine incentives, feedback, goal setting, and involvement. …………………………… 605
INTERVENTIONS IN THE FORM OF AIDS, TOOLS, AND TECHNOLOGIES ……………………………………………………………. 606
Structured group problem solving techniques. ……………………………………………………………………… 606
Computer mediated communication and virtual teams. ………………………………………………………… 609
Points to ponder: ………………………………………………………………………………………………………………. 613
CONCLUSIONS ………………………………………………………………………………………………………………………….. 614
CHAPTER 10: LEADERSHIP …………………………………………………………………………………………………….. 619
INTRODUCTION ………………………………………………………………………………………………………………………… 620
The romance of leadership vs. a scientific approach ……………………………………………………………… 621
A working definition of leadership ………………………………………………………………………………………. 622
Leadership as incremental compliance ………………………………………………………………………………… 623
Leader vs. manager or supervisor ……………………………………………………………………………………….. 623
Alternative approaches to understanding leadership …………………………………………………………….. 624
Points to ponder: ………………………………………………………………………………………………………………………… 624
EFFECTIVE LEADERS MAKE THE MOST OF THE SITUATION: SITUATIONAL APPROACHES ……………………………………… 627
Situational forces shape leader behavior ……………………………………………………………………………… 627
Situational forces shape effectiveness …………………………………………………………………………………. 628

Implications of situational theory for leader effectiveness ……………………………………………………… 629
Summary …………………………………………………………………………………………………………………………. 630
Points to ponder: ………………………………………………………………………………………………………………………… 631
EFFECTIVE LEADERS POSSESS THE PERSONAL CHARACTERISTICS OF LEADERS: THE TRAIT APPROACH ……………………… 631
Personality and intelligence ……………………………………………………………………………………………….. 631
The motivation to lead ………………………………………………………………………………………………………. 633
Physical traits …………………………………………………………………………………………………………………… 635
Sex differences …………………………………………………………………………………………………………………. 636
Is leadership inherited? ……………………………………………………………………………………………………… 637
Moderating effects of situation on relation of traits to leadership…………………………………………… 638
Summary …………………………………………………………………………………………………………………………. 640
Points to ponder: ………………………………………………………………………………………………………………. 641
LEADERS DO THE RIGHT THINGS: THE BEHAVIORAL APPROACH ………………………………………………………………… 641
Leader behaviors fulfill both task and socio-emotional demands …………………………………………….. 643
The Ohio State leadership studies………………………………………………………………………………………………….. 643
The Managerial Grid© ………………………………………………………………………………………………………………….. 644
Correlations with satisfaction and performance. ……………………………………………………………………………… 646
Summary. …………………………………………………………………………………………………………………………………… 646
Points to ponder: ………………………………………………………………………………………………………………………… 647
Effective leaders communicate effectively ……………………………………………………………………………. 648
Effective leaders use referent and expert power …………………………………………………………………… 649
Points to ponder: ………………………………………………………………………………………………………………………… 652
Effective leaders manage conflict ……………………………………………………………………………………….. 652
Points to ponder: ………………………………………………………………………………………………………………………… 653
Effective leaders are charismatic, inspirational, and transformational …………………………………….. 654
The charismatic leader. ………………………………………………………………………………………………………………… 654
The inspirational leader. ………………………………………………………………………………………………………………. 655
Transformational leadership. ………………………………………………………………………………………………………… 656
Points to ponder: ………………………………………………………………………………………………………………………… 657
Effective leaders develop high quality relationships with followers (i.e., high LMX) …………………… 658
Points to ponder: ………………………………………………………………………………………………………………………… 661
EFFECTIVE LEADERS MANAGE PERCEPTIONS: THE COGNITIVE APPROACH …………………………………………………….. 662
Categorization in leadership perception ………………………………………………………………………………. 662
Structure of cognitive categories used in describing leaders. …………………………………………………………….. 663
Effects of performance cues on perceptions of leadership………………………………………………………………… 664
Effective leaders have confidence in their followers ………………………………………………………………. 664
Leaders manage impressions ……………………………………………………………………………………………… 667
Points to ponder: ………………………………………………………………………………………………………………. 667
EFFECTIVE LEADERS DIAGNOSE AND ADAPT TO THE SITUATION: THE CONTINGENCY APPROACH …………………………… 668
Effective leaders change the situation to fit their traits (Fiedler) …………………………………………….. 668
Effective leaders modify their participativeness to fit the situation …………………………………………. 671
Effective leaders help followers attain goals ………………………………………………………………………… 677
Leader directiveness. …………………………………………………………………………………………………………………… 678
Leader supportiveness. ………………………………………………………………………………………………………………… 679
Leader participativeness. ……………………………………………………………………………………………………………… 679
Achievement oriented leadership. …………………………………………………………………………………………………. 680
Effective leaders consider the maturity of followers ………………………………………………………………. 680
Points to ponder: ………………………………………………………………………………………………………………. 682
CONCLUSIONS ………………………………………………………………………………………………………………………….. 683
CHAPTER 11: ANALYZING WORK ……………………………………………………………………………………………. 686

INTRODUCTION ………………………………………………………………………………………………………………………… 687
What is a job? ………………………………………………………………………………………………………………….. 687
What does a job analysis measure? …………………………………………………………………………………….. 691
Points to ponder: ………………………………………………………………………………………………………………. 693
WHY DO JOB ANALYSES? …………………………………………………………………………………………………………….. 694
Reducing role conflict and ambiguity …………………………………………………………………………………… 694
Design and evaluation of training ……………………………………………………………………………………….. 695
Performance appraisal and criterion development ……………………………………………………………….. 695
Job design ………………………………………………………………………………………………………………………… 695
Personnel selection …………………………………………………………………………………………………………… 695
Wage and salary administration …………………………………………………………………………………………. 696
Compliance with regulations and laws ………………………………………………………………………………… 696
Points to ponder: ………………………………………………………………………………………………………………. 697
HOW ARE JOB ANALYSES EVALUATED? …………………………………………………………………………………………….. 697
HOW IS A JOB ANALYSIS CONDUCTED? …………………………………………………………………………………………….. 699
Methods of collecting job information …………………………………………………………………………………. 699
Observation method of job analysis ………………………………………………………………………………………………. 699
Interview method of job analysis. ………………………………………………………………………………………………….. 700
Questionnaires. …………………………………………………………………………………………………………………………… 700
Using a combination of methods. ………………………………………………………………………………………………….. 700
Who provides the information? ………………………………………………………………………………………….. 701
How much information is needed? ……………………………………………………………………………………… 702
Specific methods of job analysis………………………………………………………………………………………….. 703
Behaviorally-oriented techniques. …………………………………………………………………………………………………. 704
Techniques focused on requirements. ……………………………………………………………………………………………. 709
Points to ponder: ………………………………………………………………………………………………………………………… 716
Eclectic techniques. ……………………………………………………………………………………………………………………… 716
Points to ponder: ………………………………………………………………………………………………………………………… 721
JOB EVALUATION AND THE GENDER GAP IN COMPENSATION ……………………………………………………………………. 721
Methods of job evaluation …………………………………………………………………………………………………. 722
Ranking………………………………………………………………………………………………………………………………………. 722
Classification systems. ………………………………………………………………………………………………………………….. 722
Point systems. …………………………………………………………………………………………………………………………….. 722
Evaluating internal equity in the wage structure …………………………………………………………………… 724
Comparable worth and the gender gap in wages ………………………………………………………………….. 725
Are job evaluation procedures biased? ………………………………………………………………………………… 728
How do occupations become sex-typed? ……………………………………………………………………………… 729
Points to ponder: ………………………………………………………………………………………………………………. 730
CONCLUSIONS ………………………………………………………………………………………………………………………….. 730
CHAPTER 12: CRITERION DEVELOPMENT, PERFORMANCE APPRAISAL, AND FEEDBACK ………………….. 733
INTRODUCTION ………………………………………………………………………………………………………………………… 734
WHAT CRITERIA ARE USED TO EVALUATE PERFORMANCE? ……………………………………………………………………… 735
Points to ponder: ………………………………………………………………………………………………………………. 736
WHAT IS EVALUATED? ………………………………………………………………………………………………………………… 737
Points to ponder: ………………………………………………………………………………………………………………. 739
HOW IS PERFORMANCE MEASURED? ………………………………………………………………………………………………. 740
Evaluating performance appraisal measures ……………………………………………………………………….. 740
Objective measures …………………………………………………………………………………………………………… 742
Points to ponder …………………………………………………………………………………………………………………………. 744

Evaluating employees on graphic rating scales …………………………………………………………………….. 744
Rating effects and biases. ……………………………………………………………………………………………………………… 745
Points to ponder: ………………………………………………………………………………………………………………………… 754
Cognitive structures and processes in an evaluator’s ratings of an employee. ………………………….. 754
Points to ponder: ………………………………………………………………………………………………………………………… 758
Organizational and social determinants of ratings. ……………………………………………………………….. 758
Purpose of the rating. ………………………………………………………………………………………………………………….. 758
Organizational culture/climate. …………………………………………………………………………………………………….. 759
Accountability of the rater for the appraisal. …………………………………………………………………………………… 759
Politics of the organization……………………………………………………………………………………………………………. 760
Other contextual determinants of ratings. ……………………………………………………………………………………… 760
Using rating scales to improve ratings. ………………………………………………………………………………… 761
Behaviorally anchored rating scales (BARS) and behavioral expectation scales (BES). …………………………… 761
Mixed standard scale (MSS). …………………………………………………………………………………………………………. 763
Behavioral observation scale (BOS)………………………………………………………………………………………………… 764
Forced choice measures. ………………………………………………………………………………………………………………. 765
Weighted checklist measures. ………………………………………………………………………………………………………. 767
Forcing comparisons to improve evaluations. …………………………………………………………………………………. 768
Summary of research on rating scales. …………………………………………………………………………………………… 771
Training to improve performance appraisals. ……………………………………………………………………….. 772
Points to ponder: ………………………………………………………………………………………………………………………… 773
WHO SHOULD JUDGE PERFORMANCE? …………………………………………………………………………………………….. 773
Supervisors? …………………………………………………………………………………………………………………….. 773
Peers? ……………………………………………………………………………………………………………………………… 774
Self-Appraisals? ………………………………………………………………………………………………………………… 774
Subordinates? ………………………………………………………………………………………………………………….. 775
Customers? ……………………………………………………………………………………………………………………… 775
Who agrees with whom? …………………………………………………………………………………………………… 775
360 degree evaluation and feedback. ………………………………………………………………………………….. 776
WHEN ARE PERFORMANCE EVALUATIONS CONDUCTED? ………………………………………………………………………… 776
Points to ponder: ………………………………………………………………………………………………………………. 777
HOW IS FEEDBACK GIVEN TO EMPLOYEES ON THEIR PERFORMANCE APPRAISALS? ………………………………………….. 777
Feedback giving and seeking ……………………………………………………………………………………………… 778
Processing of feedback ………………………………………………………………………………………………………. 779
Feedback in the formal appraisal session …………………………………………………………………………….. 780
Points to ponder: ………………………………………………………………………………………………………………. 783
CONCLUSIONS ………………………………………………………………………………………………………………………….. 783
CHAPTER 13: EMPLOYEE TRAINING AND DEVELOPMENT …………………………………………………………… 785
INTRODUCTION ………………………………………………………………………………………………………………………… 786
AN INSTRUCTIONAL SYSTEMS MODEL OF TRAINING ………………………………………………………………………………. 787
Points to ponder: ………………………………………………………………………………………………………………. 789
NEEDS ASSESSMENT …………………………………………………………………………………………………………………… 789
Organization training needs analysis …………………………………………………………………………………… 789
Task training needs analysis ………………………………………………………………………………………………. 790
Person training needs analysis ……………………………………………………………………………………………. 791
An example of a training needs assessment …………………………………………………………………………. 792
Specification of Instructional Objectives ………………………………………………………………………………. 793
Gagne’s learning outcomes………………………………………………………………………………………………… 793
*Intellectual skills. ……………………………………………………………………………………………………………………….. 794

*Cognitive strategies. …………………………………………………………………………………………………………………… 794
*Motor skills. ………………………………………………………………………………………………………………………………. 794
*Attitudes. …………………………………………………………………………………………………………………………………. 795
Stating instructional objectives …………………………………………………………………………………………… 795
Points to ponder: ………………………………………………………………………………………………………………. 796
TRAINING AND DEVELOPMENT PHASE ………………………………………………………………………………………………. 796
Incorporating the findings of learning research …………………………………………………………………….. 796
*Provide advance organizers. ……………………………………………………………………………………………………….. 797
*Practice and recite. ……………………………………………………………………………………………………………………. 797
*Distribute practice. …………………………………………………………………………………………………………………….. 798
*Learn the whole-task on complex and organized tasks. ………………………………………………………………….. 798
*Set goals. ………………………………………………………………………………………………………………………………….. 799
*Give positive reinforcement. ……………………………………………………………………………………………………….. 799
*Give knowledge of results (feedback). ………………………………………………………………………………………….. 800
*Provide models. ………………………………………………………………………………………………………………………… 800
*Make learning a team effort. ………………………………………………………………………………………………………. 800
Gagne’s model of instructional events …………………………………………………………………………………. 801
Points to ponder: ………………………………………………………………………………………………………………. 802
WAYS TO ENHANCE POSITIVE TRANSFER OF LEARNING TO WORKPLACE ………………………………………………………. 802
Maintain what was learned. ………………………………………………………………………………………………. 802
Increase similarity of stimulus and response in training and the workplace ……………………………… 803
Create difficulties for trainees …………………………………………………………………………………………….. 804
Instruction in how to overcome barriers in the workplace. …………………………………………………….. 807
Increase the motivation to learn. ………………………………………………………………………………………… 808
Create a climate supportive of the training ………………………………………………………………………….. 809
Type of skill learned as a moderator of transfer. …………………………………………………………………… 809
Points to ponder: ………………………………………………………………………………………………………………. 811
ALTERNATIVE TRAINING METHODS …………………………………………………………………………………………………. 811
On-the-job training …………………………………………………………………………………………………………… 811
Mentoring and coaching. ……………………………………………………………………………………………………………… 811
Internships. ………………………………………………………………………………………………………………………………… 812
Apprenticeships. …………………………………………………………………………………………………………………………. 812
Job rotation and transfers. ……………………………………………………………………………………………………………. 813
The hidden costs of on-the-job training. …………………………………………………………………………………………. 814
Integrating on-the-job and classroom training. ……………………………………………………………………………….. 814
Off-the-job training methods ……………………………………………………………………………………………… 815
Lecture. ……………………………………………………………………………………………………………………………………… 815
Programmed instruction. ……………………………………………………………………………………………………………… 816
Simulation. …………………………………………………………………………………………………………………………………. 818
Conference discussion. ………………………………………………………………………………………………………………… 818
Case method. ……………………………………………………………………………………………………………………………… 819
Role playing. ……………………………………………………………………………………………………………………………….. 819
Behavior modeling. ……………………………………………………………………………………………………………………… 820
Choosing among alternative instructional methods ………………………………………………………………. 821
Points to ponder: ………………………………………………………………………………………………………………. 821
IMPLEMENTATION OF TRAINING …………………………………………………………………………………………………….. 822
Face-to-face instruction …………………………………………………………………………………………………….. 822
Recorded presentations …………………………………………………………………………………………………….. 823
Web based instruction ………………………………………………………………………………………………………. 823
EVALUATING TRAINING PROGRAMS …………………………………………………………………………………………………. 825
Criteria for evaluating training …………………………………………………………………………………………… 825

Research designs for evaluating training ……………………………………………………………………………… 828
Pseudo experiments…………………………………………………………………………………………………………………….. 829
True experiments. ……………………………………………………………………………………………………………………….. 829
Complications associated with experimenting in an organization. ……………………………………………………… 830
Taking into account individual differences among trainees ……………………………………………………. 831
Summative vs. formative evaluation of training ……………………………………………………………………. 831
Points to ponder: ………………………………………………………………………………………………………………. 832
CONCLUSIONS ………………………………………………………………………………………………………………………….. 832
CHAPTER 14: PRINCIPLES OF EMPLOYEE SELECTION …………………………………………………………………. 834
INTRODUCTION ………………………………………………………………………………………………………………………… 835
STAFFING AND THE MATCHING STRATEGY …………………………………………………………………………………………. 835
Some caveats …………………………………………………………………………………………………………………… 836
Points to ponder: ………………………………………………………………………………………………………………. 837
SELECTION ………………………………………………………………………………………………………………………………. 837
The scientific and intuitive approaches to selection ………………………………………………………………. 838
Two examples. ……………………………………………………………………………………………………………………………. 838
Features held in common by the two approaches. …………………………………………………………………………… 838
Important differences. …………………………………………………………………………………………………………………. 839
I/O psychology’s role. …………………………………………………………………………………………………………………… 842
Adaptation to special circumstances. …………………………………………………………………………………………….. 842
Points to ponder: ………………………………………………………………………………………………………………. 843
Evaluation of selection techniques ………………………………………………………………………………………. 843
Reliability. …………………………………………………………………………………………………………………………………… 843
*Test-retest (stability) indexes………………………………………………………………………………………………….. 844
*Equivalent (alternate) forms approach. ……………………………………………………………………………………. 844
*Internal consistency (or internal reliability) indexes. ………………………………………………………………….. 845
*Interrater (or interjudge) reliability. ………………………………………………………………………………………… 845
Points to ponder: ………………………………………………………………………………………………………………………… 846
Validity. ……………………………………………………………………………………………………………………………………… 846
*Criterion-related validation. …………………………………………………………………………………………………… 847
**Validity generalization vs. situational specificity. ………………………………………………………………… 849
**Factors that lead to underestimation of criterion-related validity. ………………………………………… 849
**Meta-analytic procedures in testing validity generalization. …………………………………………………. 851
*Content validation. ……………………………………………………………………………………………………………….. 855
*Construct validation. ……………………………………………………………………………………………………………… 855
*Face validity. ………………………………………………………………………………………………………………………… 856
*Points to ponder: ………………………………………………………………………………………………………………….. 857
Fairness and legality. ……………………………………………………………………………………………………………………. 858
*U. S. Civil rights legislation. …………………………………………………………………………………………………….. 859
* U. S. Executive Order 11246 (1965). ……………………………………………………………………………………….. 860
* The EEOC Uniform Guidelines. ……………………………………………………………………………………………….. 860
*U. S. Supreme Court decisions interpreting the Civil Rights Act of 1964. ………………………………………. 861
**Griggs v. Duke Power (1971). …………………………………………………………………………………………… 861
**Albermarle Paper Company v. Moody (1975). ……………………………………………………………………. 861
*Age Discrimination in Employment Act (ADEA). ………………………………………………………………………… 862
*Americans with Disability Act (1990). ………………………………………………………………………………………. 862
*U. S. Civil Rights Act of 1991. ………………………………………………………………………………………………….. 862
*Disparate treatment vs. disparate impact discrimination……………………………………………………………. 863
*The Cleary (1968) model of test fairness. …………………………………………………………………………………. 863
*Values and fairness. ………………………………………………………………………………………………………………. 868
Points to ponder: ………………………………………………………………………………………………………………………… 868

Utility. ……………………………………………………………………………………………………………………………… 869
*Using the Taylor-Russell tables to calculate utility……………………………………………………………………… 870
*Brogden-Cronbach-Gleser utility formula. ………………………………………………………………………………… 874
Points to ponder: ……………………………………………………………………………………………………………………. 877
Using predictor measures to make selection decisions …………………………………………………………… 877
The top-down procedure. …………………………………………………………………………………………………………….. 877
The Angoff procedure in setting cut scores. ……………………………………………………………………………………. 878
Banding of cut scores. ………………………………………………………………………………………………………………….. 879
Using multiple predictor measures. ……………………………………………………………………………………………….. 883
*Compensatory strategy of combining predictors. ……………………………………………………………………… 884
*Noncompensatory strategies of combining predictors. ………………………………………………………………. 885
*Clinical vs mechanical combination of predictors. ……………………………………………………………………… 885
Points to ponder: ………………………………………………………………………………………………………………. 886
PLACEMENT AND CLASSIFICATION …………………………………………………………………………………………………… 886
RECRUITMENT ………………………………………………………………………………………………………………………….. 887
Factors that affect applicant attraction and choice. ………………………………………………………………. 888
Realistic job previews ………………………………………………………………………………………………………… 889
Moral/ethical vs. scientific questions …………………………………………………………………………………… 891
Points to ponder: ………………………………………………………………………………………………………………. 892
CONCLUSIONS ………………………………………………………………………………………………………………………….. 892
CHAPTER 15: CONSTRUCTS AND METHODS IN EMPLOYEE SELECTION …………………………………………. 895
INTRODUCTION ………………………………………………………………………………………………………………………… 896
Constructs used in selection ……………………………………………………………………………………………….. 896
Methods used in assessing applicants …………………………………………………………………………………. 897
Source: objective testing, self-report, and subjective judgments. ………………………………………………………. 897
Structure of selection methods. …………………………………………………………………………………………………….. 897
Maximal vs typical performance tests. …………………………………………………………………………………………… 898
Points to ponder: ………………………………………………………………………………………………………………. 899
PREDICTORS BASED ON STABLE WORKER TRAITS …………………………………………………………………………………. 899
Worker abilities ………………………………………………………………………………………………………………… 899
Specific cognitive abilities. ……………………………………………………………………………………………………………. 900
General cognitive ability (also called general intelligence, general mental ability). ………………………………. 903
*Controversy surrounding the construct of cognitive ability…………………………………………………………. 904
*The controversy over the fairness of cognitive ability testing. …………………………………………………….. 905
*The author’s take on the issues. ……………………………………………………………………………………………… 906
Points to ponder: ………………………………………………………………………………………………………………………… 907
Physical and psychomotor abilities. ……………………………………………………………………………………………….. 907
Occupational interests ………………………………………………………………………………………………………. 910
Work values …………………………………………………………………………………………………………………….. 911
Personality traits ………………………………………………………………………………………………………………. 912
Alternative models of personality ………………………………………………………………………………………………….. 912
The Big Five personality traits. ………………………………………………………………………………………………………. 914
Alternative types of personality tests. ……………………………………………………………………………………………. 914
Reasons for relatively low validities. ………………………………………………………………………………………………. 916
Points to ponder: ………………………………………………………………………………………………………………. 919
Knowledge and Skills Acquired Through Education and Experience …………………………………………. 919
The O*NET description…………………………………………………………………………………………………………………. 920
Experience, education, and tenure. ……………………………………………………………………………………………….. 920
Knowledge as a predictor. …………………………………………………………………………………………………………….. 922
Work samples……………………………………………………………………………………………………………………………… 924

Points to ponder: ………………………………………………………………………………………………………………. 925
PREDICTORS WHERE UNCERTAINTY SURROUNDS THE CONSTRUCTS ……………………………………………………………. 925
Interviews ………………………………………………………………………………………………………………………… 925
Unstructured interviews. ……………………………………………………………………………………………………………… 926
Structured interviews (behavioral interview). …………………………………………………………………………………. 926
Meta-analyses of interview validities. …………………………………………………………………………………………….. 929
Biographical information …………………………………………………………………………………………………… 930
Weighted application blank (WAB). ……………………………………………………………………………………………….. 930
Biographical inventory blank (BIB). ………………………………………………………………………………………………… 930
The empirical vs rational approach to constructing BIBs and WABs. …………………………………………………… 932
Assessment centers …………………………………………………………………………………………………………… 934
What is an assessment center? ……………………………………………………………………………………………………… 934
Criterion-related validity of assessment centers. ……………………………………………………………………………… 935
Construct validity issues. ………………………………………………………………………………………………………………. 935
* Do assessment center ratings contaminate criteria or vice versa? ………………………………………………. 936
* Do assessment center ratings reflect common stereotypes of the successful employee? ………………. 937
* Are assessment center evaluations self-fulfilling prophecies? ……………………………………………………. 937
* Are assessment center evaluations affected by the employee’s past performance? ……………………… 937
* Are assessment centers tests of intelligence? ………………………………………………………………………….. 937
References ……………………………………………………………………………………………………………………….. 938
Grades …………………………………………………………………………………………………………………………….. 939
Emotional intelligence ……………………………………………………………………………………………………….. 940
Three models of emotional intelligence. ………………………………………………………………………………………… 941
* The mixed model. ………………………………………………………………………………………………………………… 941
* The ability model. ………………………………………………………………………………………………………………… 941
*The personality model. ………………………………………………………………………………………………………….. 943
Criticisms of emotional intelligence. ………………………………………………………………………………………………. 943
Meta-analyses of criterion-related validities for emotional intelligence. …………………………………………….. 944
Situational judgment tests …………………………………………………………………………………………………. 945
Polygraph and integrity testing ………………………………………………………………………………………….. 948
Individual assessments ………………………………………………………………………………………………………. 949
Points to ponder: ………………………………………………………………………………………………………………. 951
COMPARISONS OF PREDICTORS ON VALIDITY, ADVERSE IMPACT, AND APPLICANT REACTIONS ……………………………… 951
Points to ponder: ………………………………………………………………………………………………………………. 956
CONCLUSIONS ………………………………………………………………………………………………………………………….. 956
CHAPTER 16: EPILOGUE ……………………………………………………………………………………………………….. 958
THREE LEVELS OF GENERALITY IN PRACTICAL PRINCIPLES ………………………………………………………….. 959
Concrete take aways …………………………………………………………………………………………………………. 959
Fit the person to the situation ………………………………………………………………………………………………………. 960
Leave room for growth ………………………………………………………………………………………………………………… 960
Set goals and then monitor and reward progress …………………………………………………………………………….. 961
Continually learn and grow …………………………………………………………………………………………………………… 961
Justice for all ………………………………………………………………………………………………………………………………. 962
Cultivate teamwork ……………………………………………………………………………………………………………………… 962
Cultivate high quality relationships: The human relations movement redux ……………………………………….. 962
Empower ……………………………………………………………………………………………………………………………………. 963
Engage the self ……………………………………………………………………………………………………………………………. 963
Manage perceptions ……………………………………………………………………………………………………………………. 964
Midrange principles ………………………………………………………………………………………………………….. 964
Define success… ………………………………………………………………………………………………………………………….. 965

Build human capital …………………………………………………………………………………………………………………….. 965
Build social capital ……………………………………………………………………………………………………………………….. 966
Diagnose the situation and the contingencies …………………………………………………………………………………. 966
Integrate the formal and the informal ……………………………………………………………………………………………. 966
More is not always better …………………………………………………………………………………………………………….. 967
Fundamental principles: …………………………………………………………………………………………………….. 967
Ethics trump efficacy ……………………………………………………………………………………………………………………. 967
Take a scientific perspective…measure, theorize, and empirically test ……………………………………………….. 968
1. Meta-analysis for better and worse. ………………………………………………………………………………………. 969
2. Keeping it simple, the danger in the proliferation of constructs. ……………………………………………….. 969
3. Taking effect sizes too seriously or not seriously enough. …………………………………………………………. 969
4. The need for multiple methods. ……………………………………………………………………………………………. 970
CONCLUSIONS ………………………………………………………………………………………………………………………….. 970
REFERENCES ……………………………………………………………………………………………………………………….. 972
FIGURE, TABLE, AND IMAGE CREDITS ……………………………………………………………………………………. 1117
TOPIC INDEX …………………………………………………………………………………………………………………….. 1145

PREFACE

Exploring Industrial and Organizational Psychology provides a comprehensive review of
the topics typically covered in courses on Industrial and Organizational (I/O) Psychology
(most often used in the U. S., Canada, England, and Australia) and Work and
Organizational Psychology (used in Germany, Netherlands, and other European
countries). The text is appropriate for a broad audience including students at the
undergraduate and graduate levels as well as those readers looking for an introduction to
the topic. Because the text is an in depth coverage of the field, some readers may find the
material challenging. There was a concerted effort to present the theory, research, and
application as it is rather than presenting a grab bag of techniques and current fads.
Strengths and weaknesses are discussed and when possible the actual results are
presented in the form of effect sizes. Maintaining scientific integrity required in many
cases a level of detail that is probably unusual for an introductory text. In short, if the
reader is looking for simple primer, this is probably not the best choice. This text was
written with several core attributes in mind, some of which are distinctive and distinguish
it from other I/O psychology textbooks.

First, as implied in the title, the scientific nature of the field is emphasized. As in any
science, there are controversies, conflicting results, and theoretical disputes. This is not
the book for the reader seeking the final word on questions of human behavior in
organizations. This is the book for those readers who seek to gain insight into a field that
is continuing to evolve and who appreciate the competition of ideas crucial to scientific
progress. I hope that the reader will share in the excitement associated with the efforts of
I/O psychologists and other organizational scientists to generate an understanding of why
people behave as they do in organizations and the implications for organizational
effectiveness. The text does cover practical applications, especially in that section
covering personnel topics. However, an underlying theme is that good practice follows
from solid science.

Second, in reviewing the science, quantitative meta-analyses are used to show the
cumulative results of research on the topics covered in this book. It is often the case that
texts provide only qualitative reviews of research findings. As a consequence, the reader
may mistakenly conclude that the relations among variables are stronger or weaker than
is actually the case. Meta-analyses provide the reader with a realistic picture of the state
of research findings and are useful in identifying directions for future research.

Third, the text provides the student with a basic understanding of the topics and
intentionally avoids overly detailed presentations of psychometric issues and statistics.
Consequently, students should be able to comprehend the material without much prior
background in psychology or statistics. Although I have avoided overloading the reader
with too much detail, the reader will also find that I thoroughly review the literature on
the topics, with each chapter containing both classic and current citations.

Fourth, Exploring Industrial and Organizational Psychology provides a balanced
coverage of personnel and organizational topics (the so-called I and O of I/O
Psychology). The reader is exposed to a full range of organizational topics such as
leadership, power, role theory, groups in organizations, conflict, and communication, as
well as the traditional personnel topics of job analysis, selection, performance appraisal,
staffing, and training.

Fifth, an effort is made to show the relevance of the research, theory, and applications to
current events. Major changes are occurring in society that will fundamentally change
careers and the way organizations manage human resources. Corporate restructuring,
globalization of business, and diversity of the workforce are only a few of the continuing
changes that are reshaping worklife. Throughout the text I use examples of how these
changes are affecting organizations and the implications of these changes for the nature
of work and the management of people at work.

Finally, the text attempts to provide an integrative review of the field. I/O psychology
texts typically present organizational and personnel topics as if they are entirely separate
areas of research. While I have maintained the split in the text between the “I” and the
“O”, I cross reference throughout the text and frequently refer to the implications of
research in O for topics in I and vice versa. Also, the organizational topics are covered
prior to the personnel topics to emphasize that the latter always occur in the context of the
organization.

The topics covered in this textbook are typical of those included in other I/O textbooks.
There is no coverage of consumer psychology, unions, or engineering psychology. These
are important topics but have become at best peripheral to the discipline of I/O
psychology. They deserve a more focused and separate treatment. Also, there is no
chapter dedicated to cross-cultural issues. The most impressive development in the field
of I/O psychology over the last three decades is the conduct of research by those outside
the United States. At one time, I/O was dominated by work in North America and
England. A perusal of the professional journals reveals the explosion of research in
China, Japan, Germany, Netherlands, and other countries. Where appropriate, attention is
given to the national and cultural differences that moderate the results of research. The
dynamics of human behavior at work and in organizations appear consistent enough
across cultures to justify the universalistic approach of this text.

The first section of the text contains three chapters that are intended to provide a
foundation for the second and third sections. The first chapter defines the field of I/O
psychology and describe the activities and career paths of those who are in this field.
Special attention is given in this chapter to how I/O psychology is part of the larger field
of psychology but yet distinct social, clinical, cognitive and the other subdisciplines. The
second chapter outlines the history and development of I/O psychology and related fields.
Here there is an attempt to identify the individual scientists and practitioners who
contributed to this history and the societal events that provided the background for the
emergence of scientific applications of psychology to work and organizations. The third
chapter is perhaps the most important in the text. Each of the subsequent chapters reviews

TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 22
the body of knowledge that has accumulated for various topics, while emphasizing that
what is known about each topic is evolving. Scientific research continues to contribute to
each topic and what is known about each topic will change as new findings emerge; but
the method of scientific research remains constant. The third chapter covers the elements
of the scientific method and the use of scientific methodology by researchers in I/O
psychology.
The second section of this text covers topics in Organizational Psychology (the O in the
I/O), including work motivation, attitudes, groups, social structure, social processes,
leadership, and stress. Much of this literature comes from the research of social scientists
who view organizations as complex social systems. By starting with organizational
psychology, the “big picture” of how organizations function is presented prior to moving
on, in the second half of the text to the more applied issues that are the focus of personnel
or industrial psychology (the I in I/O). For instance, the discussion of work motivation
and social processes will set the stage for later discussions of performance appraisal and
feedback. Despite this structure, each chapter stands alone and does not require the
student to read any previous chapters. The only possible exception is the examination of
research methods which probably should be assigned prior to the chapters in the second
and third sections.
The basic component of the organization is the individual employee. It is appropriate
then to begin this section with a consideration of work motivation and attitudes. Much of
the interest in these topics has come from the potential role that unmotivated and
dissatisfied employees may play in the economic competitiveness of an organization and
an entire nation. For instance, productivity has declined in the U. S. over the last few
decades. Is the declining motivation of the work force partly responsible? Do employees
feel that their important needs are being met in their work, and what are effects of the
frustration and satisfaction of needs on work behavior? Do employees expect that they
will be rewarded for working hard and making other contributions to the organization and
how do these expectations influence motivation? Do employees feel that they have been
fairly treated and how do feelings of equity or inequity translate into job performance?
Chapter 3 explores each of these questions and the process by which needs, expectations,
and equity influence the effort that employees invest in their work.
The answers to these questions have important implications for how organizations go
about managing their employees. Traditional management practices often assume that
employees are passive and lack the ability to contribute creative ideas. The human
relations approach has promoted the opposite view that all employees want challenging
work and the opportunity to develop their skills. Research on motivation has empirically
examined these assumptions and has provided a more realistic and complex view than
provided in either the traditional or human relations approaches. Much of this research
has focused on content issues such as what employees want, whereas other studies have
addressed the process of motivation. Included in the chapter is a discussion of recent
attempts to pull together the various theories of motivation and the practical implications
of these efforts.

Motivation leads directly to the topic of attitudes toward work and the organization in
Chapter 4. Some critics of modern management practices have claimed that today’s
workforce has become increasingly alienated and “turned off” by dehumanizing aspects
of their jobs. There is a chapter on the work-related attitudes of job satisfaction,
organizational commitment, and job involvement that examines alternative measures of
these work-related attitudes, and the findings of research with these instruments. As
shown in this chapter, the findings of research have not supported some widely held
beliefs, such as the assumption that job satisfaction leads directly to good performance.
Nevertheless, worker attitudes are crucial to understanding and predicting other important
behaviors in the job, such as turnover.

I/O psychology is concerned with the well-being of the employee in addition to the
performance of the organization and chapter 6 delves into the stress that employees
experience as they attempt to fulfill the demands of their work roles. Work stress is
considered to be epidemic among employees in many occupations but is it really as bad
as the popular press makes it out to be? Chapter 6 will examine the personal and
situational factors that can determine the physiological and psychological reactions of
employees to unusual demands (sometimes called stressors). In addition to examining the
antecedents and consequences of stress, this chapter also evaluates research on coping
strategies and the applications of this work to helping employees manage stress in the
workplace.

In contrast to the chapters on attitudes and motivation, which are concerned with
behavior of individual employees, chapters 7, 8, and 9 are concerned with social behavior
in the workplace. The importance of this topic becomes apparent when you consider that
an organization is essentially a group of groups. Much of the activity in an organization
consists of people in the process of communicating, influencing, competing, fighting, and
helping. Social structures, such as roles and norms, emerge to add predictability to these
processes. Much of the social behavior that can be observed in an organization comes as
the result of the formal departments and other subunits that management has put together
to get the work done. Just as important, however, are the informal relationships that
emerge and sometimes conflict with formal social arrangements. Whether formal or
informal in nature, groups often fail to achieve their potential. This chapter concludes
with an examination of some of the interventions that are being used to improve group
effectiveness in organizations.

Leadership involves all of the topics in the course and is perhaps the most controversial
of the social psychological aspects of organizations. Although those who hold the formal
positions of leadership are usually held responsible for the success of the organization,
some organizational theorists have questioned whether leaders are really as important as
most people seem to believe (see). Is leadership mainly in the mind of the beholder such
that anyone who manages the right impressions can become a leader? Is leadership an
outdated concept reflecting mainly the needs of followers to believe that someone is in
charge? If leadership is important to the success of the organization, what can be done to
improve the effectiveness of leaders? Is it primarily a matter of selecting the people with
the right traits? Can I train people to show the right behaviors regardless of their traits?

TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 24
Do great leaders transform rather than simply manage their followers? In exploring the
research that has addressed these questions, I will show how the simplistic views of what
makes an effective leader have been replaced by more complex contingency models.
Essentially, these models state that the effectiveness of leadership depends on whether
the traits and behaviors of the leader are appropriate to the situation. Different situations
call for radically different types of leaders.
The third section of the text covers topics that are usually lumped under the title of
personnel or industrial (the I in I/O) psychology. Included in this section are work
analysis, employee selection, training, and performance appraisal. In contrast to the more
theoretical slant of organizational psychologists, personnel psychologists tend to be more
practically oriented, with much of the research devoted to developing techniques that can
be used to improve the fit between employees and the organization.
Providing a “good fit” is in part an attraction problem in that you want to entice the most
qualified people to seek employment in your organization. It is a selection problem in
that once you interest people in applying for positions, you want to be able to choose the
best qualified people. It is a placement problem in that once hired, you want to put people
where they belong. It is a training problem in that once hired and placed, you want to
instruct your employees so that they can perform the work properly. Work analysis is the
foundation for all these activities. Chapter 11 covers the topic of work analysis. In other
textbooks, this topic is usually part of the section on the “I” in I/O psychology. All the
subsequent topics build on basic questions about work. What is a position, a job, a career,
and occupation? What are the basic dimensions along which work can be described and
compared? What are methods of analyzing work?
Another important step in providing a good fit is determining the standards of
performance that employees are expected to attain in their work. Chapter 12 deals with
alternative measures of employee performance and their uses in an organization.
Measures of performance can be used in a variety of human resource management
activities, including the rewarding, punishing, promoting, firing, and developing of
employees. They also are used as criteria in research evaluating human resource
management interventions. Several options are available to employers seeking to appraise
performance. Performance appraisals can be objective or subjective, can be conducted by
supervisors, peers, subordinates, customers, or the employees themselves, and can focus
on results, process, or traits of the employees. As shown in the chapter, much of the
performance appraisal that occurs in organizations relies on subjective ratings. The large
amount of research on subjective rating scales has provided insight into the pros and cons
of specific scales such as Behavioral Anchored Rating Scales (BARS). A point that
emphasized, however, is that more attention needs to be given to the process of judging
performance as opposed to developing the right scale. A particularly important
application of performance appraisals is in providing feedback to employees. The chapter
concludes with a consideration of some of the methods of providing feedback that appear
to have great potential for improving performance effectiveness.

No matter how good staffing procedures might be in a company, the employees that a
company attracts and selects will still have some “rough edges’ that need to be smoothed
to provide a good fit to the work. The next chapter is concerned with training, a personnel
function that is aimed at instilling the knowledge, skills, abilities, and attitudes needed to
perform the work. As important as training is to the organization, it is often haphazard
and wasteful in the way it is implemented. Chapter 13 is organized around a systems
model that is presented as an ideal approach to training. An essential point of this model
is that training is a subsystem within the organization and should be evaluated on the
basis of how Ill it fulfills the needs of the larger organization. According to this model,
training programs should be designed to fulfill objectives that are derived from a careful
assessments of needs and then evaluated against how well they fulfill these needs and
modified accordingly. A variety of on-the-job and off-the-job training techniques
currently exist, and each has its advantages and disadvantages. The chapter reviews the
pro’s and cons of these various techniques, along with interventions to improve
acquisition and transfer of what is learned. Although the training literature is not as well
developed as the work on staffing, this is changing. The chapter on training will cover
some of the exciting new developments that are likely to have dramatic effects on the
practice of training in the future.

Staffing, which is the focus of chapters 14 and 15, is the chief activity of many I/O
psychologists. A variety of selection techniques have been developed over this century
that have proven quite valuable in screening prospective employees and predicting their
future performance. These include measures of mental and physical abilities, biodata,
personality questionnaires, work samples, assessment centers, structured interviews, and
others. A major point in this chapter is that organizations should conduct research to
evaluate the sensitivity, reliability, and validity of selection techniques. In actually using
a selection instrument in making decisions about potential employees, other factors need
to be taken into account, including how many people are currently succeeding, how many
applicants you can reject out of those who apply, the cost of the procedure, and the
fairness of the decisions. Much of this chapter will concentrate on selection, but in
closing I will also give some attention to employee placement and recruitment.

The final chapter provides an overview of some major conclusions to be drawn from the
previous chapters. As is always the case with any scientific endeavor, there are no final
answers. Everything is a work in progress. I do not measure the success of this book
against the number of definitive answers it provides to crucial questions such as how we
motivate workers. Rather, I consider the standard of success to be the extent that readers
are challenged, intrigued, and even provoked to seek their own answers. I sincerely hope
that the reader shares the excitement associated with this relatively new field of
psychology and the search for answers to questions surrounding human behavior at work.

This text builds on four previous books. William (Bill) C. Howell wrote a short volume
entitled Essentials of I/O Psychology that was first published in the 1970s, followed by
two revised editions that I co-authored. In the 1990s, I became first author of a new
textbook, Understanding I/O Psychology: An Integrated Approach with Carla S. Smith
and William C. Howell as second and third authors. Unfortunately, before the text could

be revised, Carlla became seriously ill and passed away. I considered a revision but
without Carlla, neither Bill nor I had the heart to launch the revision. So the text remained
dormant until Bill’s untimely death in 2013, when I decided to try my hand at a new text.
The product of these efforts is Exploring Industrial and Organizational Psychology.
Although this is a new text, the contributions of both Carlla and Bill are apparent
throughout. To the extent that this text proves intellectually stimulating, entertaining, or
in some other way beneficial, they should share in the recognition. To the extent that the
text falls short, I assume total responsibility and they deserve none of the blame.

CHAPTER 1: WHAT IS I/O PSYCHOLOGY?

Introduction

Industrial and Organizational (I/O) Psychology (also called Work and Organizational
Psychology in Europe or simply Organizational Psychology) is the scientific study of
how individuals and groups behave in the performance of work activities and in the
context of organizations. It is also the application of this research to improving the
effectiveness and the well being of people and the organizations in which they work. As
implied in this definition, Industrial and organizational (I/O) psychology is a specialized
field within the larger discipline of psychology. It is part science, contributing to the
general knowledge base of psychology, and part application, using that knowledge to
solve real-world problems. The business of I/O psychology is NOT simply business. I/O
psychology is concerned with a variety of types of work including service,
manufacturing, professional, intellectual, physical, paid and unpaid. I/O psychology is
concerned with a variety of organizations including, but not limited to, business,
governmental, military, religious, educational, union, and not-for-profit organizations.
I/O psychology is concerned not only with how to increase the productivity and
efficiency of workers in the performance of their tasks and the effectiveness of the
organizations to which they belong, but also the health and well-being of individuals who
perform work activities and society in general.

Psychology is not the only social science discipline that is concerned with these issues.
Other social sciences including anthropology, sociology, political science,
communications, and economics have made important contributions. Also, other
specialties within psychology are concerned to some extent with the topic, most notably
social psychology, personality, and psychometrics. The term Organizational Science is
often used to refer to the interdisciplinary research and theory concerned with
understanding the behavior of organizations and the people in these organizations. I/O
psychology is one of several social sciences and psychological specialties contributing to
the multidisciplinary effort. Despite the interdisciplinary nature of the field, I/O
psychology is the dominant specialty that has taken a psychological approach to
understanding human behavior at work and in organizations and is the focus of this text.

I/O as a Specialty within Psychology

It is one thing to write a concise definition of I/O psychology and quite another to convey
in concrete terms what those words actually mean. We have given you the definition;
now we will try to make it come alive. A technical “field” such as I/O psychology really
consists of several things. It is a core body of knowledge, a set of tools used to expand
and apply that knowledge, a setting or context in which that expansion and application
take place, a collection of people who identify with the field, and a variety of institutions
that hold all these components together (e.g., professional organizations, journals,
meetings, training programs).

This chapter first examines the professionals who identify themselves as I/O
psychologists. What do they do? Where do they work? How much do they earn? The
chapter then considers how I/O as a subdiscipline fits into the larger discipline of

psychology and other related fields outside of psychology. A final consideration is the
defining characteristics of I/O psychology as a profession: the ethical principles and
standards that I/O psychologists espouse, the content of the training in I/O psychology,
and the professional associations that represent I/O psychologists.

What Do I/O Psychologists Do?

If you were to follow some I/0 psychologists around to learn what they do for a living,
you would find them engaged in a wide variety of activities and working in a wide
variety of organizations, as illustrated in table 1.1.

MANAGER, Organizational Development
DIRECTOR, Management, Education, & Leader Development
CAREER COACH
CONSULTANT in
Organizational Development
Organizational Effectiveness
Talent Management
Employee Relations
Evaluation and Assessment
Testing Programs
Learning and Performance
Selection Systems
Leadership Development
Training and Development
PRESIDENT Consulting Group, University, Corporation
SENIOR RESEARCH ASSOCIATE
BUSINESS PSYCHOLOGY
ASSOCIATE DEAN
SENIOR PERSONNEL PSYCHOLOGY
PERSONNEL PSYCHOLOGIST
PERSONNEL RESEARCH PSYCHOLOGIST
ORGANIZATIONAL EFFECTIVENESS MANAGER
CONSULTING HR PROFESSIONAL
HUMAN RESOURCES DIRECTOR
VICE PRESIDENT, GLOBAL LEARNING
DIRECTOR – TALENT PLANNING
ASSESSMENT MANAGER
ORGANIZATIONAL DEVELOPMENT MANAGER
SENIOR PERSONNEL ANALYST

Table 1.1: Sample Titles Held by I/O Psychologists

In their research and practical applications, I/O psychologists tend to specialize in one of
two general areas often referred to as the “I” and the “O” in I/O psychology. They differ
chiefly in the kinds of psychological content and techniques they use as well as in their
general approach to organizational problems. Those taking the “I” approach tend to
emphasize measurement and individual differences. They use quantitative methods to
help organizations make the best use of their people by getting precise measures of both
human and work characteristics and then applying those measures in human resource
management. You will find them engaged in the design (and management) of recruiting,
selection and training, the design and evaluation of compensation systems, and the
analysis, description and evaluation of jobs. In contrast, those taking the “O” approach
tend to rely on broad psychological theories in diagnosing and trying to correct
organizational problems. Theories of social behavior and human motivation figure
prominently in this strategy. For example, a common theme in “O” practice is that non-
management employees ought to have some influence in decisions that affect them
because people will work harder and stay more committed to goals that they understand
and help set. Organizational psychologists help managers interact more effectively with
the individuals and groups they manage, conduct attitude surveys of the workforce,
design and evaluate procedures for communicating and motivating employees, implement
and evaluate organizational development (OD) interventions such as team building, and
help individuals cope with stress. Although there is a tendency for I/O psychologists to
sort themselves into one or the other camp (the I and the O), it is in many ways a false
dichotomy. Interventions that involve I-related activities almost always require attention
to the larger organizational context. Likewise, O-related activities inevitably involve
measurement and other I-related activities. A well-trained I/O psychologist is willing to
crossover.

Approximately 60 – 70% of I/O psychologists work as practitioners with the remainder
working mostly in academic settings as researchers and teachers. A SIOP survey
provided more detail on the specific activities that distinguish practitioners and non-
practitioners (Silzer, Cober, Erickson & Robinson, 2008). The members of the Society
for Industrial and Organizational Psychology were surveyed and 1005 respondents
replied, a 36% response rate. Respondents were asked the percentage of time they spent
as a practitioner of I/O, as educators, and as researchers/scientists. Sixty-one percent
indicated that 70% or more of their time was spent in practice and were defined as full-
time practitioners. Approximately 10% identified themselves as non-practitioners. Of
those who were non-practitioners, the mean percentage of time spent in science/research
was 43.5 and the mean percentage of time in education was 39%. Of those who were
practitioners, the mean percentage of time in science/research was 4.2% and the mean
percentage in education was 1.2%. The respondents were given practice activities and
asked how important each was to their work as a practitioner. The percentage of
practitioners and nonpractitioners stating that each activity was important or very
important is summarized in figure 1.1. The six most important activities for full time
practitioners were

• Building relationships
• Consulting with and advising clients
• Managing work projects and administrative activities
• Implementing and delivering programs and / or tools
• Making presentations
• Developing and designing systems, methods and/or programs

The least important activities among practitioners were:
• Writing for a scientific journal
• Teaching courses or training programs
• Writing reports, articles, chapters
• Conducting primary research and data analysis

Figure 1.1: Activities Performed by I/O Psychologists

Where Do I/O Psychologists Work?

Some I/O psychologists work full-time within one organization in performing these tasks.
Others would perform some of the same functions but as consultants to a variety of
0 20 40 60 80 100 120
consulting and advising clients
Building relationships
Managing work projects and…
Implenting and delivering…
Making presentations
Developing and designing…
Coaching others and providing…
Developing strategy and policy
Leading and managing others
conducting selection and…
writing business proposals
Leading change management…
Conducting primary research…
Writing reports, articles and…
Teaching courses or training…
Managing a business
Writing for a scientific journal
Percentage selecting activity as very important or important to success
A
ct
iv
it
y
p
e
rf
o
rm
e
d
in
p
o
si
ti
o
n

Non-practitioners
Full time practitioners

clients rather than as full-time employees of one organization. Moreover, their services
might include diagnosing an organization’s management problems and, when necessary,
recommending and helping to bring about company-wide changes. Some of these
consultants are totally self-employed, work for a consulting firm, or hold full-time
academic appointments while consulting on the side. Still other I/O psychologists are
found in university psychology laboratories, corporate research and development groups,
advertising agencies, marketing departments, and government research organizations.
Those so engaged would most likely work as scientists whose primary function, like that
of biologists, physicists, or other social scientists, is generating knowledge. Some might
try to answer simple applied questions such as what test to use to select employees, while
others are concerned with “basic” issues such as what motivates people to work. Most are
trying to answer questions that lie somewhere between these extremes. Regardless of
scope or focus, all of them would use scientific methods such as those presented in the
next module to gather new information, rather than rely on what they already know to
solve problems. And finally, of course, some teach in a university psychology department
or business school, management training or executive development program, or some
other education or training setting. Most academic I/O psychologists combine teaching,
research, and consulting. The chart below is a breakdown of the specific places of
employment based on a survey of the SIOP membership.

Figure 1.2: Where do U. S. I/O Psychologists Work?
How Much Do I/O Psychologists Earn?

Of all the psychological specialties, I/O psychologists are the highest paid. A survey was
conducted in early 2013 in which 4,073 members of the Society for Industrial and
Organizational Psychology were contacted and asked to complete a survey regarding
Universities and
Colleges
42%
External Consulting
28%
Private industry
20%
Government,
military, nonprofit
10%
Place of employment of SIOP members

TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 33
their 2012 incomes (see figure 1.3). There was a response rate of 32.3%. The analyses
were limited to 1,120 respondents who worked full-time. When broken down by place of
employment, the median primary salaries ranged from a high of $196,051 reported by
those involved in self-employed consulting to approximately $104,00 to $103,00 reported
by those employed in nonprofit organizations and in universities/colleges.
* I/O psychologists in Colleges, Universities, and Professional Schools are typically on
nine-month contracts and frequently supplement their academic income through
consulting and other applied activities.
Figure 1.3: Median Salaries of I/O Psychologists
It is important to note that the salaries for university/college faculty are typically nine-
month salaries. Also, some employed as faculty supplement their primary income with
consulting activities. The salaries are substantially higher in business schools than in
psychology departments, and where Doctorate degrees are granted than in schools where
only bachelors or masters degrees are granted (Khanna, Medsker, & Ginter, 2012).
Having observed this wide variation in activities and work settings, you might well
question whether it makes sense to lump all these professionals together under the title of
“I/O psychologist.” What, if anything, do they share that distinguishes them from other
professionals who do many of the same things? The answer, very simply, is psychology.
While it is true that psychological principles and methods are used widely by other
specialists and that many I/O psychologists focus on particular organizational functions,
I/O psychologists share an important bond with the larger discipline of psychology. By

training, outlook, and professional affiliation, even the most specialized practitioner is
first a psychologist. To understand why this connection is important, we must now take a
closer look at the parent discipline.

The largest and most widely recognized specialties are clinical and counseling
psychology. They provide similar mental health services to individuals and groups for
somewhat different populations (“deviant” vs. “normal”) and in somewhat different
settings (e.g., mental hospitals vs. counseling centers). By tradition, four psychology
specialties are currently recognized by most professional and legal bodies as official
practice fields: clinical, counseling, school, and I/O. Collectively, they account for over
90% of the psychologists who offer services to the public. Of these, I/0 is by far the
smallest (approximately 3,000 practitioners) and most distinct from the others (Rosen-
feld, Shimberg, & Thornton, 1983). Psychologists consist of those who provide direct
healthcare services (health-care providers or HCPs) and those who do not (Howard and
Lowman, 1985). By this definition, I/O psychologists are not HCPs and are not involved
in the diagnosis of and treatment of behavioral and mental disorders; clinical, counseling,
and many school psychologists are. We shall have more to say about this in a moment.

Psychology also is differentiated according to field of specialization. But because there is
no universally accepted basis for defining and classifying specialties, psychology has a
rather long and confusing list of such fields. Some, such as child psychology, are defined
largely by a focal population; others, such as developmental psychology, by the emphasis
on how people change over time; others, such as experimental psychology, by the
scientific methods used; and still others, such as school psychology, by the setting in
which the field is practiced. Like school psychology, I/O psychology is identified
primarily by its focus on particular settings in which the human behavior is studied. In the
case of I/O, these settings are work and the organizations in which the work is performed.

The Scientist-Practitioner Model

There is a scientific component and a practice component to I/O psychology. The same is
true of psychology as a discipline, and this is one of its most distinctive features. By
contrast, other practice fields, such as engineering or medicine, are largely independent of
the scientific disciplines (such as physics and biochemistry) on which they depend for
basic knowledge. Psychology is concerned broadly with how people think, act, and
feel—in more technical terms, with cognition, behavior, and affect. Its scientific
component seeks to understand these functions; its practice component seeks to use what
is learned to benefit individuals, groups, and society. The experimental, developmental,
or social psychology, are dominated by science but a distinguishing characteristic of the
applied psychological specialties such as clinical and school psychology is the scientist-
practitioner model (Howard, Pion, Gottfredson, Flattau, Oskamp, Pfaffdin, Bray, &
Burstein, 1986). Similar to other applied psychological specialties, I/O psychologists
have always taken pride in maintaining a balance between the two orientations, both in its
scope of activities and in the training of its professionals. It prepares the I/O psychologist

to assume either role, and promotes the point of view that a mix of the two is vastly
superior to either for even the most practical of purposes.

The training model promoted by SIOP is that I/O professionals should receive training as
both scientists and practitioners. To enter the field, most I/0 psychologists earn the PhD
degree and acquire supervised experience in one of the nonacademic work settings
described earlier. For years the scientist-practitioner model was the unchallenged
philosophy behind all professional training in psychology, for the health-care practitioner
as well as for the other applied specialties. Today, however, the health-care practitioner
(HCP) community is badly split between those who favor the traditional model and a
growing constituency who prefer a more practice-oriented “scholar-practitioner” or pure
“practitioner” model culminating in a professional doctorate—the PsyD degree (Stricker,
1975). Trivial as arguments over training models and degree labels appear on the surface,
they reflect profound philosophical differences that threaten to reshape the whole
discipline of psychology (American Psychologist, 1987; Schover, 1980). Although these
differences also exist within I/O, the I/O field still stands apart from most of the
psychological specialties in maintaining a firm commitment to combining science and
practice. As a practice, I/O contributes to practical interventions to improve the
effectiveness of people and organizations. As a science, I/O contributes to psychology’s
larger mission of understanding human cognition, behavior, and affect. The importance
of understanding the role of work to a basic understanding of human psychology is made
apparent when one considers that almost 90% of the adults in the industrialized nations
work for organizations (Perrow, 1986). We spend nearly half our waking hours at work,
and what happens there directly affects the way we think, act, and feel (London & Moses,
1990).

If understanding human behavior and experience is its goal, the science of psychology
must concern itself with work-related phenomena. Questions arising from the workplace
are many and fundamental. What does work mean to people? How important are the loss
of a job or opportunities for advancement to their self-esteem? Are attitudes changing on
the relative importance of security, pay, personal growth opportunities, the social
atmosphere at work, and other job features? What constitutes good management? How do
multiple roles affect women employees? What is the effect of dual breadwinners on the
family? How do corporate mergers and acquisitions affect the stress level of employees?
The list is virtually endless. In their role as scientists, many I/O psychologists carry out
research on these basic issues. In addition, by exposing and highlighting such
problems, they encourage the interest of other branches of psychology.

I/O is a Profession as Well as a Job

For a long time, profession was used to refer to a handful of occupations such as the law,
medicine, the clergy, engineering, and accounting. The boundaries between what are
considered professions and nonprofessions are increasingly blurred as a wider variety of
occupations attempt to professionalize (Gorman & Sandefur, 2011). Rather than saying
some occupations are professions and others are not, a more accurate statement is that
professionalism is a continuum with occupations varying on professionalism. Three

essential characteristics define this continuum: a body of specialized knowledge,
technical autonomy, and a normative orientation to the service of others. I/O psychology
is clearly a profession that embraces all of these attributes.

A body of specialized knowledge: expertise
I/O psychology is based on a body of knowledge consisting of the accumulated findings
of scientific research and theory. This knowledge provides the basis for the solution of
practical problems. In contrast, those who base their practical actions on the internet,
commonsense, personal experience, or authority figures are not acting professionally. An
entire chapter in this text focuses on what constitutes science and non-science. I/O
psychology is clearly a science. Not all research and theory becomes part of the body of
knowledge providing the core of I/O psychology. As in any respectable scientific
discipline, the assertions of I/O psychologists are subjected to peer review and rigorous
scrutiny. Most of what I/0 researchers discover is published in journals such as
the Journal of Applied Psychology, Personnel Psychology, Administrative Sciences
Quarterly, Academy of Management Review, Academy of Management
Journal, Organizational Behavior and Human Decision Processes, International Review
of Applied Psychology, Journal of Organizational Behavior, Journal of Business
Psychology, and Journal of Occupational Psychology. Other specialties have their own
professional journals. Articles published in these and the other major I/O journals are
reviewed by other scholars and only accepted for publication if they meet the highest
standards of research excellence. Only 10 – 20% of manuscripts submitted to these
journals are actually accepted and published. Once published, research and theory of I/O
psychologists are subject to challenges that can lead to modification or outright rejection.

A normative orientation to the service of others: ethics
I/O psychology also takes a normative orientation in the form of a code of ethics. Not
everyone who offers psychological services to the public is trustworthy. Most potential
clients, including large organizations, lack sophistication in psychology and are
vulnerable to claims by unscrupulous practitioners (primarily nonpsychologists, the
author is happy to say) who have opportunistically promoted questionable products and
services. Managers faced with chronic problems in morale, production, or turnover find
the promise of a quick and affordable cure irresistible even when the cure is unrealistic.
Consequently, charlatans have prospered, usually at the expense of the organization,
society at large, and the reputation of legitimate psychology. A case in point is the
proliferation of tests and claims made for the validity of these tests. The claims made for
tests that purport to identify dishonest or unproductive employees are frequently
exaggerated, and many decent people have suffered personal indignities and financial
loss as a result of their misuse. As a profession, I/O psychologists in the United States
adhere to the ethical principles of the American Psychological Association, the oldest and
largest professional organization representing psychologists. These general principles are
described as follows:

Ethical principles and standards of the American Psychological Association. The code of
ethical conduct of the American Psychological Association states five general principles

and ten categories of ethical standards. What follows is quoted from the general
principles section of the code (American Psychological Association, February 20, 2010,
pp. 3 – 4).

“General Principles, as opposed to Ethical Standards, are aspirational in nature. Their
intent is to guide and inspire psychologists toward the very highest ethical ideals of the
profession. General Principles, in contrast to Ethical Standards, do not represent
obligations and should not form the basis for imposing sanctions. Relying upon General
Principles for either of these reasons distorts both their meaning and purpose.”
“Principle A: Beneficence and Nonmaleficence. Psychologists strive to benefit those with
whom they work and take care to do no harm. In their professional actions, psychologists
seek to safeguard the welfare and rights of those with whom they interact professionally
and other affected persons, and the welfare of animal subjects of research. When
conflicts occur among psychologists’ obligations or concerns, they attempt to resolve
these conflicts in a responsible fashion that avoids or minimizes harm. Because
psychologists’ scientific and professional judgments and actions may affect the lives of
others, they are alert to and guard against personal, financial, social, organizational, or
political factors that might lead to misuse of their influence. Psychologists strive to be
aware of the possible effect of their own physical and mental health on their ability to
help those with whom they work.”

“Principle B: Fidelity and Responsibility. Psychologists establish relationships of trust
with those with whom they work. They are aware of their professional and scientific
responsibilities to society and to the specific communities in which they work.
Psychologists uphold professional standards of conduct, clarify their professional roles
and obligations, accept appropriate responsibility for their behavior, and seek to manage
conflicts of interest that could lead to exploitation or harm. Psychologists consult with,
refer to, or cooperate with other professionals and institutions to the extent needed to
serve the best interests of those with whom they work. They are concerned about the
ethical compliance of their colleagues’ scientific and professional conduct. Psychologists
strive to contribute a portion of their professional time for little or no compensation or
personal advantage.”

“Principle C: Integrity. Psychologists seek to promote accuracy, honesty, and
truthfulness in the science, teaching, and practice of psychology. In these activities
psychologists do not steal, cheat, or engage in fraud, subterfuge, or intentional
misrepresentation of fact. Psychologists strive to keep their promises and to avoid unwise
or unclear commitments. In situations in which deception may be ethically justifiable to
maximize benefits and minimize harm, psychologists have a serious obligation to
consider the need for, the possible consequences of, and their responsibility to correct
any resulting mistrust or other harmful effects that arise from the use of such
techniques.”

“Principle D: Justice. Psychologists recognize that fairness and justice entitle all persons
to access to and benefit from the contributions of psychology and to equal quality in the
processes, procedures, and services being conducted by psychologists. Psychologists
exercise reasonable judgment and take precautions to ensure that their potential biases,
the boundaries of their competence, and the limitations of their expertise do not lead to
or condone unjust practices.”

“Principle E: Respect for People’s Rights and Dignity. Psychologists respect the dignity
and worth of all people, and the rights of individuals to privacy, confidentiality, and self-
determination. Psychologists are aware that special safeguards may be necessary to
protect the rights and welfare of persons or communities whose vulnerabilities impair
autonomous decision making. Psychologists are aware of and respect cultural,
individual, and role differences, including those based on age, gender, gender identity,
race, ethnicity, culture, national origin, religion, sexual orientation, disability, language,
and socioeconomic status and consider these factors when working with members of such
groups. Psychologists try to eliminate the effect on their work of biases based on those
factors, and they do not knowingly participate in or condone activities of others based
upon such prejudices.”

In addition to the general principles, the American Psychological Association’s ethical
code sets forth ethical standards which are enforceable.

1. Resolving ethical issues. Psychologists take reasonable steps to resolve ethical
issues. Particularly relevant to I/O psychologists is standard 1.03. According to
this standard, “If the demands of an organization with which psychologists are
affiliated or for whom they are working are in conflict with this Ethics Code,
psychologists clarify the nature of the conflict, make known their commitment to
the Ethics Code, and take reasonable steps to resolve the conflict…Under no
circumstances may this standard be used to justify or defend violating human
rights (American Psychological Association, 2010, p. 4).

2. Competence. Psychologists stay within their areas of competence in providing
services, teaching, consulting, and conducting research. When they lack
competence in an area, they take steps “to obtain the training, experience,
consultation, or supervision necessary to ensure the competence of their services,
or they make appropriate referrals (American Psychological Association, 2010, p.
5).

3. Human relations. Psychologists do not engage in unfair discrimination and sexual
harassment and do not harm, demean, or harass those with whom they have
professional relationships. They also avoid multiple relationships that might
impair, exploit, or harm persons with whom they have professional relationships.
They provide informed consent to research participants and clients. Especially
relevant to I/O psychologists employed by organizations provide information to

clients prior to the services that includes the nature and objectives of the services,
who will have access to information gathered, and any limits to confidentiality.

4. Privacy and confidentiality. Psychologists take reasonable steps to protect the
confidentiality of any information gathered in their research and consulting. If
they disclose confidential information they obtain the consent of the persons
involved. There are exceptions such as where the law mandates or permits
disclosure and disclosure is necessary to fulfill valid objectives such as protecting
the client or providing services. In their public communications in lectures, media
interviews, and writings, they take steps to conceal the identity of clients,
including research participants, and organizations or obtain consent for the
disclosures.

5. Advertising and other public statements. Psychologists can advertise their services
but they take care to convey truthful information and avoid false, deceptive, or
fraudulent statements. They take steps to ensure that their public statements are
used appropriately by the media and they make sure that their training, education,
licensing, and other credentials are accurately described.

6. Record keeping and fees. Psychologists create and maintain records to facilitate
the provision of services by other professionals, to allow replication of research,
to ensure accuracy in billing, and to comply with the law.

7. Education and training. Teachers and other educators make sure that their courses
and programs are accurately described, provide the appropriate knowledge and
experiences, and convey the content set forth in course syllabi and program
descriptions. Instructors protect the confidentiality of students and in their
teaching convey current, accurate information on the topic of the instruction.
“Psychologists do not engage in sexual relationships with students or supervisees
who are in their department, agency, or training center or over whom
psychologists have or are likely to have evaluative authority” (American
Psychological Association, 2010, p. 10).

8. Research and publication. Researchers obtain informed consent from those
participating in their research. The informed consent states the purpose of the
research, possible consequences, any potential risks or harm, and provides
persons who can be contacted if there are questions. In informed consent
participants can refuse to participate or withdraw during the study. Both the
benefits and costs of the research are described, and any limits to confidentiality.
There are some exceptions in which informed consent is not required such as
where the informed consent might create harm, naturalistic observations, or
research on educational practices or curricula in educational settings. Also,
informed consent is not required in “the study of factors related to job or
organization effectiveness conducted in organizational settings for which there is
no risk to participants’ employability, and confidentiality is protected” (American
Psychological Association, 2010 p. 10).

9. Assessment. When psychologists convey opinions, recommendations, and
assessments they base their statements on information and techniques that are
sufficient and adequate to substantiate their conclusions. Where they cannot
obtain adequate information, they make clear the impact on the validity and
reliability of their judgments and limit the conclusions they draw. “Psychologists
use assessment instruments whose validity and reliability have been established
for use with members of the population tested. When such validity or reliability
has not been established, psychologists describe the strengths and limitations of
test results and interpretation” (American Psychological Association, 2010, p.
12). Psychologists obtain informed consent from participants who are assessed
when it is mandated by law or governmental regulations or “implied because
testing is conducted as a routine educational, institutional, or organizational
activity (e.g., when participants voluntarily agree to assessment when applying
for a job)” (American Psychological Association, 2010, p. 12). Psychologists
explain assessment results to those assessed. When this is not possible the lack of
explanation is explained to those assessed prior to the assessment.

10. Therapy. This is relevant to clinical psychologists, but is not relevant to most I/O
psychologists. Engaging in therapy violates the boundaries of the competence for
most I/O psychologists, unless they obtain special training and supervision in
clinical psychology. Likewise, clinical psychologists do not engage in I/O
psychological practices unless they have the necessary training and experience.

Ethical principles in other countries. Similar to the U. S. Psychological Principles
and Standards, psychologists in many other countries have ethical codes to guide
their research and practice. The International Union of Psychological Science has
identified over sixty ethical codes associated with professional organizations
representing psychologists around the world (see
http://resources.iupsys.net/iupsys/index.php/ethics/compendium-of-codes-of-
ethics-of-national-psychology-organizations). Although the content of these codes
and the procedures used in enforcing them varies across professional
organizations, the basic objective is the same: psychologists impose on
themselves standards and principles of conduct as part of their self-governance.

Technical autonomy: self-governance in professional associations
The members of a profession attempt to exert autonomy in their research and practice by
joining together in professional associations. These associations establish professional
standards and ensure that members adhere to them. The Society for Industrial and
Organizational Psychology (SIOP) is the major professional association representing I/O
psychologists in the United States. Similar associations are found in other countries.
SIOP is a division within American Psychological Association (APA) and adheres to the
ethical and other professional standards that apply to all APA members such as the
ethical principles and standards. In addition, SIOP has generated other guidelines and
standards that apply more specifically to I/O psychologists. These include guidelines for
training in I/O and principles for the validation of selection procedures.

http://resources.iupsys.net/iupsys/index.php/ethics/compendium-of-codes-of-ethics-of-national-psychology-organizations

Table 1.2 lists the competencies that I/O psychologists should have as recommended by
the Society for Industrial and Organizational Psychology (SIOP), the field’s chief
professional organization. As you can see, the list runs the gamut from HRM functions
(e.g., staffing, compensation) to OD functions (e.g., attitude surveys, conflict resolution).
HRM and OD approaches originated in different philosophies, but they have moved
closer together in practice.

Core competencies:

1. Ethical, Legal, and Professional Issues
2. Statistical Methods/Data Analysis
3. Measurement of Individual Differences
4. Criterion Development
5. Job Analysis Job and Task Analysis
6. Performance Appraisal
7. Employee Selection,
8. Training:
9. Multivariate Statistics
10. Work Motivation
11. Attitude Theory
12. Small Group Theory and Process
13. Organization Theory
14. Organizational Development
15. Research Methods

Optional (desirable but not essential):
1. Career Development Theory
2. Human Performance/Human Factors
3. Consumer Behavior
4. Compensation and Benefits
5. Industrial and Labor Relations

Table 1.2: Competencies that an I/O Psychologists Should Possess According to the
Society for Industrial and Organizational Psychology

A competent practitioner might prefer one set of techniques to another, but it is unlikely
that he or she would fail to consider the possibilities afforded by the alternative set. There
is no standard way of packaging either set of functions within the formal structure of
work organizations. Some firms have personnel departments that perform both OD and
HRM functions; some house HRM activities in one office (e.g., personnel) and OD
functions in another (e.g., management development); some maintain only one set of
functions; some maintain neither and rely heavily on outside consultants or Inside
nonprofessionals.

The practice of psychology also is subject to both legal regulation, like medicine or law.
State laws require that psychologists meet certain educational and competency standards
to practice, and special licensing boards set and enforce these standards. Additionally, a
professional ethics code mandates staying within one’s specialized areas of competence.
Thus, at least some assurance of competence is available merely by determining whether
a practitioner is licensed or certified as a psychologist by the state. Since states differ in
their policies regarding licensure for nonhealth care-practitioner (HCP) specialties, only
about a third of the I/O psychologists in the U.S. are licensed (Howard, 1990). From the
consumer’s standpoint, a simple call to the ethics committee of the state psychological
association or the American Psychological Association (APA) can often establish the
legitimacy of services or credentials being offered by a practitioner.

Becoming an I/O Psychologist

The trend in psychology, like most scientific disciplines, has been toward increasingly
narrow specialization, a condition that has limited this sort of healthy interaction
(Altman, 1987). If one goes back earlier in the history of I/O psychology, one will find
far fewer educational programs dedicated specifically to the field of I/O. One will also
find that many self-proclaimed I/O psychologists did not even specialize in the field
during their graduate training. Today most of those identifying themselves as I/O
psychologists received their training in one of the I/O Ph. D. or master’s programs.

So what do you need to do if you want to pursue a career in I/O? First of all, you will
need a graduate degree. Although there are positions available to those with only a
bachelor’s degree in some Human Resource Management positions, becoming an I/O
psychologist requires either a masters or a doctorate degree. The typical Ph. D. program
requires about 5 years. The first two years consists of course work culminating in a
Master’s Thesis in year 2 or 3. Some programs require an internship along the way. In
year 3 or 4 students must pass comprehensive exams (sometimes referred to as qualifier
or candidacy exams). In years 4 and 5 students propose and conduct dissertation research
and defend the written dissertation before their committee. The doctoral committee
consists of four or five faculty persons, one of whom supervises the student. Masters
programs are usually 2 or 3 years in duration and consist primarily of coursework and
typically an internship or some other practical experience. Some terminal master’s
programs also require a written thesis consisting of original research and defended before
a master’s committee.

If you are considering a masters or Ph. D. program in I/O psychology you should take an
I/O psychology course as an undergraduate. Not all graduate programs require this but
taking the course will provide a realistic preview of what the field is about and what I/O
psychologists do. If after taking the course you are still interested in becoming an I/O
psychologist, you should prepare with a well-rounded set of undergraduate courses. A
good combination would be a psychology degree with courses in statistics and research
methods. Experience in research is also a plus. What is not necessary is a specialization
in I/O at the undergraduate level. One, possibly two courses in I/O are enough. A broad

liberal arts background is better preparation as an undergraduate. Specialization comes
once you are in graduate school. Majoring in psychology is not always required,
although it seems to be a requirement in the majority of graduate programs in I/O. If you
do not major in psychology, you should at least take introductory, social, statistics, and
research methods.

It is also important to acquire some research experience as an undergraduate, perhaps by
completing an undergraduate thesis or working in a faculty person’s laboratory either for
coursework or on a volunteer basis. Any area of psychology will typically suffice. Most
graduate programs in the author’s experience do not require that you have research
specifically in I/O psychology. Because letters of reference are important in gaining
admission to a program, you should become involved in activities that will establish
contact with faculty. Join Psi Chi (the honorary psychology society) and the Psychology
Club in your department. Volunteer to help out in departmental events, interact with
graduate students, and attend departmental presentations by faculty in your department
and invited guest speakers. Having some practical experience in a business or some other
organization is a nice touch but in the experience of the author, admission decisions are
primarily based on GPA, GRE scores, research experience, and faculty references.

Needless to say you will need good GRE scores and grades. It is strongly recommended
that any undergraduate thinking of applying to graduate schoolwork as a teaching and/or
research assistant for a faculty person. Most graduate programs require the Graduate
Record Exam and a few also require the specialized exam in psychology. It is wise to
take the GRE in your junior year so that you will have time to repeat the exam if your
scores the first time around are low. Apply to graduate programs in the fall of your senior
year. Most schools have their deadlines from December through February. Particularly
important are your recommendations. Letters from graduate students, friends, ministers,
and family count for little in graduate admissions. This brings me back to the importance
of working for a faculty person as a research and/or teaching assistant. Working for an
I/O psychologist is not especially important. More important is being able to show that
you have worked in a lab collecting, coding, and analyzing data.

Also important is the personal statement in the application. Obtain feedback from peers,
graduate students, and faculty advisors. This statement should address why you are
interested in I/O psychology, your research experiences, and your goals. Applicants
should probably be overly specific in stating future goals. If you are absolutely sure that
you want to go to work for a large computer firm in the HR department, go ahead and
state that. But most students modify their aspirations as they progress through their
graduate programs. Allow some room for changing your mind. If you are interested in
working with a particular faculty person, it is worth mentioning this, but it is often wise
to mention more than one in the event that this individual is not accepting students. Do
not try to be cute in your personal statement. In my experience, personal statements that
go into too much personal detail can hurt the applicant’s chances. Admissions
committees are more interested in how likely the applicant will demonstrate seriousness
and dedication to research and coursework as a graduate student. They are not interested
in applicant rather than his hobbies, travels, and love life.

A variety of social science disciplines are involved in the attempt to understand human
behavior at work and in organizations in addition to psychology. These include
anthropology, sociology, economics, communications science, labor relations, and
management, to name a few. Each of these fields offers specialization that resemble in
some respects the focus of I/O psychologists on work and organizations. Business
schools offer programs in Organizational Behavior (OB) and Human Resource
Management (HRM) that attempt to provide an interdisciplinary coverage by employing
as faculty not only I/O psychologists but a variety of other social scientists. If you are
especially interested in the application of psychology to organizations and work,
obtaining a degree in I/O psychology is the preferred route to take. However, if you are
more interested in an interdisciplinary approach to the topic and want more exposure to
the fields of business and management, then you should seriously consider an OB or
HRM degree in a business school.
The Future Looks Bright for I/O Psychology

As noted in Figure 1.5, the United States Bureau of Labor Statistics declared in 2014 that
I/O psychology ranked number one in the 20 fastest growing occupations in the United
States (http://www.bls.gov/emp/ep_table_103.htm). This ranking was based on
projections of where the occupation was projected to be in terms of number of employees
by 2022. This projected growth reflects in part the fact that I/O is the smallest of the
occupations listed below, so that large percentage increases can be shown from relatively
small absolute increases. However, the attention given to I/O also reflects the growing
importance in a variety of organizations of the skills that I/O psychologists possess.

The growth of I/O psychology is also indicated by the rapid globalization of the
applications of scientific psychology to work and organizations. In addition to all the
European and North American countries, one can find I/O research and application
underway in these countries and many others around the world
(http://www.siop.org/tip/apr12/12thompson.aspx).

*** China (http://www.siop.org/tip/oct11/16thompson.aspx),
*** India (http://www.siop.org/tip/April10/20thompson.aspx),
*** Peru (http://www.siop.org/tip/april11/11thompson.aspx)
*** Greece (http://www.siop.org/tip/oct09/12thompson.aspx)
*** Croatia (http://www.siop.org/tip/july12/16thompson.aspx)
*** Ghana (http://www.siop.org/tip/Jan13/13_spotlight.aspx)
*** Philippines (http://www.siop.org/tip/oct12/13spotlight.aspx)

http://www.bls.gov/emp/ep_table_103.htm

http://www.siop.org/tip/apr12/12thompson.aspx

http://www.siop.org/tip/oct11/16thompson.aspx

http://www.siop.org/tip/April10/20thompson.aspx

http://www.siop.org/tip/april11/11thompson.aspx

http://www.siop.org/tip/oct09/12thompson.aspx

http://www.siop.org/tip/july12/16thompson.aspx

http://www.siop.org/tip/Jan13/13_spotlight.aspx

http://www.siop.org/tip/oct12/13spotlight.aspx

Figure 1.4: I/O Psychology is the Fastest Growing Occupation

One of the strengths of I/O psychology is that there are a wide variety of job
opportunities available to those who enter the field. Academic I/O psychologists are
found in not only psychology departments, but also in business schools, communication
departments, medical schools, and even law schools. Indeed, there is concern among
some in the I/O field that most graduates in I/O are pursuing business school positions
where the salaries are higher than psychology faculty positions. Although business
schools do sometimes prefer that faculty have degrees in business areas, most of the top
business schools have hired I/O psychologists because of their expertise in research and
the quality of their training. I/O psychologists who pursue non-academic careers are
found in a very diverse range of work settings including consulting firms, business
organizations, the military, schools (as human resource management or organizational
development specialists), and even in religious organizations. The minimum degree that
you need to practice I/O psychology in nonacademic settings is a master’s degree, but

many applied positions require a Ph. D. This is especially true of the management
consulting firms. Masters level I/O psychologists are more often found in positions
involving a narrower application of technical skills such as in test development. Most
academic positions require a Ph. D. The exceptions are community colleges and some
teaching oriented colleges.

Some points to ponder

1. Organizational Science was identified as a multi-disciplinary effort to understand the
behavior of humans at work and the organizations in which they are employed. How do
you think a psychologist might differ in his or her approach to this endeavor from a
sociologist, an economist, a political scientist, or an anthropologist?
2. Review the activities of I/O psychologists listed in figure 1.1. Which of these would
you most and least like to do?
3. Explore the websites for the following three professional associations of psychologists
involved in the scientific study of the psychology of work and organizations:
a. the Society for Industrial and Organizational Psychology (http://www.siop.org)
for U. S. psychologists.
b. the European Association of Work and Organizational Psychology (EAWOP)
for European psychologists (http://www.eawop.org).
c. the College of Organisational Psychology (COP) for Australian psychologists
(http://groups.psychology.org.au/cop/)
Identify similarities and differences in how these websites describe the profession.
4. Review some of the ethical codes of psychologists around the world. Here are 60
codes. What are some of the similarities and differences among the codes?
http://resources.iupsys.net/iupsys/index.php/ethics/compendium-of-codes-of-ethics-of-
national-psychology-organizations
5. Assume for the moment that you have decided to become an I/O psychologist. Outline
your strategy for preparing yourself for this career.
6. Why do you think I/O psychology is now among the fastest growing occupations
according to the U. S. Bureau of Labor Statistics?

Conclusions

I/O psychologists are a diverse lot—educators, researchers, consultants, and technical
specialists who perform a variety of OD and HRM functions for management in virtually
every kind of organization. What binds them together and distinguishes them from others
who fill similar roles is their grounding in the specialized knowledge, methods,
institutions, and culture of psychology. You now have a general understanding of what
I/O psychology is and what it is not. It is a specialty within the general field of
psychology and it is this emphasis on psychology that distinguishes it from related fields
such as Organizational Behavior and Human Resource Management. It adheres to the
scientist-practitioner model and as such is concerned with not only science but also
practice. As a profession, it adheres to a code of ethics governing both research and
practice. And the future looks bright with I/O psychologists found in a wide variety of
work settings and involved in a diverse set of activities. The next chapter provides an

http://www.siop.org/

http://www.eawop.org/

http://groups.psychology.org.au/cop/

http://resources.iupsys.net/iupsys/index.php/ethics/compendium-of-codes-of-ethics-of-national-psychology-organizations

overview of the history of I/O psychology and how it developed in the context of
historical events and societal trends.

CHAPTER 2: A HISTORY OF I/O PSYCHOLOGY

Introduction

This chapter reviews the development of I/O psychology from the first crude attempts at
application in the late 1800s to the sophisticated profession that constitutes modern I/O
psychology. Understanding how the field developed requires a consideration of the
contexts in which I/O psychologists have conducted their activities at various points in
time over the last 150 years. These contexts include the historical and societal events,
management theory and practices, and the development of the larger discipline of
psychology.

Figure 2.1 summarizes this evolution of I/O psychology in the context of historical and
societal events, the development of psychology as a discipline, and thinking about
organizations and the management of people in these organizations. The boxes outside
the circle anchor four periods (early, middle, post WWII, modern) and the historical and
societal events occurring during these years. The chapter begins with the early years from
1870 – 1920 defined by the industrial revolution and the rise of the corporation. This era
ends with WWI, a very significant event in the development of I/O psychology. The end
of WWI signals the beginning of the middle years. During these years there was the
economic boom of the 1920s followed by the economic bust of the Great Depression in
the 1930s and still another war, WWII. After the end of WWII came the postwar years in
which the U. S., the dominant economic and military power in the world was engaged in
a cold war with the Soviet Union and the People’s Republic of China. This competition
among the great powers spawned space exploration and unprecedented progress in
science and technology. Also occurring during this period was the Civil Rights
movement, which had a huge impact on I/O psychology. Finally, the end of the cold war
and the end of the Soviet Union marked the beginning of the modern period. This period
continues to the present time and is characterized by global competition, the development
of computer technology and the internet, diminishing resources, and population
explosion. Although the great powers are no longer threatening each other with nuclear
annihilation, a new threat has emerged in the form of terrorism.

Over these four periods psychology emerged as a major discipline and I/O evolved as a
subdiscipline within psychology. The development of I/O is represented by outer edge of
the circle and is punctuated by some major contributions of I/O psychologists to meeting
the needs associated with historical and societal events occurring at the time. For
instance, with the outbreak of WWI psychologists developed the first paper-and-pencil
tests of intelligence to allow screening of prospective soldiers. The inner circle
summarizes influential theories in psychology and management that were associated with
what was happening in the field of I/O and the world at large. This timeline provides a
context for the development of I/O psychology, but do not interpret this as a rigid
chronology. For instance, the classical theories of management were dominant during the
early years and less so today. Yet, this approach to thinking about management of people
in organizations did not suddenly end with the middle years. Indeed, classical theory and
scientific management still influence thinking about how to manage people in
organizations.

*The dominant theoretical orientations are in the inner circle. Some of the
major contributions at each period in history are in the outer circle.

Figure 2.1: The Evolution of Industrial and Organizational Psychology in
the Context of Historical Events and Theory

Still another caveat to consider when reviewing this chronology of events

in I/O psychology is its U. S. centric perspective. This admitted bias is because
researchers and practitioners from the United States overwhelmingly dominated I/O
psychology in the early and middle years. This has changed dramatically over the last 2 –
3 decades. I/O psychology is today a multi-national and cross-cultural discipline. There
are active I/O societies on all the continents and the number of articles authored by those
outside the U. S. and published in the major I/O journals are increasing at a rapid rate.
However, because the focus of this chapter is on the past rather than the future of I/O, the
developments occurring in the U. S. are the primary focus. For an excellent and detailed
account of the development of I/O psychology in the U. S. and around the world, the
reader is referred to A History of Industrial and Organizational Psychology (Bryan and
Vinchur, 2012).

The Early Years (1880 – 1920)

Industrial and organizational psychology emerged at the time of the transition of the U. S.
economy from an agrarian society to an industrial society. These early years were marked
by mass production, the rise of the corporation, consumerism, and immigration to the
United States to meet the demand for labor. At the same time psychology was emerging
as a discipline and some of the early psychologists looked for applications to the rapidly
changing workplace. The dominant thinking about managing people in the organization
was what McGregor would call Theory X. Classical and scientific management theorists
emphasized efficiency and felt that employers could obtain the cooperation of workers by
assigning them to specialized tasks, closely supervising their performance of these tasks,
and motivating them with financial incentives.

What were the major forces shaping work during the early years?

Industrialization and mass production. The Industrial Revolution began in the late 1700’s
with technological innovations such as the cotton gin and steam gin
(https://en.wikipedia.org/wiki/Industrial_Revolution). It continued through the 1800s and
early 20th century with the invention of the telephone, the sewing machine, the
automobile, the incandescent light bulb, the diesel engine, the airplane, the Bessemer
process and open hearth in steel making, and the spread of the telegraph. Unlike in
Europe where railroad systems connected existing cities, the rapidly growing network of
railroads was responsible for creating cities in the United States.

Although the invention of new machines was a driving force of the industrial revolution,
innovations in the management of people were as important as advances in hard
technology. A crucial development in management was the idea that the work that
employees perform is best divided into narrow, specialized tasks. This was not
“discovered” in the 1700s but is found in one form or another throughout history. What
was unique about what happened in the 1700s was that the division of labor principle was
refined, systematized, and formalized. One of the earliest applications was by Josiah
Wedgewood who revolutionized the manufacture of pottery in the 1700s (McKendrick,
1961). The old methods of pottery making were becoming obsolete in meeting consumer

https://en.wikipedia.org/wiki/Industrial_Revolution

demand. “The family craftsman stage had already given way to the master potter with his
journeymen and apprentices recruited from outside the family, and this in turn was
becoming inadequate to deal with the growing complexity of potting production”
(McKendrick, 1961, p. 31). Wedgewood introduced one of the first examples of the
assembly line. “His designs aimed at a conveyor belt progress through the works: the kiln
room succeeded the painting room, the account room the kiln room, and the ware room
the account room, so that there was a smooth progression from the ware being painted, to
being fired, to being entered into the books, to being stored. Yet each process remained
quite separate” (McKendrick, 1961, p. 32). Rather performing all the tasks that
constituted this assembly line, each worker specialized in one specific task. “His
workmen were not allowed to wander at will from one task to another as the workmen
did in the pre-Wedgwood potteries. They were trained to one particular task and they had
to stick to it” (McKendrick, p. 32). In addition to division of labor, Wedgwood also
innovated the use of piece rate incentive systems in which workers were paid by the unit
of production, the organization of supervision in the hierarchical layers, and the use of
rules in the discipline of workers. The concepts of division of labor and specialization of
labor had become well known enough that Adam Smith (1776) in Wealth of Nations, his
famous treatise on economics, attributed much of the prosperity achieved in the civilized
nations to the application of these principles by manufacturers.

The division of labor, the assembly line, and other management innovations introduced
by Wedgewood and others were notable but were not widespread. It was not until the late
1800s and early 1900s that these techniques emerged as dominant methods of organizing
and managing work. The most publicized and famous of the applications of the assembly
line was in 1913 when Henry Ford introduced a moving assembly line procedure at his
Highland Park, New Jersey plant in the U. S. (see photo below; see also
http://www.pbs.org/wgbh/aso/databank/entries/dt13as.html).

Assembly line in Ford plant

Numerous books and articles have attributed to the assembly line the increase in
productivity and decreases in costs that occurred in manufacturing during the first few
decades of the 1900s. Reductions in the cost of what was produced allowed the purchase
of all sorts of goods by the working and lower socioeconomic classes and fueled an
economic boom. A bundle of innovative management practices that included incentives,

http://www.pbs.org/wgbh/aso/databank/entries/dt13as.html

rules, and supervisory approaches accompanied the introduction of the assembly line at
the Ford plants (Wilson & Mckinlay, 2010; Williams, Haslam, Williams, Adcroft, &
Johal, 1993). Part of the bundle of changes introduced by Ford was an increase in wages
for workers who worked on his lines. In 1914 he decided to double the wages of his
assembly line workers to $5.00 a day, a remarkable increase in wages for that time. He
also reduced the workday to eight hours a day from the 10 – 12-hour day that was more
typical in the early 1900s. The increased wages and reduced working hours attracted huge
numbers of workers to his plants and along with the assembly line procedures allowed the
mass production of vehicles. The increased wages also created customers for these
automobiles. These innovations were criticized severely by the business community.
After seeing the positive impact on productivity, other corporations were soon imitating
Ford’s increase in wages and reduced working hours. (Open this link to read more:
http://www.thedailybeast.com/articles/2014/01/06/henry-ford-understood-that-raising-
wages-would-bring-him-more-profit.html).

Massive waves of immigration. The factories that were being built in the booming cities
of the U. S. needed workers. The demand for labor led to a huge wave of immigration of
Europeans to the U. S. (27.5 million immigrants between 1865 and 1918, 89% from
Europe) and a movement of farmworkers within the U. S. to the cities to work in the
mills. Throughout the world, but especially in the United States, technological
innovation transformed what had been a predominately rural and agricultural existence
into societies dominated by large cities and manufacturing.

Rise of the corporation. As important as the technological innovation and the dominance
of manufacturing and city life was the rise of the large corporation. Prior to the industrial
revolution, the dominant work organization was a small entrepreneurial firm in which the
owners were the managers. To deal more efficiently with the larger scale of operation
that characterized industry in the mid to late 1800s, a new organizational entity emerged
in which the owners were shareholders with little contact with the firm and managers
were hired to run the firm as agents of the shareholders. This new form of organization is
called the corporation. The corporation as a legal entity was controversial when it first

http://www.thedailybeast.com/articles/2014/01/06/henry-ford-understood-that-raising-wages-would-bring-him-more-profit.html

emerged 150 years ago and remains controversial. For portions of a recent documentary
on these controversies see:

A highly significant event occurring in the U. S. in 1886 was the Santa Clara County V.
Southern Pacific Railroad Supreme Court decision that gave corporations the same rights
and protections given to real people under the U. S. Constitution (http://www.truth-
out.org/opinion/item/375:unequal-protection-the-deciding-moment). This included the
right of management to use the corporation’s wealth to influence the government. It is
this decision that set the precedent for the U. S. Supreme Court’s recent decision to
declare attempts to reform campaign finance as a violation of the constitutional right of
free speech. During the latter part of the 1800s, corporations were subjected to very little
oversight by government and could take ruthless actions to eliminate their competition
and create monopolies. They could also crush attempts to form unions by firing,
harassing, or even physically threatening workers who tried to organize. The period in the
United States that followed the end of the Civil War in 1865 and continued until the early
1900s has been called the era of the “robber barons” such as Andrew Carnegie, John D.
Rockefeller, and Jay Gould. These men amassed enormous wealth by destroying their
competitors, subjugating their employees, and ignoring public interests.

Rise of consumerism. During the late 1800s and early 1900s, American industry was
enjoying the combined benefits of a host of positive business conditions. These included
technological innovations such as mass production, seemingly unlimited natural
resources, cheap labor immigrating from Europe, strong government support, expanding
markets, and a social climate that lionized material success: All the ingredients for rapid
growth and prosperity were in place. The great American experiment was working!
Evolving business organizations, like the country itself, saw themselves as virtually
invulnerable to outside forces. If a firm, or an individual for that matter, did not succeed,
only it (or he) was to blame. Initiative and growth could overcome virtually any obstacle
that might arise. With natural resources and cheap labor available and virtually no
restrictions on corporations, the U.S. became the economic, industrial, and agricultural

http://www.truth-out.org/opinion/item/375:unequal-protection-the-deciding-moment

leader of the world. The rise of the large urban area and manufacturing brought pollution,
poverty, crime, and disease and large economic inequalities, but also dramatically
improved the economic status of the industrial worker. The average annual income of
nonfarm worker improved dramatically with the average annual income rising by 75%
from 1865 to 1900, and by 33% from 1901 to 1918 (U.S. Bureau of the Census,
Historical Statistics of the United States (1976) series D726 and D736 pp. 164–5). The
availability of cheaper goods coupled with the rising income of workers, led to
consumerism. Consumerism is often defined as an economic and social phenomenon in
which people are driven to purchase more than they need. Consumerism has had critics
(e.g., the sociologist Thorstein Veblen (1934/1899) coined the term conspicuous
consumption to refer to irrational buying behavior of people seeking status by means of
possessions). Despite the critics, there is no doubt that consumerism fueled the economic
growth and increasing prosperity of the late 1800s and early 1900s. Supporters of
consumerism argued that the creative waste of money on goods that were luxuries fuels
economic progress and raises the living standard of all socioeconomic classes. The
practice of purchasing on installment plans became a popular form of buying. To create
the demand that would lead people to go beyond their means in making purchases,
advertising and marketing emerged as professions. Advertising changed from merely
informing consumers of the existence of various products, to creating the need for these
products. Psychology became the source of ideas for advertising, and early psychologists
such as Walter Dill Scott (http://faculty.frostburg.edu/mbradley/psyography/bioscopes_walterscott.html) and
John B. Watson (http://advercloud.com/Movers-shakers/John-B-Watson.php) were pioneers in the study
of consumer behavior.

What were the dominant theories of how to organize and manage work?

As a consequence of what was occurring in the 1800s and early 1900s, a variety of
disciplines began to give attention to what constitutes the most effective ways of
managing work. To successfully compete in the early years required that the management
of organizations were able meet the demands for producing goods in the most efficient
way possible. Classical Organizational Theory (also called Administrative theory) and
Scientific Management emerged as the two dominant theories. There were differences
between the two types of theory. For instance, the Classical Organizational Theories
focused more how to structure reporting relationships and the entire organization,
whereas Scientific Management was focused more on how to organize tasks and motivate
the employees performing these tasks. Despite these differences, the two approaches
were quite similar. They both were closed systems approaches that focused inward
oriented and ignored the world outside the organization or the context in which workers
performed their tasks. Both approaches tended to state universal rules for managing work
and organizations that were believed to apply in all situations. Both stressed the
management of the individual worker and mostly ignored the fact that workers belonged
to groups. Both emphasized monetary incentives and paid little or no attention to the
other factors that might motivate a worker. Organizations and the people that belonged to
these organizations were assumed to operate in a rational manner. The assumption was
that if management structured the situation so that workers benefitted from working hard
to achieve organizational goals, these workers act rationally and do what management

http://en.wikipedia.org/wiki/Thorstein_Veblen

http://faculty.frostburg.edu/mbradley/psyography/bioscopes_walterscott.html

http://advercloud.com/Movers-shakers/John-B-Watson.php

wants them to do. Perhaps the most important common feature was the belief that
productivity and efficiency were achieved with a strict hierarchy of authority in which
managers had the right to give orders and subordinates were obligated to obey.

Classical Organizational Theory and Scientific Management reflected what some
corporate leaders were already doing to increase efficiency and productivity. Thinking
about to organize and manage work had been around for centuries (e.g., read Exodus 18:
17 – 27 in the Bible to see classic organizational principles at work). After all, work
organizations have existed throughout recorded history, and it is not surprising that
advice would exist on how to create and manage these organizations. The advice on how
to management organizations, however, was not seriously investigated until the
nineteenth and twentieth centuries. The classical and scientific management approaches
that emerged during 1870 – 1920 were explicit and systematized systems of thought that
delved more deeply into the underlying dynamics of organizations and work than any
previous theorizing. These first formal theories of organization were presented by both
practitioners, such as chief executive officers, and scholars from a variety of academic
disciplines, including psychologists, sociologists, economists, and political scientists.
Three theories that are most representative of the Classical and Scientific Management
approaches were Henri Fayol’s Administrative Management Theory, Frederick Taylor’s
Scientific Management, and Max Weber’s Ideal Bureaucracy.

Fayol’s administrative management theory. The existing knowledge about work
organizations derives from a variety of sources, both scientific and experiential. First,
came the attempts by practicing managers to capture their insights in the form of
universal management principles. Probably the most influential of these was a book first
published in 1916 by a French manager, Henri Fayol, Administration Industrielle et
Generate, in which he spelled out 14 principles that he saw as key to successful
management (Fayol, 1936). Included were such basic notions as

** Division of labor: every employee performs a specialized set of activities.
** Clear chain of command: require that all communications flow through the hierarchy
so that no subordinate bypasses his or her boss but goes through that boss
** Subordination of individual interests to those of the organization.
** Unity of command: each employee should report to one and only one supervisor.
** Parity of authority and responsibility: every employee is given just enough authority
to carry out his or her responsibilities..no more, no less.
** Centralization: those higher in the organizational hierarchy call the shots and have the
largest amount of authority.
Fayol’s book was followed by a number of other books that carried much the same
message. This philosophy of management became known as classical organization (or
administrative or management) theory (Bedeian, 1980; Dessler, 1980).

Henri Fayol
Collectively these works formed the basis of academic training in management from its
beginnings in the 1920s through the l950s. The classical philosophy focused on the
formal structure of organizations: clear and efficient division of labor, lines of authority,
and communication. The idea was to divide work up into clearly defined jobs, put them
together in the most efficient way, and make sure workers performed according to plan
by imposing a military-like system of authority over them. People were seen as mere
instruments of production and productivity the only legitimate goal. Ensuring that the
organization was productive required that management provide clear direction and tight
control.

Scientific management. As early management theorists formed these ideas, classical
theory received an important boost from a most unlikely source: a mechanical engineer
named Frederick Taylor who sought to extend engineering principles to the design and
management of work (Taylor, 1911). His approach was to determine through precise
measurement (“time and motion studies”) the most efficient way to carry out work
operations, to standardize them through clearly specified procedures, and to enforce them
through the exercise of explicit management functions, i.e., managers were there to issue
orders and workers were to follow those orders. Scientific management, as it came to be
called, provided concrete techniques for implementing the classical philosophy. Needless
to say, managers came to love it more than did the workers, many of whom saw scientific
management as just another attempt by management to exploit them, despised it (Gies,
1991).

Frederick Taylor

Value judgments aside, Taylor’s approach was anything but scientific except in the
narrow sense of measuring things and using data in place of opinion. True science, as
discussed in the next chapter, is much more than that: its aim is to understand natural
events. “Scientific management” was concerned with putting into practice a specific
theory of management that its proponents accepted as valid. Like classical theory, it had
nothing whatsoever to do with a search for knowledge, and in fact represented the
antithesis of scientific inquiry.

Taylor’s scientific management was founded upon four premises (Dessler, 1980):

1. Finding the “one best way” to perform the job. This typically meant simplifying the job
so that employees could more easily and quickly accomplish their tasks. Time and motion
studies were crucial to determining the most efficient way of performing the tasks in the
job.
2. Systematic personnel selection and placement to match the best worker to each job.
Well-developed selection and training programs that consider workers’ strengths and
limitations were emphasized.
3. Strict division of labor between management and workers. Managers were supposed to
manage and plan while the workers accomplished the work; boundaries between labor
and management were rigidly guarded.
4. Monetary incentives to attract and motivate workers to perform optimally. Systematic
selection and simplification of jobs served as only a baseline for efficiency. The profit
motive provided the balance.

Unlike the typical, despotic management practices of the nineteenth century, Taylor’s
ideas were grounded in an ethic of worker-management cooperation and increased
worker benefits. Even if his motives were irreproachable, his assumptions about human
motivation and attitudes were not. Taylor assumed that if workers were provided with a
decent job and wages linked to productivity, they would strive for peak performance. As
the researchers later discovered, however, this simplistic notion was grossly inaccurate.
Management could not increase productivity merely by lining workers’ pockets with
money. Taylor’s work inspired numerous scientific management practitioners. Among
the most notable of these were Frank and Lillian Gilbreath who pioneered the use of time
and motion analysis and the use of motion pictures (for a fascinating silent movie
showing the Gilbreaths and their early time and motion studies,
https://en.wikipedia.org/wiki/Time_and_motion_study). The book Cheaper by the Dozen
written by their son, Frank Gilbreath Jr., is a hilarious account of daily life with Frank
and Lillian and their 12 children.

A video showing the early applications of scientific management to the assembly line is
at link below. Go to this link now and view the video:

https://en.wikipedia.org/wiki/Time_and_motion_study

Lillean and Frank Gilbreath: “Thereis no waste in the world that equals the waste from needless, ill-directed and
ineffective motions” (Frank Gilbreath)

Weber’s ideal bureaucracy. Max Weber (1946) was a noted German sociologist whose
ideas, though consistent with the other classical theories and published about the same
time, developed independently. Because they were published in German and represented
a more academic than practical orientation, Weber’s notions had little impact on the
management literature until they were translated into English after World War II.
Undoubtedly the strained relations between the U.S. and Germany through two world
wars had something to do with this delay, a rather vivid illustration of how societal events
have influenced both the theory and practice of management.

The essence of Weber’s position is that the most defensible way to design and manage an
organization is through a rational-legal authority system. The key elements of a
“bureaucratic ideal” include
1. Division of labor,
2. Clear specification of roles,
3. A hierarchical (military-like) authority structure,
4. Explicit rules and regulations, and
5. Assignment of people to positions based on merit rather than social position, charisma,
nepotism, or political clout (the chief ways in which people landed good jobs at the time).

Max Weber in 1894
Though very similar to the prescriptions set forth by Fayol and Taylor, Weber’s views
were based on scholarly philosophical argument and historical analysis rather than sheer

practical experience (Weber, 1946). Thus, when social science became embroiled in
issues of organization and management, Weber’s arguments gave scholarly credibility to
the classical philosophy.

Weber saw in the bureaucratic ideal advantages that alternative organizing schemes
lacked, such as fairness and efficiency. It is ironic that these are the very criteria for
which modern bureaucracies are most vigorously criticized. To most of us, the word
“bureaucracy” has come to mean red tape, waste, and mediocrity, the direct opposite of
the efficiency and fairness Weber had in mind. Thus, it is important to distinguish the
concept, as envisioned by Weber, from the way the concept has often been implemented.
Perrow (1986) reminds us that despite its imperfections, and the prediction by its
midcentury critics that its days were numbered (Bennis, 1966), bureaucracy has not only
survived, it has outlived all the competition. It stands today as the dominant form in both
capitalistic and socialistic industrialized societies.

The emergence of psychology as a separate discipline

As industrialization, the corporation, and classical theories of management were
emerging, Psychology was evolving as a separate intellectual discipline. For centuries,
philosophers and other scholars have argued over the nature of mind and how it
functions—the phenomena of mental life. But the idea of using scientific methods to
resolve such issues was what distinguished modern psychology. In 1879 Wilhelm Wundt,
a professor at the University of Leipzig in Germany, opened the first laboratory devoted
exclusively to the study of psychology. Experiments were carried out under carefully
controlled conditions to answer basic questions about the human mind. Because most of
the important first-generation psychologists studied under Wundt, the new discipline
began as a basic science concerned mainly with laws of conscious experience.

Wilhelm Wundt in his Leipzig lab with his students and colleagues.
At about the same time that Wundt was trying to describe the general laws that govern
mental life, a British scholar and cousin of Charles Darwin, Sir Francis Galton, was
studying how people come to differ from one another mentally. He devised the mental
test as a means of indexing individual differences in mental ability. Although Galton’s
attempts at measuring mental ability were unsuccessful and misguided, his work spawned
a tradition in psychology that developed largely in parallel with Wundt’s mental science

(Hothersall, 1984). It has often been called the “mental testing movement.”

Testing for individual differences and doing experiments to discover general mental
principles have remained fairly distinct approaches in psychology despite periodic efforts
to draw them together (Cronbach, 1957, 1975). Wundt’s mental science approach
dominated early psychology in the United States, but the practical possibilities of testing
are what captured industry’s interest. Hence, the individual differences approach became
dominant in early industrial psychology and remained so for the next several decades.

The early years of psychology were characterized by debates among grand, theoretical
perspectives. Early on the emphasis was on describing the general laws that govern
conscious experience. Using introspection as a research tool, psychologists such as
Wundt attempted to analyze and describe the contents of conscious experience (e.g.,
sensations, images, feelings) similar to the way a chemist might analyze the elements of
matter. This paradigm was often referred to as structuralism because of its focus on the
structure of mental content.

During this period, Sigmund Freud’s theories on the nature and treatment of mental
disorders (and the role of unconscious events in mind and behavior) were also beginning
to attract a great deal of attention, eventually helping to establish the health-care branch
of psychology (Hothersall, 1984).

Sigmund Freud

Freud’s work on the unconscious, shifted attention to the more dynamic functional
properties of mind. From the late 1920s until the 1940s, therefore, the emphasis was on
the mental and biological underpinnings of functions such as motivation, emotion,
learning, and perception. The primary concern of functionalists was how humans and
other living organisms adjust to their environment.

Sigmund Freud
This school of thought constituted America’s first unique paradigm and became known as
American functionalism (Hothersall, 1984). It was a much less restrictive school of

thought than structuralism and much more appreciative of applications of psychology to
practical programs. Functionalism in psychology and pragmatism in philosophy provided
a context in which individual differences approach and applied psychology to develop
within it. William James was a Harvard professor who was perhaps America’s most
influential psychologist in the early years. He was among the founders of American
functionalism and the leading proponent of philosophical pragmatism. He also recruited
to the faculty at Harvard a German psychologist by the name of Hugo Munsterberg who
is known as the father of I/O psychology.

William James promotes pragmatism:
”Truth in our ideas means their power to work” (From Pramatism, 1907)

Behaviorism was an offshoot of functionalism and developed as a distinct area of
psychology under the leadership first of John B. Watson, and later, B.F. Skinner
(Hothersall, 1984). Behaviorism held that both the content and functions of mind are
unsuitable subjects for scientific study because neither is open to public view. Because
science only advances through objective observation, psychologists should limit their
research to behaviors (responses) that organisms exhibit and the environmental
conditions (stimuli) that control them. In its most radical form, behaviorism denied the
very existence of mind and maintained that all behavior is reducible to simple stimulus-
response (S-R) laws.

John Watson: “Give me a dozen healthy infants, well-formed, and my own specified world to bring them up in and I’ll
guarantee to take any one at random and train him to become any type of specialist I might select – doctor, lawyer,
artist, merchant-chief and, yes, even beggar-man and thief, regardless of his talents, penchants, tendencies, abilities,
vocations, and race of his ancestors” (Watson, J. B., 1930. Behaviorism (Revised edition). Chicago: University of
Chicago Press, p. 89).
The early applications of psychology to work

In these early years, psychology was largely confined to the realm of academic research
and application was held in disdain. Many psychologists felt that attempt to develop a
field of applied psychology would harm the development of psychology as a science.
Nevertheless, the pressures to use the findings and methods of this new discipline to
assist in the management of work proved irresistible. The potential uses of psychology
were obvious, and in the 1800s and psychologists rose to the challenge by applying the
methods, theories, and body of knowledge to practical problems of industry. The work on
individual differences, behaviorism, and functional psychology nicely complemented the
principles of Scientific Management. Job analysis techniques were developed to help in
studying the work and identifying the best ways of doing the job. Finding the right person
to do the work was facilitated as psychologists generated a variety of tests used to assess
and select among workers. Training techniques were developed to make sure that
workers learned the best ways of doing the work. The schedules of reinforcement
discussed by Watson and B. F. Skinner were compatible with what Taylor said about
paying people by the piece of work produced. Above all, scientific management proposed
that research should guide how people are managed, and psychology provided the
research tools for evaluating the effectiveness of alternative approaches.

The early years in which psychology was applied to problems of management tended to
reflect the emphasis dominant at the time on productivity and efficiency. Anything one
could do to improve human reliability, interchangeability of employees, and increase
work output was believed to have a profound impact on profitability. The principles of
work standardization and simplification coupled with tight management control made
sense in the societal and historical context of this period. And the prospect of using
paper-and-pencil testing to weed out the most incompetent or unreliable workers prior to
hiring them was especially attractive. Classical organization theory, scientific
management, and applied psychology provided tools for lowering unit production costs

just as surely as new machinery did. Applied psychology also provided tools for
identifying how to best persuade consumers to buy one’s product or service.
Interestingly, John Watson not only established the behaviorism school of thought in
psychology but later in his career also pioneered applications of psychology to marketing
and advertising.

The “fathers” of I/O psychology. The applications of psychology to industrial problems
began in earnest in the early 1900s with the work of two students of Wundt, Hugo
Munsterberg and Walter Dill Scott. Both wrote important books on industrial applications
of psychology. Munsterberg (1913) was concerned primarily with industrial efficiency
and Scott (1908) with advertising. Between them they touched on many of the topics that
have occupied I/O psychologists ever since.

Hugo Munsterberg writes the first I/O textbook in 1913: “For the sciences of the mind the time has come when theory
and practice must support each other. An exceedingly large mass of facts has been gathered, the methods have become
refined and differentiated, and however much may still be under discussion, the ground common to all is ample enough
to build upon.”

Walter Dill Scott publishes Increasing Human
Efficiency in Business- A contribution to the psychology of business (1911).

The first standardized mental ability test. Alfred Binet was a French psychologist who

was another pioneer in applying psychology to practical problems
(http://en.wikipedia.org/wiki/Alfred_Binet). He is best known for introducing the basic
procedures associated with intelligence testing and gave us the concept of IQ
(intelligence quotient). Binet, who is shown testing a student, devised an objective
method of assessing the intelligence of elementary school students so that teachers could
identify low intelligence students and provide them with special education. Previously,
this decision was based on the subjective and often biased judgments of teachers. For
instance, a teacher might have decided that some students from a lower socioeconomic
class or a minority ethnic group was incapable of learning and would assign them to
special education. Binet constructed an objective, standardized test that he believed
eliminated these biases. His test served as the prototype for all future mental testing.

Alfred Binet with student
World War I provided the setting for the first large scale application of psychology. The
United States and the other countries on the both sides were faced with the daunting task
of processing millions of people to serve as soldiers. U. S. President Woodrow Wilson
called upon psychologists to help in this endeavor and in May 1917 his administration
formed the Psychological Examination of Recruits Committee consisting of the top
people in psychological research on individual differences at the time (Dahlstrom, 1985).

Within two months they had constructed the Army Alpha test, which was a written paper-
and-pencil test, intended for recruits who could read and write. Another test, the Army
Beta, was a nonverbal test for illiterate and non-English speaking recruits. By the end of
WWI over two million army recruits had taken the Army Alpha and Beta (McGuire,
1994). Although the Army Alpha and Beta were not subjected to the research scrutiny
that characterizes modern testing, these were the first practical tests that educators could
administer to large groups of people. From this point on, intelligence testing became a
focus of attention of popular culture. An unfortunate consequence of intelligence testing
was that it provided the data on racial differences in intelligence that the eugenics
movement used to encourage discrimination against ethnic minorities and immigrants
(Gould, 1981). This led to controversy over whether intelligence testing is racially biased
that continues to this day (Fancher, 1985).

http://en.wikipedia.org/wiki/Alfred_Binet

Psychometric theory. At the same time that psychologists were developing measures of
intelligence that management could use in making practical hiring decisions, other
psychologists and statisticians were working on the theoretical foundations for these
applications. The conceptual framework that emerged during this early period was
Classical Test Theory (CTT). CTT would dominate thinking about the use of individual
difference measures for decades to come and continues as an influential framework in I/O
psychology. CTT presumes that for every psychological construct there is a true score.
Psychological measures reflect not only the true score but also some error. These
measures are fallible to varying degrees and psychometric theory provides the conceptual
foundation for practical means of assessing the fallibility of psychological measures.
Reliability and validity, the two primary bases for evaluating psychological measures,
were the products of classical test theory.

Psychometric theory was made possible by the emergence of the discipline of statistics.
The historical roots of the field of statistics are traceable to as far back as the 1600s, but it
was during the 1800s and the early 1900s that the discipline emerged as the quantitative
foundation for a variety of social sciences including sociology and psychology. Earlier
there was a discussion of Gauss and his notion of the normal distribution. He also
introduced such concepts as standard deviation, correlation, and regression, statistical
concepts and procedures that are discussed throughout this text. In the late 1800s and
early 1900s, Karl Pearson, Ronald Fisher, and Charles Spearman pioneered the statistical
procedures that continue to dominate, for better or worse, how I/O psychology uses
statistics. The author of this text says “for better or worse” to acknowledge the serious
doubts raised about how statistics are used to determine whether relationships among
variables deserve attention (Krueger, 2001).

*Left to right, front row, Edgar Doll, Henry Goddard, and Thomas Haines. Left to right,
rear row: Frederic Wells, Guy Whipple, Robert Yerkes, Walter Bingham, and Lewis
Terman.

The Middle Years (1920 – 1945)

The classical theories of management and scientific management, along with the practical
tools developed by psychologists, succeeded in helping managers control labor costs.
Moreover, the emphasis on hierarchical structure and strict obedience to the boss were
effective in managing the masses of workers and large scale manufacturing so dominant
during the early years. Labor unions were emerging during the early years but they were
weak and seemed manageable. Where necessary, corporate tycoons could hire private
armies to prevent unions from forming, and if that failed, management could bring union
members into line by threating to dismiss them. If a strike still occurred, management
could simply replace the striking workers. Were some do-gooding politician or
newspaper editor to take issue with these methods, as some did, corporate leaders could
handle that as well by the proper application of money or pressure in the right places—
notably to the politician’s sources of funding or the newspaper’s ownership. In short, there
was little apparent reason to worry about societal change, and the only real issue in
managing people was ensuring that their humanness interfered as little as possible with
their output.

In the years between the end of World War I and the end of the World War II, there were
huge changes that forced managers to reconsider how they dealt with people. The
included the “roaring twenties” and the economic crash of the great depression years.
During this period there were dramatic increases in the educational levels of American
workers, the rise of populism, socialism, and unionism and World War II. I/O psychology
continued to grow during these years building on the successes of the early years but
rather than focusing solely on the traditional concerns of I/O during the early years,
another way of thinking about the management of workers emerged. McGregor (1960)

would call this new way of thinking “Theory Y” and contrasted it with Theory X, the
way of thinking reflected in the Classical and Scientific Management approaches.

Economic boom and bust

Few periods in American history have influenced government structure, size, and
economy more than the “Roaring Twenties” and the “Great Depression”. The optimism
and greed of the immediate post WWI years bear a striking resemblance to the recent
economic booms in the 1990s and early 2000s as does the economic collapse that
occurred in the 1930s. The events of this period also shaped thinking about management
of people and the field of I/O psychology.

The roaring twenties. Cities in the 1920s were profoundly impacted by the preceding war.
European immigration had almost come to a standstill during the fighting. New
immigration restrictions in 1921 and 1924 left cities with insufficient European
immigrants to man factories. Southern blacks, looking for a better life, migrated to urban
areas and exchanged sharecropping for factory jobs. Widespread economic prosperity
ensued as cities eclipsed small towns and rural areas in overall population.
Mass-produced technology products were within the reach of the majority of Americans,
including returning WWI soldiers. Every home was a market for electrical appliances of
all kinds—refrigerators—washing machines—vacuum cleaners, etc. Advertising
convinced Americans they needed products, some of which they had never heard of
before. Electricity production soared throughout the 1920s’ decade as more of the United
States joined the electric grid, and most coal powered industries switched to electricity.
Assembly lines produced Henry Ford’s automobiles at prices most Americans could
afford. Mobile Americans enjoyed freedom as they took to thousands of miles of new,
paved roads. Charles Lindbergh traveled the highway in the sky in 1927, becoming the
first person to fly solo across the Atlantic.

Actions in the 1920s led to the depression in the 1930s, which led to the New Deal in
1933 through 1938. The economic causes of the Great Depression originated in the
events surrounding World War I and its aftermath. World War I exhausted people. Grief,
loss, and hard work made the people ready for a change. The 1920s brought this in a
number of ways. Consumerism reached its height in this time period. The “flapper”
epitomized consumerism with her makeup, ready made clothes, and cigarettes.
Entertainment soared and morals began to slide. All of these factors covered the inflation
that was happening in prices. Worker’s salaries increased but did not match the rate of the
price increases and farmers were forced further into debt. National economics also
suffered with income taxes being the lowest for those who were the wealthiest. In 1924,
immigration became limited, reducing the number of workers in the economy. When the
stock market crashed in 1929 all the economy’s weaknesses became clear. This was not,
however, the start of the Great Depression but just a signal of its onset.

Also rising during the 1920s and 30s was the educational level of the worker. As noted in
the graph below, there was a steady increase in the proportion of 5 – 19-year-old
population enrolled in school from 1850 to the 1990s. Especially dramatic was the steep

rise in education that occurred in the 1920s and 1930s. Increased education raises the
expectations of people for the type of jobs they will hold and how management will treat
them. An educated workforce is also less submissive to autocratic management and more
likely to question such practices.

The Great Depression. The other event that brought attention to the need to take into
account human rights and attitudes was the Great Depression. The dramatic worldwide
economic collapse of the 1930s most drastically affected national moral by eliminating
jobs. In 1929 only three percent of the workforce was unemployed, by 1933 this had
reached twenty-five percent. Since a quarter of the nation was out of work, people
stopped spending money. All durable goods such as automobiles, housing and appliances,
didn’t sell and forced people who worked in those industries out of their jobs. People
with jobs had to accept pay cuts and many felt lucky just to have a job. Banks lent out
more money than they had and many people lost their savings. Family life also changed
drastically, with marriage and birth rates falling rapidly while divorce rates grew. All of
this kept the economy in a shambles with no hope of relief. Then came the New Deal.

After taking office in 1933, President Franklin Delano Roosevelt sought to bring change
to the nation by increasing the role and power of the government in ways that the
previous President, Herbert Hoover, hadn’t. Hoover had been reluctant to use welfare
programs to help those in need and did not establish many public works projects.
Roosevelt, however, used both of these programs generously. At this time the US also
abandoned the gold standard and put more money into circulation, meaning there was
more cash circulating than there was gold bullion. Trust in banks was one of the first
things Roosevelt sought to restore. This happened with the “bank holiday” and the
Emergency Banking Act. The Federal Deposit Insurance Corporation also helped to
restore trust by ensuring the safety of people’s money in case the bank failed. The Public
Works Administration appeared at this time as well. Dams, bridges and other public
buildings were built, creating more jobs for unemployed people. Work camps were also
set up to help people find jobs. The New Deal was influential in that it helped alleviate
the misery that overtook the people but it also expanded the role of government in the
United States, Great Britain, and other nations. Assistance provided by charitable
organizations such as churches to help the unemployed, homeless, and displaced now
became government programs.

Figure 2.2: 5 to 19-year-olds in the U. S.
Enrolled in School, 1850 to 1991.

The financial crash of 1929 followed by the depression of the 1930s shattered both of
these myths together with the national spirit of rugged individualism and self-confidence.
The suspicion arose and spread that perhaps unfettered capitalism was not the answer to

all organizational and social problems. Clearly there were forces at work in society
powerful enough to bring down even the biggest and strongest. U. S. society began to
recognize that neither individuals nor organizations were the complete masters of their
fate. The Great Depression of the 1930s was a worldwide economic phenomenon that led
to a host of societal changes. For example, it undoubtedly contributed to conditions in
Germany that culminated in the next epoch-making event, World War II. Further, it
reinforced the growing recognition that for all its success, the U. S. was not immune from
foreign influence. This, unfortunately, has remained a difficult lesson for Americans to
learn in the twenty-first century.

For present purposes, it is sufficient to mention some direct consequences of the
depression for corporate America in the 1930s. First, government influence became much
stronger, becoming a force that corporate leaders had to reckon with in business decisions
affecting union relations, personnel management, and consumers. President Franklin D.
Roosevelt’s sweeping reform programs (the National Recovery Act, his reshaping of the
Supreme Court, the New Deal, etc.) put serious restrictions on the free enterprise system
and created a favorable climate for organized labor. During the 1930s union membership
grew to nearly 30% of the nonagricultural workforce (Kochan & Barocci, 1985). No
longer could management dominate labor-management relations as it had in the past.

Second, public support for these government programs signaled a dramatic shift in
society’s attitudes. For the first time the public could identify with the plight of the
downtrodden worker. Because misfortune had touched everyone, it was no longer limited
to what society perceived as the ignorant, the lazy, and the largely “foreign element” that
dominated the working class (who in society’s view probably deserved mistreatment).
Corporate greed and abuses became an alternative explanation, and sentiment shifted
from supreme confidence in big business to distrust and a plea for government
intervention.

Finally, organizations began to feel pressure from a variety of sources to shoulder more
responsibility for employee welfare. Massive unemployment did nothing to hurt their
favorable labor market, i.e., labor became even cheaper and the desperate economic times
made people even more willing to submit to authority. Nevertheless, there were mounting
pressures from government and changing social norms on the business community to
address the rights and welfare of their employees. In addition to social and government
pressures, the research and theory of anthropologists, sociologists, and psychologists led
to fundamental changes in the way managers thought about work.

The rise of unionism, populism, and socialism. Other sources of the increased concern for
worker rights and welfare were the three “isms”…. populism, unionism, and socialism.
Populism can come in the form of both right-wing and left-wing movements and reflects
a rejection of elitism (e.g., the “Hollywood elite” by the right and the “1%” and
“corporate elite” by the left) and an assertion of the rights of the ordinary working person
(http://gildedage.lib.niu.edu/populism). Populism had begun prior to the World War I in
the form of the progressive movement and culminated in among other things the winning
of the vote for women in the United States in 1920. With the greed and corruption of the

http://gildedage.lib.niu.edu/populism

gilded age during the late 1800s social critics brought attention to the abuses of large
corporations and were instrumental in mustering public opinion in support of greater
government oversight of corporations. Theodore Roosevelt was the first U. S. president
to act as a champion for the Progressive cause. In the Square Deal, the name of his
domestic plan, he advocated Federal regulations and laws to prevent the unfair business
practices of the large monopolies.

Also responsible for an increased concern with worker rights and welfare was the
growing strength of organized labor. Early examples of labor-management clashes in the
U. S. were The Great Railroad Strike of 1877 organized by the Knights of Labor, the
1892 strike at Carnegie steel works in Homestead, Pennsylvania led by the Amalgamated
Association of Iron, Steel and Tin Workers, and the strikes by Pullman Palace Car
Company employees organized by the American Railway Union. Socialism and powerful
unions such as the Industrial Workers of the World (IWW) or “Wobblies”
(https://en.wikipedia.org/wiki/Industrial_Workers_of_the_World) raised the fears of
corporate leaders that unions would force them to negotiate wages, work conditions,
working hours, safety, and other issues with their employees. Not only was this seen as a
threat to their profits but more importantly as a threat to their power.

Even more threatening to corporate leaders was the rise of socialism, which was a system
of thought advocating social ownership of property. There are many gradations in this
philosophy, ranging from the extreme in the form of communism where there is total
ownership of property by the state to the milder forms of socialism that led to progressive
income tax, social security, and universal education. The fears of corporate leaders
around the world were fed by the communist overthrow of the Russian czar and the
imposition of a Marxist state. Post WWI labor party reforms in Great Britain further fed
their fears. These reforms included the nationalism of a large proportion of industry in
Great Britain, a National Health Service, and public housing. Socialism as bloody

https://en.wikipedia.org/wiki/Industrial_Workers_of_the_World

totalitarian overthrows and as milder democratically guided economic reforms seemed on
the march in the 1920s and 1930s and appeared to threaten an eventual end to capitalism.
((http://en.wikipedia.org/wiki/Socialism). Confronted with unionism, populism, and
socialism, corporate leaders began to give serious attention to providing better working
conditions for employees and reconsidered the iron-fisted, autocratic form of supervision
that had dominated management practices during the 1800s and early 1900s.

World War II leads to widespread applications of I/O psychology. Ultimately, it was not
the New Deal that ended the Depression but World War II. In 1940 the unemployment
rate was still at fifteen percent. World War II produced many more jobs and finally got
the U. S. and other nations back to normal. The evolution of I/O psychology from a
minor offshoot to a recognized specialty within the field of psychology was a
consequence of several influences, but none more important than WWII. The demands of
this massive war effort, representing dramatic changes in the doctrine and technology of
warfare, called for radical changes in the management of human resources. Psychologists
of all kinds were summoned to help, and in the course of doing so they discovered a great
deal about the potential applications of their specialized knowledge and techniques. All
branches of psychology made huge advances under the exigencies of war—an ironic
twist for a discipline pledged to the promotion of human welfare.

December 7, 1941 attack on Pearl Harbor, Hawaii
WWII provided massive evidence of the value of testing for selection and assignment of
people according to job requirements. Naturally, industry recognized the potential
applications of these techniques, and the use of testing increased dramatically after both
wars. In addition, World War II produced important advances in techniques used to tram
people, many of which found peacetime application. Educational, experimental, and
industrial psychologists all contributed significantly to the design of these training
programs. One notable example was simulation training, a technique in which trainees
could practice the basic operations required in a job (e.g., flying an airplane) on a replica
of that job’s environment (i.e., a simulator) without the risks and costs associated with
learning in the actual situation. Today, the simulation approach has spread to the training
of managers, professionals, and skilled technicians, and it remains a cornerstone of pilot
training, both military and commercial.

http://en.wikipedia.org/wiki/Socialism

In addition to selection, classification, and training developments, World War II fostered
the creation of a whole new field based upon an entirely different strategy for improving
human effectiveness. The essential idea behind it was that “human error” or poor
performance is not always a matter of incapable or poorly trained personnel, but often
reflects insufficient consideration of human characteristics in the design of the machines
that people must operate. The field survived the war, and today constitutes a formally
recognized discipline known in the United States as human factors and in the rest of the
world as ergonomics. It encompasses a variety of human-oriented and design-oriented
specialties, the psychological component of which is often called engineering or human
factors psychology (Howell, 1991).

WWII also further reinforced the growing concern for worker rights and welfare. To
succeed, it was vital that the vast defense industry keep running smoothly without
interruptions. The manufacture of airplanes, guns, ammunition, uniforms, and all the
other stuff of war could not afford strikes, walkouts, slowdowns, and other forms of
resistance so often associated with autocratic management. To maintain harmony and
avoid costly labor-management conflicts, the federal government imposed rules
and procedures and created entire bureaucracies dedicated to implementing these rules
and procedures. It was during the war years that many of the current wage and salary
administration practices were created to ensure fairness in compensation of employees. It
was also during the war years that labor arbitration and mediation began to avoid and
resolve labor-management disputes. Human resource management dramatically changed
during WWII from a primarily record keeping function that was focused mostly on
reducing costs and increasing productivity, to a function that had a broader concern with
human relations. Not surprisingly, psychologists had a lot to say about how this new
human resource management function should operate.

The Human Relations movement and neo-classical theories of management
What were social scientists and in particular, psychologists, doing during the post WWI
and WWII years? One important development during this period was the conduct of
empirical research to test alternative approaches to management. This was in contrast to
the prescriptions of classical theory and scientific management, which seldom used
rigorous scientific methods. One program of research, the Hawthorne studies, stretched
over two decades and fueled a rethinking of the Classical and Scientific Management
theories that dominated the early period. Two streams of thinking emerged in the middle
years: the Human Relations movement and the Neo-Classical theories. Both schools of
thought urged more attention to the needs of employees and questioned the assumption
that employees will necessarily obey and submit if the principles of Classical Theory and
Scientific Management were implemented. While raising questions, the Human Relations
and Neo-Classical theorists accepted for the most part the organization as it was and the
need for management to make the decisions. They were not avid proponents of allowing
employees to participate in important decisions or decentralizing to empower those at the
lower levels of the hierarchy. Rather, both approaches advocated ways of reducing
resistance to managerial orders and taking a more people oriented approach was seen as
key to achieving this objective.

The Hawthorne studies. The true scientific approach made a tentative entry into the field
as early as the 1920s in the “Hawthorne Studies” conducted at the Chicago Hawthorne
plant of Western Electric (Roethlisberger & Dickson, 1934). These studies began in 1924
and continued through the early 1930s. A variety of researchers participated including
Clair Turner, Fritz J. Roethlisberger, W. J. Dickson, and Elton Mayo.

The illumination studies began in 1924 as a typical industrial engineering question about
the relationship between worker productivity and the level of illumination in the work
environment. The researchers were guided by the prevailing organizational philosophy of
that era which assumed that workers are motivated solely by tangible factors, such as
money and working conditions (illumination).

Elton Mayo: “…..problems of absenteeism, labor, turnover, ‘wildcat’ strikes, show that we do not know how to ensure
spontaneity of cooperation; that is teamwork. Therefore, collaboration in an industrial society cannot be left to
chance….”

Fritz Roethlisberger: “…the behavior of no one person in an industrial organization, from the very top to the very
bottom, can be regarded as motivated by strictly economic or logical considerations…noneconomic motives, interests,
and processes, as well as economic, are fundamental in behavior in business.”

In the initial experiment, workers in three departments were exposed to varying lighting
levels. To their surprise, the researchers discovered that productivity did not always
decline when they lowered illumination. Intrigued, the researchers conducted two more
illumination studies, each with tighter experimental controls than the first. Even with the
systematic manipulations of lighting levels, worker productivity did not decline relative
to the amount of illumination present. In one group, productivity continued to increase as
illumination decreased, even to the point that the workers complained that they were
hardly able to see what they were doing.”

These results provoked the researchers to implement the second set of experiments, the
relay assembly test room studies. In these studies, the goal was to isolate small groups of
workers to control for factors that were believed to have potentially biased the
illumination studies, such as personnel changes and departmental policies. The job the
researchers chose to study was the assembly of telephone relays. Putting together the
assembly fixture was highly repetitive but required considerable motor skill. During the
course of these experiments, the researchers tested the effects of many different types of
work-rest cycles. For example, manipulations included 5-minute breaks in the morning
and afternoon, 10-minute breaks in the morning and afternoon, a series of six 5-minute
breaks, and a 15-minute break in the morning and a 10-minute break in the afternoon.
Regardless of the type of rest break introduced, productivity increased over the baseline
levels found prior to the interventions. In the next phase of the relay assembly test room
studies, the researchers investigated how shortening the workday influenced worker
productivity. The manipulations included shortening the workday by 30 minutes (with
rest pauses) and the workweek by one-half day (with rest pauses). In both cases
productivity did not decrease. Again, productivity rose regardless of the changes in the
length of the workday.

The tendency of workers to increase their productivity regardless of the changes made in
the rest breaks and length of the work day suggested to the researchers that social
processes had influenced work output. To test this possibility they launched the bank
wiring room studies. Based on observations of fourteen men engaged in the assembly of
wiring banks, the researchers found that the workers had developed a concept of a
“proper day’s work” and pressured any peer who did not comply with the group’s
standard. This social pressure was particularly evident for those workers who attempted
to exceed the group’s output quota, i.e., the “rate busters”. These findings were quite
surprising to the researchers and management because the organization’s incentive system
was based on higher pay for higher output. The rational course of action was produce
more to make more money but the workers apparently went against their own best
interests to win the approval and avoid the criticisms of their co-workers. These and other
observations led them to conclude that social factors indeed were a powerful influence on
individual worker productivity.

The research produced a new perspective on work organizations (Dessler, 1980). The
results of these studies made behavioral scientists aware, for the first time that monetary
incentives and the physical work environment are not the only factors that affect worker
behavior. Indeed, the results suggested that peer pressure, informal group dynamics and
recognition, and personal feelings of freedom and self-worth are more important than
impersonal structural features, such as work methods and management control. Although
these studies were flawed, they provide a fascinating example of how psychologists can
apply the scientific method to organizational questions. The cumulative findings of the
entire series of studies conducted at the Hawthorne plant led to several radical insights
that departed from classical and scientific management (Roethlisberger & Dickson,
1939):

1. Employee attitudes and morale are crucial to employee productivity. A satisfied
worker is a productive worker.
2. Employees work more for acceptance and social recognition than for money.

3. The informal work group was a major determinant of productivity.
4. It is employee’s perceptions of the situation that are important, not objective reality.
For instance, some employees after being asked what they thought of the food in the
cafeteria perceived improvements when nothing had been done.
5. The technical and human components of the organization are interrelated and
interdependent. Management must take both into consideration when implementing
changes.
6. In dealing with worker resistance, management does not need to relinquish control or
change the organization. Instead, good human relations practices are used to reduce
resistance and gain acceptance of the workers.

The early Human relations movement. Mayo, Roethlisberger and the other researchers in
the Hawthorne studies were known as the founders of the Human Relations Movement.
According to the human relations school, the manager should possess skills for
diagnosing the causes of human behavior at work, interpersonal communication, and
motivating and leading workers. The focus became satisfying worker needs. If worker
needs were satisfied, wisdom held, their productivity would rise. Thus, the human
relations school focuses on issues of communication, leadership, motivation, and group
behavior. The human relations school of thought still influences management theory and
practice, as contemporary management focuses much attention on human resource
management, organizational behavior, and applied psychology in the workplace.

Some basic ideas of the human relations movement (also called Mayoism) were:
1. Supervisors should act as friends and counselors to the workers. They should use
humor to engage workers in friendly interactions. By satisfying employee needs for
acceptance and security, such treatment will motivate them to cooperate with
management.

2. The primary concerns of managers are the people, not productivity. Showing a people
concern will achieve higher levels of productivity in the end than an overriding concern
for productivity. The Mayo formula is: humanistic supervision plus morale equals
productivity.

3. Managers should ask workers how they feel about their work — and their supervisors.
They can survey worker attitudes through informal conversations as well as in formal
interviews and surveys.

4. Management should consult workers before changing things in the workplace.
Allowing employees to participate will reduce resistance to management.

5. Managers must accept and encourage group dynamics in the workplace. Groups are
not an impediment to productivity but are vehicles for achieving high levels of
productivity.

The Human Relations movement especially in the form of “Mayoism” did not escape
criticism. Questions were raised about the soundness of both the science and the thinking

underlying this approach (Bruce & Nyland, 2011). To the extent that the movement relied
on the findings of the Hawthorne studies, critics have noted that the researchers gave a
one-sided interpretation to their data and an overemphasis on worker-management
harmony. Critics also have targeted the flaws in the design and execution of the
experiments. The investigators in these studies deviated in several respects from good
research design. Perhaps the most glaring error was that there was never an attempt to
randomly assign workers to experimental and control groups. In fact, workers were often
chosen because they were congenial and productive. In the course of the experiments,
these already superior workers were given preferential treatment by both the researchers
and management. At no time was the output of workers in the experiments compared
with the output of other, comparable workers in the organization. Also, the number of
workers actually studied was so small that sample size precluded any generalizations to
the larger organization.

Other criticisms of the Human Relations movement included:

1. It was a “cow psychology” in which workers were treated as contented cows to
increase their milk production.
2. It did not acknowledge that there were genuine conflicts of interest between
management and employees within organizations. These conflicts required serious
negotiation and were not simply smoothed over through good human relations.
3. It was manipulative. The human relations approach often accepted management
decisions as givens and the objective of human relations was to reduce resistance to these
decisions.
4. It was based on assumptions about people that were degrading. Workers were viewed
as irrational children who management could coax into high levels of productivity
through human relations.

Bruce and Nyland (2011) describe Mayo and the human relations movement as treating
“workers as irrational, agitation-prone masses susceptible to ‘socialistic’ radicalism and,
so, unfit for ‘voice’ in the workplace. For Mayo, the industrial unrest of his time was
caused by ‘factors of unreason’ or ‘factors of irrationality’ and this irrationality among
workers was the product of ‘disassociated’ or ‘obsessional reveries’. Applying the new
medical psychology, he likened workers to shell-shocked soldiers in need of serious
psychological/psychiatric attention. Crucially, Mayo dismisses workers’ calls for better
wages and conditions and a ‘voice’ over same as ‘socialistic radicalism’ and as
symptomatic of some deeper psychosocial maladjustment” (p. 289).

In spite of these well-warranted criticisms of the Hawthorne studies and the Human
Relations approach both had a profound and enduring influence on the study of worker
behavior. Their significance lays not so much in the specific results or conclusions, but in
the ways they shaped future research. In fact, this chapter on job attitudes would probably
not exist—at least not in its present form—if the Hawthorne studies had not convinced
behavioral scientists of the pervasive effects of job attitudes on both organizational and
personal functioning.

The neoclassical and other notable organizational theorists.

Other theorists emerged during the middle years who questioned whether management
should accept the Classical and Scientific Management approaches hook, line, and sinker.
Similar to the Human Relations movement none of these theorists rejected outright the
ideas of centralized authority, specialization, and many of the other principles of the early
years. However, they all pointed to limits of these approaches and advocated a kinder,
gentler, and somewhat looser form of the Classical and Scientific Management
approaches. Some of these theorists were called neoclassical theorists. As indicated in the
“neo” they approached organizational theory from a somewhat new perspective but
basically kept many of the tenets of the classical approach.

1. The unanticipated consequences of bureaucracy. During the 1930s, 40s, and 50s
several social scientists, including Merton (1940), Selznick (1949), and Gouldner (1954),
identified unintended consequences of implementing the recommendations of the
classical and scientific management theorists. They based their comments on intensive
case studies of organizations. Merton (1940) observed that efforts to control behavior of
employees in an organization following the dictates of Classical and Scientific
Management generally succeed in increasing the reliability and predictability of worker
behavior and improving performance in meeting customer or client demands. In the long
run, however, the same factors that lead to improvements can have unanticipated
consequences that eventually work against meeting client needs. These unanticipated
consequences include rigidity, defensiveness, and concern with organizational status.
These unfortunate outcomes reflect the impersonal treatment of employees, the tendency
of employees to internalize the rules of the organization, and the stereotyping of problems
so that there is an unwillingness to consider alternative solutions to these problems.

Selznick (1949) observed that in the classical model of the organization there is
delegation and specialization in the design of the organization. For instance, Fayol and
others recommended as part of their principles that the CEO create functional units or
departments (e.g., marketing, production, human resource management, accounting, etc.)
and delegate to the heads of those departments the authority to make important decisions.
Of course there are general strategies and policies that management must follow but as
long as the middle and lower management ranks work within these guidelines, they make
the day-to-day decisions and thus alleviate the top person from making every single
decision. Specialized duties and functions are assigned to employees and training them in
the performance of these activities, and each set of employees have subgoals that they
pursue that help fulfill the overall organizational goals. All of this sounds fine, but an
unanticipated consequence is that the employees within these specialized units and roles
tend to lose sight of the overall organizational objectives and place increasing importance
on their own subgoals. Conflict among units ensues. Each unit attempts to justify and
rationalize actions in a way that distorts reality and detracts in the long run from effective
performance.

Gouldner (1954) focused on the unanticipated consequences of general and impersonal
rules. Specific rules as to what employees should, and should not do, improve the

reliability of performance but will also identify the minimum acceptable behavior. For
instance, if a salesperson is told that the rule is that three cold calls are required per day,
the employee may then focus on this as the minimum acceptable performance and only
do three cold calls and no more than three cold calls. This unanticipated consequence
will then provoke closer supervision of the employee that, in turn, increases the
interpersonal conflict and tension between supervision and employees. The conflict will
only further increase the closeness of the supervision and the cycle will continue with an
ultimate decay of performance.

2. Barnard’s cooperative systems view. Chester Barnard was an executive with extensive
experience in both public and private sector management. Yet unlike the early classicists,
his views were shaped as much by the emerging work of social scientists as by personal
experience. In fact, his principal contribution was the attempt to reconcile the
contradictory classical and human relations views through his concept of natural
cooperation. For Barnard (1938), the essence of organization is the existence of a shared
goal or purpose. An organization is effective to the extent that all employees regardless of
position are loyal to the goals of the organization. Unless members put the organizational
goals ahead of personal goals, the organization has little to offer society or its members
and is unlikely to survive. A basketball team exists primarily to win games; a business, to
make a profit; a government, to serve its public. If the members are preoccupied with
scoring points, amassing personal wealth, or achieving political power to the detriment of
the shared objective, no one wins. While this is consistent with the classical position, an
organization cannot survive if it fails to recognize the needs of its members— goals that
are often in direct conflict with the shared purpose. Moreover, formal authority is of little
value unless those to whom it is applied choose to accept it. Disgruntled workers can
restrict production, engage in sabotage, or simply quit. It is the organization’s
responsibility, therefore, to give individuals a fair return on their personal investment in
terms that matter to them. In so doing, it ensures their compliance. It is management’s
responsibility to “sell” subordinates on the collective purpose and to help them see their
stake in it. The college coach must convince her recruits that they must pass off more
often and concentrate on defense if they are to win at the college level even though her
players were all star point scorers in high school.

Barnard believed that organizations need rules, procedures, reporting relationships,
managerial authority, and all the other recommendations of the classical and scientific
management approaches. But these are not enough to ensure success unless the recipients
of those orders and directions are willing to cooperate. Bernard distinguished among
three types of orders.

*First there are those that lie within what he called the “zone of indifference” and are
accepted without question and are always obeyed. There are many orders involving
routine duties that are in an employee’s job description and are performed without
resistance.

*Second are those that may lie within the zone of indifference but come close to being
unacceptable. These orders that employees may or may not follow. Much depends on

characteristics of the employee and what the employee’s informal organization defines as
acceptable and unacceptable.

*Third are those orders that are outside the zone of indifference and are not acceptable to
employees. They will result in disobedience. The key point in Barnard’s theory is that all
facets of an organization (individual and social, formal and informal) are concerned with
achieving cooperation in pursuit of a common goal. In essence, both workers and
managers take their orders from the requirements of the situation. The good leader seeks
the middle-of-the-road positions where cooperation is assured or attempts to extend the
zone of indifference through persuasion and good human relations. They avoid those
issues that are far outside the zone of indifference.

Chester I. Barnard publishes the Functions of the Executive, 1938: “The responsibility of the executive is (1) to create
and maintain a sense of purpose and moral code for the organization; (2) to establish systems of formal and information
communication; and (3) to ensure the willingness of people to cooperate.”
3. Mary Parker Follett (1868-1933). She was a Harvard professor of social work who was
among the very few women in the early years who rose to prominence as a management
consultant. Consistent with Barnard, she advocated a focus on human relations in which
managers would use power as a principle of “with” rather than a principle of “over”. The
power of the manager does not accrue as the result of asserting rewards and punishments
of the legitimacy of the office held. Rather the power of the manager derives from how
the manager functions in the position. She is given credit for first identifying the
groupthink effect in which committees engage in conformity, brainstorming as a
creativity exercise, an early form of MBO (Management by Objectives), and the notion
of Total Quality Management/Continuous Quality Improvement (TQM/CQI). She was a
strong advocate of empowering employees by sharing power with them. Although
Follett did not have nearly the impact on U. S. management practices as
Barnard, Japanese managers discovered her writings in the 1950’s and give her credit for
helping them rethink their management styles.

Mary Parker Follett publishes the “Creative Experience” (1924): “The best leader knows how to
make his followers actually feel power themselves, not merely acknowledge his power.”
The Post WWII Era (1945 – 1965)

In the immediate post WWII years (1945 – 1965) the field of I/O psychology matured
into a distinct sub-discipline of psychology along with other more established sub-
disciplines such as clinical, social, and experimental psychology. During this period, I/O
psychologists increasingly identified themselves with either the human resource
management wing (the I of I/O) or the organizational development wing (the O of I/O).
Both were led by the scientific method and during this period there was an explosion of
research on leadership, group dynamics, personnel selection, performance appraisal, and
all the other topics covered in the text. Much of the work during the two decades
following WWII built on the contributions of psychologists during the 1920s, 30s, and
WWII. For instance, some psychologists were busy developing and evaluating
psychological tests and constructing increasingly sophisticated statistical tools. Other
psychologists were engaged in research testing notions about leadership and group
behavior that sprung from the human relations movement.

Rapidly accelerating change in a period of progress and turmoil

Beginning in the late 1960s and continuing through to the present day, more complex
views of organizations and how to manage people within organizations evolved. In
contrast to the closed systems assumptions of the classical/scientific management and the
human relations movement, organizations were viewed as open systems. To understand
and effectively manage people in organizations one had to take into account the cultural,
political, economic, and technological contexts of the organization. Moreover, managers
had to approach the organization as an interdependent set of systems consisting of people,
groups, organizational structure, and technology. Changes in one part of this system had
implications for other parts of the system. Rather than asserting one-best way solutions,
the research identified the conditions under which a classical/scientific management
approach or a human relations movement is effective. Rather than universalistic theories,
contingency theories of organizations and management tended to dominate thinking and
practice. These changes did not occur in a vacuum but were influenced by historical,
social, and economic changes. Three events seem particularly important in the two

decades following WWII: economic growth and the dominance of the U. S., the cold war,
and the space program.

Economic growth and the dominance of the U. S. The postwar period ushered in a
sustained economic boom that was to extend, with minor fluctuations, over most of the
next three decades. The birth rate, which had been depressed during the 1930s, rebounded
dramatically (the so-called baby boom). Labor unions continued to grow in membership
and influence and the standard of living in the U. S. became the envy of the world
(Kochan & Barocci, 1983). Buoyed by military success, yet conscious of the devastation
that the war had wrought on friend and foe alike, the old spirit of self-confidence
returned, together with a healthy sense of compassion.

Isolationism, the idea that America could do nicely without the rest of the world and
should not meddle in its affairs, was prevalent before the war but was overwhelmed by
subsequent events. Comprehensive economic aid plans for war-torn Asia and Europe
were implemented and supported enthusiastically by the public. Undoubtedly, forward-
looking business leaders also realized that such restoration represented an initial step
toward opening vast new markets for their wares, although this was of little immediate
consequence given the pent-up demand for goods and services at home. What even the
most enlightened failed to anticipate was that the restoration of Japan and Western
Europe would also bring serious economic competition.

The cold war. Among the most significant of all the consequences of the war was the
economic and political division of the world along capitalistic-communistic lines, the
emergence of the Soviet Union as a superpower, and the threat of nuclear war. For the
next 45 years the cold war isolated Eastern from Western societies and created on both
sides a sustained atmosphere of fear, political competition, and military escalation.
Military spending became a dominant force in the economies of both the U.S. and the
Soviet Union as President Dwight Eisenhower foresaw when he warned in his farewell
address against uncontrolled growth in the military-industrial complexes. Psychological
research was a beneficiary of this spending as the military services sought behavioral
means to improve the effectiveness of fighting forces.

Space exploration. Closely associated with the Cold War was the space exploration race
between the former Soviet Union and the United States. It the 1950s, the apparent winner
was the Soviet Union and the threat of Soviet military platforms orbiting the earth and on
the moon triggered a massive effort on a national level to upgrade education at all levels.
Fueled by fears that the Soviet Union would dominate space, President John F. Kennedy
in 1961 set as a goal the landing of a man on the moon by the end of the decade
(https://www.youtube.com/watch?v=g25G1M4EXrQ). NASA funded research on such
topics as team behavior, human stress, and other topics critical to the success of manned
spaceflight and was a major stimulus to I/O psychology research, theory, and application.
In the end the United States won the race to the moon, but the application of I/O in this
area continues.

The National Aeronautic Space Administration (NASA) was created and teams of
engineers, scientists, and technicians engaged in what seemed an impossible task. NASA,
which was charged with developing and launching manned spacecraft and exploring
outer space was faced with not only a daunting technological challenge but also with
challenges related to human performance and organization. A spinoff of the NASA
achievements was a new way of conceiving of organizational structure. The old classical
model in which managers and employees were held to a strictly hierarchical chain of
command did not work in coordinating the efforts of engineers and scientists on a task as
complex as transporting men to the moon. An organizational design was needed that
would allow teams of employees to coordinate their efforts in the various details of the
space mission, encourage creativity, and at the same time maintain control over the entire
process. The type of organization that emerged was the matrix organization which
required a much more complex view of the organization than those set forth in the
classical and human relations approaches.

One consequence was the matrix organizational structure. The matrix organization is an
attempt to capitalize on the advantages of organizing around projects and functional
organization while minimizing the downsides to each. In the traditional functional
organization, each part of the organization (e.g., marketing, research, and manufacturing)

may lose sight of the products or services that the organization is trying to generate. In
the project organization, which is based around products or services, there are usually
inefficiencies in coordinating the various functions. A matrix organization attempts to
avoid the downsides of each type of organization by combining the project and functional
structures. Each project manager in this structure is responsible for completing the project
and must work with functional managers who assist in mustering the functional resources
needed at various stages of the project. In violation of the classical principle that each
employee reports to one and only one manager, each employee in a matrix organization
reports to two managers. One example is the matrix organizational structure of the NASA
Jet Propulsion Lab. Along the right side of the organizational chart are the projects of the
JPL, Mars Exploration, Earth Science and Technology, Solar System exploration, etc.
Each of these consists of project teams led by project managers. Along the top of the
chart are the functions performed by JPL including Business Operations, Engineering and
Science, and Safety. Another example is illustrated in the figure below. There is a
functional structure represented by research, production, sales, and finance and product
structure and product teams. Employees are assigned to both a function and a product.
Each one reports to a functional manager as well as a product manager. Matrix structures
are complex and require new approaches to managing people
(http://www.buzzle.com/articles/matrix-organizational-structure.html).

Figure 2.3: A Matrix Organizational Structure
The civil rights movement. In the mid-1960s, large numbers of baby boomers entered
early adulthood; the U. S. and other nations finally were forced to come to grips with
their discriminatory treatment of minorities and women. Although the immediate post-
WWII period was one of sustained prosperity in the United States, the contrast between
the “haves” and “have nots” was never more visible. And most disturbingly of all, a
society that prided itself on its orderly, civilized institutions of governance and justice
suddenly found itself replacing leaders through assassination and forcing change through
violent protest. The whole fabric of U. S. society seemed to come apart at the seams.
Violent disagreement over the Viet Nam War, civil rights, economic policy,

http://www.buzzle.com/articles/matrix-organizational-structure.html

environmental policy, consumer affairs, and many other issues split society into polarized
factions. The baby boomers, the most vocal demographic group, forced the U. S. and
other societies to take a hard look at the collective values that had been taken for granted.

The implications of all this for organizations and I/O psychology were profound. The
prevailing attitude among young people about to enter the workforce was that profit-
making organizations and many other institutions were immoral. They saw them as part
of the establishment that gave us war, poverty, discrimination, and pollution purely to
satisfy selfish motives. Companies found recruiting very difficult, and the young people
they were able to attract were difficult to assimilate into the traditional corporate culture.
This new breed of employees had its own ideas about personal appearance, lifestyle, and
corporate values. They tended to question authority. It was not an easy time for
managers raised in the classical tradition. Other longstanding traditions were challenged
as well, such as the organization’s prerogative to hire, promote, and fire whomever it
wished. Out of the turmoil came a mass of legislation and case law on civil rights, and a
federal agency, the Equal Employment Opportunity Commission (EEOC), to implement
it. Companies were forced to break old patterns of discrimination against women and
minorities, although in most cases not without considerable resistance.

Not surprisingly, the management philosophies and personnel management techniques
that had earlier been dismissed as impractical resurfaced under these chaotic
circumstances. The climate was right for organizations to consider new ways of
managing their affairs. In many instances they had no choice. The new generation of
employees, civil rights laws, declining productivity, and foreign competition were the
new realities; something had to be done.

The rise of organizational psychology

I/O psychologists in the decade following WWII focused their attention on human
resource management (the I side). They conducted research on assessment tools and

programs that management could use in job analysis, compensation, performance
evaluation, personnel selection, and training. This work lacked a theoretical foundation
and led to accusations of “dustbowl empiricism” in which psychologists accumulated
data without any unifying conceptual frameworks to provide coherence to the mass of
findings. With the emergence of organizational psychology, the focus shifted to theory
and research that examined the interplay among individual, group, organizational, and
contextual factors. Organizational psychologists realized that no one discipline had the
complete answer either to the question of what organizations are or how management
should design and manage them. They drew liberally from disciplines besides
psychology, including anthropology, sociology, political science, and even economics.
Three organizational perspectives were dominant in the post WW-II years: the modern
human-relations theorists, the decision theorists, and the open systems/contingency
theorists.

Modern human relations theorists. In the post-world war II years the old human relations
movement evolved to incorporate more complex perspectives on the organization.
Perhaps the most important difference between the modern human relations theories and
the earlier movement was that the modern approaches advocated changing the
organizational structure to increase employee participation and involvement whereas the
earlier approaches tended to see human relations as a way of gaining cooperation without
changing the organization. The applied field of Organizational Development emerged
using the modern human relations theories as the conceptual foundation and advocating
that the organizational psychologists serve as a change agent. Douglas McGregor, Rensis
Likert, Chris Argyris, and Warren Bennis were four of the more important of these
theorists.

Douglas McGregor was a professor at the MIT Sloan Management School who proposed
that assumptions about human nature were responsible for the ways that employees were
managed in organizations. The classical and scientific management approaches were
based on a Theory X set of assumptions in which people were considered to be
untrustworthy, lazy, and motivated primarily by money. These assumptions were
probably appropriate at the time but in the post-World War II years, McGregor believed
that a new set of assumptions were needed. Theory Y assumptions that people seek
challenge, autonomy, and responsibility. If management involves creates a work
environment in which employees can grow and take responsibility, employees will
demonstrate a high level of motivation and creativity. Unlike the older human relations
movement, McGregor proposed fundamental changes in organization to incorporate
Theory Y. These included allowing employees to participate in decision making,
enriching jobs, elimination of traditional performance appraisal, the creation of
autonomous work teams, and the development of employees. He stated his views in The
Human Side of Enterprise, a best seller and one of the most influential management
books of the 20th century (Head, 2011).

Rensis Likert (pronounced Lick-urt), a University of Michigan professor and social
psychologist, was among the most influential change agents of the era. An essential
component of his theory was that better performance would result if the whole
organization moves toward a more open, participative system in which employees are
allowed to become involved in important decisions. Among the successes attributed to
this group was a complete overhaul in the management style of a failing textile
manufacturer, the Weldon Company. Although various explanations have been offered
for what happened at Weldon, its profitability did improve and the increased
supportiveness and participation of management were widely cited as reasons (Likert,
1961; Marrow, Bowers, & Seashore, 1967). Consequently, both Likert and his theory
received a lot of attention. In fact, a play loosely based on this case, The Pajama Game,
became a major Broadway hit! (see https://www.youtube.com/watch?v=c2Xi_IyHYNc
for a scene from the movie).

Rensis Likert (1903-1981) is also well known for inventing the 7-point rating scale, also
so-called “Likert scales” and for his research and theorizing on leadership in the

“University of Michigan studies.” In general, he advocated more employee-oriented
leadership and supportive management. Consistent with the human relations movement
started by the Hawthorne studies, Likert recommended that managers engage employees
in participative decision making and provide emotional and social support.

Chris Argyris (1964) was another psychologist who built on the human relations
movement. He theorized that the normal healthy progression in human development is
usually reversed when adults enter an organization. Healthy human growth is from the
passivity that exists as an infant in which there is dependence on the parents, to
increasing self-determination and assumption of responsibility during the adult years.
When a person enters an organization, he or she is assigned to a narrow set of tasks and is
forced to regress to the state of dependence more characteristic of a child than an adult.

This frustrates the inherent needs of humans to grow as individuals and as a defense
mechanism the employee becomes passive, withdrawn, and apathetic. Consequently, the
lazy worker stereotype of the autocratic manager is a self-fulfilling prophecy in which the
autocratic actions of the manager are responsible for the perceived laziness of the worker.
Similar to Likert and others in the human relations movement he recommended that
managers involve employees in decision making and that they also attempt to create
openness and authenticity in their relationships with employees.

Warren Bennis (1969) is a fourth organizational psychologist who built on the work of
the human relations movement. He proposed that with an increasingly complex society
and technological innovations, it is inevitable that organizations will need to become
more democratic and engage employees as partners in decision making rather than as
subordinates. He described democratic management in terms of the following
characteristics:

* Authentic and open communication among all employees, regardless of their rank.

* Resolving conflict through problem solving and consensus rather than by forcing
solutions or avoiding conflicts.
* Influencing people through personal expertise and technical competence rather than
using rewards, punishments, and one’s office or title.
* Creating an organizational climate where people feel free to express emotion.

According to Bennis, “The study of leadership will be increasingly collaborative because
it is precisely the kind of complex problem—like the genome—that can only be solved
by many fine minds working together” (2007, p. 4).

Other scholars in addition to these individuals could be discussed, including Alfred
Marrow, John French, Kurt Lewin, Stanley Seashore and John Coch who contributed to
the growth of organizational psychology (the O of I/O psychology) in the two decades
after WWII. Although they all conveyed a strong human relations theme in their work,
unlike the early human relations movement they were, for the most part, researchers as
well as theorists. Also, in contrast to the early these theorists recommended training
supervisors in showing understanding and supportiveness in the management of
employees. The organizational psychologists of this period advocated Organizational
Development (OD) interventions in which the target of change was the entire
organization. Although still adhering to the humanistic principles of the human relations
movement, they were moving in a more open systems direction by bringing attention to
the fact that simply changing supervisor behavior is unlikely to succeed without changing
other aspects of the organization.

The decision theorists. Other theorists in the first decades after WWII built not only the
human relations movement but on information and decision making theories. These ideas
are partly economic, partly psychological view of organizations as decision-making
entities. Classical economic theory assumes that both individuals and organizations are
rational economic entities that attempt to maximize positive outcomes in the choices they
make. In contrast to this traditional view, a school of thought championed by Herbert
Simon, a Nobel-prize-winning psychologist/economist, and James March, an
organization theorist, argued that neither people nor organizations make strictly rational
decisions (Cyert & March, 1963; March & Simon, 1958). For one thing, they have
limited information-handling capability, have only a vague sense of what they really
want, and have limited opportunities to gather information. People “satisfice” rather than
“optimize,” to use Simon’s (1957) terms. People make decisions on the basis of what
Simon (1957) called “bounded rationality.”

For its part, the organization behaves even less like a rational, goal-directed entity than
individuals do. Composed of a host of individual and collective goals, many in direct
conflict with each other, it functions as a loose, shifting coalition that makes do in the
present rather than following a charted course to some future goal. When it succeeds, it
does so by incremental improvements over its present condition, not by systematic
pursuit of a well-defined objective. March and Simon’s (1958) view suggests, among
other things, that a person must understand human cognition, motivation, and influence
mechanisms to manage effectively. Moreover, designing organizations to function

optimally in a rational sense cannot begin to achieve the hoped-for results.

A theorist who based his own theory of organizations on these developments in decision
making and information processing was the sociologist James D. Thompson (1967). He
proposed in Organizations in Action that the design of organizations, the management of
people in these organizations, and the decision making of those responsible for these
organizations are driven by a basic human desire to reduce uncertainty. Closed systems
models of organizations in the form of the Classical/Scientific Management and the
Human Relations Movement reflect this tendency to reduce uncertainty by imposing
universal assumptions (e.g., all people are motivated by money and management must
closely supervise them or all people are basically motivated by social needs and
management should make them feel wanted and loved).

To quote Thompson:

“the organization has a limited capacity to gather and process information or to predict
the consequences of these alternatives. To deal with situations of such great complexity,
the organization must develop processes for searching and learning, as well as for
deciding. The complexity, if fully faced, would overwhelm the organization. Hence, it
must set limits to its definitions of situations; it must make decisions in bounded
rationality. This requirement involves replacing the maximum-efficiency criterion with
one of satisfactory accomplishment – decision-making now involves satisficing rather
than maximizing (Thompson, 1967, p. 9).”

The open systems/contingency approach. So far the chapter has discussed two contrasting
approaches that emerged from the experiences, research, and theorizing of managers and
social scientists in the first half of the 20th century. As already discussed, the Classical
and Scientific Management theories were dominant in the early years. The organization
was depicted as a machine that runs best when there is authoritarian leadership, highly
specialized tasks, technical supervision, a focus on individual employees, a hierarchical
structure, centralized decision making, and the use of monetary incentives. The early

human relations movement (i.e., Mayoism) suggested psychological ways of gaining
acceptance and reducing resistance to this approach. The modern human relations
approach emerged during the post-World War II years and viewed the organization
differently. This approach stated that organizations are not machines but collections of
people whose needs and attitudes are important. From this perspective the ideal
organization is participative, engages employees in interesting and challenging work,
emphasizes the human relations skills of supervisors, is focused on managing groups of
employees rather than individual employees, uses an organizational structure that
contains relatively few levels and is decentralized, and uses higher level human needs
such as social and self-esteem needs to motivate employees.

Both of these approaches had insights into the management of people in organizations
that were correct, useful, and fit the societal events occurring at the time they emerged.
Both were also simplistic in light of the turbulence and instability of the post-world war
II period. In a closed system there is relative certainty about what is happening and what
managers should do because there is little need to worry about unexpected events inside
and outside the organization. Closed systems thinking is tempting because it makes
decision maker feel better about their choices. Such thinking probably made sense in
simpler times, but in the context of the increasingly complex and unstable environments
with which organizations must contend, closed system thinking is vulnerable to
disastrous, unexpected events. Gradually, an understanding of the true complexity of
work organizations began to emerge, and the insufficiency of simplistic prescriptions
such as those offered by either the mechanistic or the human relations movement became
apparent. The research in the two decades after WWII made clear that while each was
appropriate in some situations, neither is universally correct or incorrect. One cannot
ensure effective management by making workers happy and capitalizing on informal
group processes any more than one can by planning and directing their every move. The
best way of designing and running an organization depends on a host of considerations at
the individual, group, and macro-organizational levels. This general philosophy is the
open systems/contingency approach. This third approach combines many of the elements

of the organizational psychologists of the 1945-1965 period and the decision theorists. It
evolved directly from the knowledge accumulated from scientific research and remains
dominant to this day (Scott, 1987). The primary attributes of the open
systems/contingency approach are a rejection of universality and the assertion of
contingencies, attention to the environments in which organizations operate, and the
interdependencies among subsystems and between the organization and the environment.

1. Contingencies. Open-systems/contingency approaches assume that there are alternative
ways of reaching the same goal and reject the universality of the solutions proposed in
the Classical and Scientific Management as well as the Human Relations movement. In
the open-systems/contingency approach, the best approach is contingent on the situation.
What works in one situation may not work in another and in deciding what to do one
must carefully consider the situational factors at work. Determining these situational
factors requires research. Much of the discussion in this text is devoted to the findings of
research that led to contingency statements about work motivation, work-related
attitudes, motivation, work groups, leadership, selection, performance appraisal, stress,
and other topics in I/O psychology.

2. Open-systems are embedded in an environment. As depicted in figure 2.4, the open-
systems approach depicts the organization as existing in various environments. The
environments of the organization were largely ignored in the Classical and Scientific
Management theories as well as in the early Human Relations movement. The
mechanistic (classical and scientific management) approaches and the human relations
approaches assumed that organizations were closed systems and ignored what is outside
the organization and outside each subsystem.

The environment of the organization refers to all those things that are outside the
organization that influence how the organization functions, including the economic,
political, legal, cultural, technological, and other contexts in which the organization must
operate. Unlike a closed systems approach, an open systems perspective recognizes what

is outside the boundaries of the system and the organization’s attempt to adapt to the
environment. If the customer base, labor market, or prime interest rate changes, or the
government introduces a new environmental regulation, the system knows about it and
responds. In fact, it attempts to anticipate the changes so as to respond early enough to
cushion or avert the blow. For example, an auto manufacturer might conduct market
surveys to determine how potential buyers (customers) would like a proposed change in
the design of its automobiles. If the results indicated a strong negative reaction, the
company might think twice before implementing it and avoid a costly mistake. It is
important to recognize that the boundary separating the system from its environment is
more than just the physical limits within which it operates. A boundary can have
symbolic, cultural, and psychological features as well. Political action groups such as
Mothers Against Drunk Driving (MADD) or anti-abortionists (right-to-life movement)
are primarily distinguished by these conceptual boundaries, which they work constantly
to expand.

Environments also exist at the subsystem levels of the organization including the
individual worker, work group and other departmental units, the technologies used, and
other component levels of the organization. At the level of the people in the organization
the environment includes contexts within the organization such as the work group, the
department, and the organization as a whole, as well as contexts outside the environment.

3. Interdependencies among subsystems and with the environment. A supervisor using
closed-systems thinking can decide how to manage his workers without worrying about
other workers in the plant, other work units, the nature of the technology, or events
occurring outside the organization. The CEO using closed-systems thinking can decide
how to run the entire organization without worrying about outside forces or the complex
interdependencies within the organization. In contrast to a closed system approach, open
systems approaches depict individuals, groups, technology, leadership, and other
component elements of the organization as interdependent with each other and
interconnected with forces outside the system.

This brings us to another important characteristic of open systems: interdependency.
Interdependency means that everything depends on and affects everything else.
Interdependency exists between the internal components of the organization and the
environment of the organization. Increased foreign competition is an environmental
factor that may require that the organization change the size and composition of its
workforce, its job descriptions, its compensation schedules, its management structure,
and a host of other internal characteristics to continue functioning effectively. The
existence of a strong labor union is another potential environmental factor that would
influence the people, technology, structure and leadership of the organization. Still
another potential environmental factor is government regulation of the selection, training,
compensation, supervision, and other aspects of the management of employees in an
organization. Other environmental factors, in addition to these, shape the people,
technology, strategy, structure, and leadership of the organization, and are in turn
influenced to some degree by what happens inside the organization. There is also
interdependency among the internal components of the organization (e.g., the people,

technology, structure, strategy, and leadership of the organization).

Changes in any one component have implications for the other components. Also,
changes in any one component (e.g., enrich or simplify the jobs, change leadership,
introduce new technology, change the structure of the organization in the direction of
more or less centralization) requires taking into consideration the other components.

Figure 2.4: Interdependence among subsystems of the organization
and between the organization and external environments.

4. Predictions from an open systems/contingency approach. Open systems/contingency
approaches propose that organizations will naturally evolve to cope with and adapt to the
environment. How well the organization performs depends on whether the ways that they
evolve fit the environment. Organizations characterized by high uncertainty tend to
adopt structures and practices more in align with a human relations or humanistic model.
On the other hand, organizations operating in environments characterized by certainty
and routine tend to adopt mechanistic structures such as those proposed in the
classical/scientific management theories. Not only do organizations tend to adopt these
structures but also those organizations that deviate from these tendencies do not appear to
perform as effectively as those that conform to these tendencies. Lawrence and Lorsch
(1967) provided one of the earliest demonstrations. These researchers compared the
structures adopted by successful and unsuccessful firms in two different industries. The
successful firms in the container industry, which operated in more stable environments,

tended to have structures consistent with a mechanistic model (classical/scientific
management). The successful firms in the plastic industry which had to contend with
more unstable conditions, tended to have structures more in line with a humanistic model
(i.e., a model consistent with a human relations approach).

Organizational theorists such as Tushman and Nadler (1978) would agree with the
decision theorists that how effectively managers process information and make decisions
are critical to organizational effectiveness. In performing these two tasks one could use a
vertical approach as advocated in Classical Theory and Scientific Management or a
horizontal approach as advocated in the Human Relations (Humanistic) approach.

Figure 2.5: Vertical and Horizontal Mechanisms of Information Processing

In organizations operating in environments where there is relatively little competition, a
low level of technological innovation, and other characteristics of a simple, stable
environment, the vertical mechanisms dictated in the Classical and Scientific
Management approaches may well suffice. Consistent with these approaches, top level
managers might use (1) vertical means for processing information, requiring that all
important matters pass through their hands for consideration, (2) impose rules,
procedures, and other bureaucratic tools, (3) design work with specialization in mind, and
(4) motivate workers with monetary incentives for doing the job in the one best way that
research has identified. Vertical means of processing information emphasize efficiency of
operations at the cost of flexibility and innovation. Vertical means of processing
information also imply respect for authority as those higher in the organization dictate to
those below how they should perform their tasks. In summary, in a stable, low
complexity environment managers might make the decisions, employees are required to
follow rules and procedures on all matters, with each member of the organization
performing a specialized task and expected to submit to a central authority.

In a relatively simple, stable environment all of this may work quite well. What happens,
however, as environmental complexity and instability increase? One likely consequence
is that top level managers are unable to process the information or keep up with the rapid
changes occurring in the environment of the organization. At this point they attempt to
process the information through horizontal procedures more characteristic of the modern
human relations approaches. Workers are consulted and allowed to participate in decision
making. Managers may encourage more open communications rather than adhering
strictly to the chain of command. These deviations from the hierarchical rule of an
organization rely on the cooperation of the people involved. and to gain this cooperation,
managers will need to pay attention to worker needs and morale. The modern Human
Relations approach is consistent with the loosening of the way organizations are run. The
use of horizontal information processing mechanisms also implies opportunities for
learning as employees interact more freely with others in the organization and in the
process acquire new competencies. In summary, when operating in a complex, unstable
environment management will need to allow more open and horizontal communication,
more participation of employees in decision making, rely more on groups, and place less
emphasis on rules, procedures, and specialization.

5. Examples. To illustrate consider three hypothetical examples. The Couch Potato TV
store is run by Ralph. At the start of the store in the 1950s, Ralph supervises three
employees and operates in a rather stable situation. There are few competitors in the
neighborhood, and there is little uncertainty about the types of TVs to sell or the nature of
those TVs. They are black and white and only 2 or 3 companies manufacture them. The
employees are low wage part-time employees who are easily replaced if need be. Ralph
could run this store as one person and for the most part call the shots as he sees fit. Let us
assume that Couch Potato TV is successful and as a consequence, Ralph decides to
expand. Now color TV has entered the market and there are many more product lines and
manufacturers. The basic nature of the technology has not changed dramatically but
demand has increased and to meet this demand Ralph moves his business to a larger
store, hires more employees, and expands his product line. With increased demand and
product line running the shop as a one-person show is more difficult. It’s simply too
much for one person to handle and as a consequence he implements some changes to help
deal with the increased work load and to make sure his employees satisfactorily sell the
TVs and perform their duties correctly and efficiently. Ralph has just read the writings of
Fayol, Weber, and Taylor and constructs a set of rules and procedures to govern how
things are done in the store. So in essence, Couch Potato TV has implemented a
mechanistic model following the dictates of Classical/Scientific Management theorists.

All is fine, but as time passes, Ralph finds that the stable environment in which his store
operates is becoming increasingly unstable and complex. There is an increased focus on
energy conservation, environmentally friendly practices, diversity in the workforce and
consumer tastes, and increased rules and regulations at the local, state, and federal level.
There is more competition. Other TV stores are moving into the neighborhood. The
variety of TVs has dramatically increased from the old days when only three or four
could serve as the product line. The technology changes at an accelerating pace with the

introduction of solid state technology, desk top computers, and the internet. Government
regulations have increased such that Ralph must pay much more attention to how he
manages his workers. Things have changed within the organization as well. In the past,
workers seemed grateful to have a job and never questioned his authority. The employees
who now work in his shop are not as submissive and frequently question the ways things
are done. At the same time, Ralph is not an engineer or a computer expert, and he is
increasingly dependent on the expertise of the employees.

At this point the reliance on a hierarchical, rule driven, and efficiency focused
management style simply does not work in processing the amount of information
available. To deal with the complexity he also decides to have employees participate in
decision making and encourages full communication among them rather strictly adhering
to the chain of command. This loosening up of Couch Potato leads to a recognition that
employees have a variety of needs in addition to money and that a more human relations
approach is needed in which employees’ feelings and perceptions are taken seriously.
Ralph has not thrown out the rules, procedures, and many of the other components of the
Classical model but has added horizontal means of processing information to these
vertical means of information processing.

Now let us consider another hypothetical company…Wizard Manufacturing. Wizard
consists of four major departments: marketing, production, research and development,
and finance and accounting. The CEO of Wizard. Homer Simpson…runs this firm with
an iron hand. There is an emphasis on following the rules and procedures. Workers are
expected to obey and submit to their supervisors and little room is given for individuals
to veer from the methods that have worked so well for Wizard in the past. The firm is
also run with an emphasis on the individual rather than the team. There are individual
financial incentives for good performance. Each worker is given a specialized set of tasks
to perform and he or she is responsible for these tasks without asking or giving assistance
to other workers. Workers are expected to keep social interaction to a minimum during
work hours and to concentrate on their individual duties. Communications within the
organization pass up and down the line with little horizontal interaction. Workers are
expected to go through their immediate supervisor first when attempting to communicate
with other workers in the organization. If one were to use the McGregor distinction
between Theory X and Theory Y, Wizard is definitely a Theory X firm in its
management styles, communication patterns, decision making, mechanisms for
motivating workers, and bureaucratic structures.

This way of operating has worked well for Wizard but now there is increasing
competition and growing uncertainty due to changing consumer demand and
technological innovations. A mid-level manager at Wizard…Dudley Dooright… believes
that a Theory Y approach is more effective in coping with these changes. With the
permission of the CEO, Bart decides to organize his department around teams of workers.
These teams are allowed to participate in important decisions. Moreover, rules and
procedures are deemphasized. Rather, workers are expected to show initiative and to
innovate. Dress requirements are loosened as well work schedules. Now workers can
often work at home or vary their schedules. These new procedures work very well. The

productivity of Dudley’s department is very high and much higher than prior to the
changes. However, complaints soon arise from workers in other parts of the company
about the freedom that Dudley’s group has been given. Dudley’s group comes up with
many new and creative ideas but the rest of the company remains skeptical and ignores
most of them. There is growing pressure on the CEO to bring Dudley’s department back
in line with the rest of the organization. Eventually, the CEO orders Dudley to restore his
operation to the way it previously operated.

Open systems/contingency theory provides an understanding of what is happening in this
case at two levels…the organizational level and the level of the subsystem. From the
perspective of the company as a whole, classical/scientific management has worked well
and probably fits the situation. However, that situation is changing and Dudley Dooright
is probably right in wanting to move the company in the direction of more horizontal
mechanisms that would allow for more flexibility and adaptability. At the level of the
subsystem, one sees open systems/contingency theory at work. Wizard is a system and as
such there is interdependence among the various subsystems. One cannot change one
subsystem without taking into account the other subsystems. Turning Dudley’s
department into an island of Theory Y in the midst of a sea of Theory X simply will not
work. Dudley’s way of operating, while perhaps the best way of dealing with change and
uncertainty is inconsistent with the larger organization and is eventually rejected. To
return to a key characteristic of open systems/contingency theory, organizations are
complex systems and changes in one part of the system require taking into consideration
the other parts of the system.

6. Summary of the open systems/contingency approach. In summary, one can distinguish
between vertical means of dealing with the uncertainty stressed by the classical/scientific
management approaches and the horizontal means stressed by the human relations
movement. Whether a more traditional, classical approach to the management of people
or a more complex, human relations approach is effective depends on the stability and the
complexity of the organization’s environment. In the latter half of the 20th century and
the first decade of the 21st century, all organizations have experienced increasingly
complex and changing environments. However, there are still individual differences with
some organizations having much more stable and simple environments than others. The
approach that management takes to designing an organization and managing the people
within the organization depends on the uncertainty imposed by the complexity and
stability of the environments.

The Modern Era (1990 – now)

Anyone 18 or older is familiar with the major events of this period. While the first two
decades after WWII were characterized by U. S. economic and military dominance,
unprecedented economic growth and optimism in the future, other events beginning
around 1965 and continuing to the present time are reshaping both I/O and thinking about
organizations. The “modern era” begins in 1990 when the Cold War ended. The major
forces during this time that shaped I/O psychology include global competition, computer
technology and the internet, terrorism, and increases in population and resource scarcity.

Global competition

So far discussion has focused on the dominance of the United States, politically,
economically, and militarily. Although the U. S. is still the most powerful nation in the
world, it now has competitors. Foreign competition in trade began to make significant
inroads into the U. S.’s unchallenged leadership as supplier of the world’s manufactured
goods. Former vanquished foes in WWII, the Japanese and Germans, are now major
economic competitors. China and other nations in the Pacific Rim have emerged as the
fastest growing economies in the world. In search of cheaper labor, many U. S.
manufacturers have moved their operations overseas resulting in a major decline in
employment in the areas of the U. S such as Pennsylvania, Michigan, Ohio, and Indiana
that were the centers for the production of automobiles, steel, and a variety of other
goods. In the past corporations around the world were identified with one nation (e.g., U.
S. Steel with the U. S., British Petroleum with Great Britain). The economic foundation
of U. S. affluence was beginning to erode. The growth in productivity that the U. S. took
for granted declined dramatically (Kochan & Barocci, 1985). Most corporations now
have a strong international component with operations and employees around the world.
Among the implications of the globalization of trade is that I/O psychologists can no
longer afford to look at human behavior in the workplace through exclusively North
American lenses. I/O psychologists and other social scientists involved in the study of
organizations were almost exclusively North American and British prior to the 1980s, but
over the last three decades, academic programs in I/O and other organizational studies
have emerged in Asia, Europe, and even the Mideast. Also occurring in the last three
decades is a large increase in cross-cultural research on all the topics covered in this
course.

Computer technology and the internet

Another remarkable event is the accelerating change in computer technology in the
workplace. Computers, which were once incredibly expensive machines occupying entire
buildings, became tools accessible to the general population. Almost overnight in the
1980s, the computerized checkout counter became a standard retail store fixture and the
computer workstation replaced the typewriter on every secretary’s desk. Now workplace

technology includes computer networking, mobile technologies, and virtual reality, to
name only three. It is hard to fathom the impact that computer technology has had on
organizations and the workplace and even harder to project the future effects.

The ways that organizations and work are designed and managed are driven by the need
to process information and make decisions. The classical form of organization was one in
which people are arranged in a pyramid with those at the top possessing the most
authority and power. Orders and directives flow downward and communications are
required to follow chain of command. The more complex the tasks and environments of
the organization, the more levels in the hierarchy and the more specialization of the tasks
were required to deal with the massive amount of information to process. Those at the
bottom of the organization were responsible for day-to-day decisions about small matters,
following the rules, procedures, and policies imposed from on high. Exceptions to the
routine way of operating that were not covered by existing rules, procedures, and policies
were kicked upstairs. This allowed those at the top to deal only with the important
matters and avoid an overwhelming load of information and decisions. The fact that there
were often many links separating those at the top from those doing the work at the bottom
reduced the capacity of the organization to respond quickly to change or to deal with
unexpected events. In a simpler time, the hierarchical structure served us well, but the
increasing complexity of modern times have forced a move away from a strict
hierarchical form of organizational structure toward structures such as the matrix
organization of NASA that can provide more effective processing of information.
Transportable, affordable computers with massive computing power has allowed those at
the operational level to instantaneously communicate with those at all other levels of the
organization and to collaborate across vast distances. Take, for example, the development
of networking in the military (http://www.defencetalk.com/drones-to-take-network-
integration-evaluation-by-swarm-65538/). A soldier on the ground in Afghanistan carries
a computing device and can feed information to operators in command centers in the
United States who launch and control drone weapons. These innovations in computer
technology and the networking applications of this technology are having and will have
dramatic effects on how leadership, teamwork, personnel selection and training, and a
variety of other HRM and OD functions are viewed (e.g., Bias & Bogue, 2015; Lin,
2011; Miles, Snow, Fjeldstad, Miles & Letl, 2010; Thomas, 2014).

Drones to Take Network Integration Evaluation By Swarm

Terrorism

With the end of the cold war, the democratization of the former Soviet Union, and the
embrace of capitalism in socialist societies such as China, one might have expected a
period of peace and security. However, the threat of nuclear annihilation by the great
powers which was dominant in the two decades after WWII was replaced by a new
threat…the use of terrorism by small groups with radical agendas. The bombing of the
Federal building in Oklahoma City by right wing terrorists in the U. S. and the 9/11
attack by fundamentalist Islamic groups are only two of the many acts of terrorism that
have occurred over the last three decades. The attempts of the U. S. and other countries
to increase homeland security and to protect against such attacks are having a continuing
influence on I/O psychology and the management of organizations. One important
function of I/O psychology has always been the reduction of uncertainty associated with
the use of people in the workplace. When employers hire people for jobs they are never
entirely certain of how the person will perform. Will the hires meet performance
standards or will they prove less than adequate? Even more of a concern in many
organizations since 9/11 is whether the persons hired are trustworthy and honest
individuals or potential threats to the well-being of the organization, customers, and other
employees. Since 9/11, I/O psychologists have been called upon to device assessment
tools that can screen out those who might perform terrorist acts and to devise training
procedures for the increasing number of employees who are engaged in security roles
(e.g., Halbherr, Schwaninger, Budgell & Wales, 2013). Computer technology is now
routinely used to monitor the everyday activities of employees and customers. This has
raised concerns about privacy and civil rights and research is underway on the impact on
employee attitudes (Holland, Cooper & Hecker, 2015). I/O psychologists and other social
scientists are beginning to conduct research on the impact of the threat of terrorism on
organizations and employees (e.g., Malik, Abdullah & Uli, 2014; Niaawn, Nielsen,
Solberg, Hansen & Heir, 2015; Shirom, Toker, Shapira, Berliner & Melamed, 2008). The
ultimate impact on the way organizations are structured and managed remains to be seen.

The population explosion, diminishing resources, and climate change

Other important pressures for change were presented as a consequence of diminishing
natural resources…especially fossil fuel. An embargo in the Middle East in the 1970s
shocked the United States into the realization that it was dependent on foreign
governments, many of them unstable at best, for its most vital source of energy. Gasoline
shortages, dramatic increase in the price of everything related to energy (gasoline,
electricity, heating oil, manufactured products, etc.), and the search for alternative
sources of energy became issues that some argued should constitute the highest national
priority. Moreover, the energy crisis exacerbated a worsening economic trend that saw
inflation growing, productivity declining, national debt soaring, and the balance of
payments (value of exports over imports) becoming negative. The U. S. was about to
relinquish its position as the world’s largest lender nation to become the world’s largest
debtor nation. The era of prosperity was clearly headed for trouble. Although oil prices
dropped in 2015 due to a temporary glut in supply, the long-term prospects are for
shortages and an eventual return to the difficulties encountered in the 1970s.
The climate changes created by the growing use of fossil fuels in not only the U. S. but in
the developing countries presents still another challenge to organizations as they move
forward into the 21st century. Increasingly, individuals and organizations are adopting
more environmentally sustainable practices to deal with the decline in resources and to
slow the harmful effects that carbon emissions are having on the climate. The effects of
climate change are potentially catastrophic and eventually may lead to a host of negative
consequences including the extinction of species, water scarcity, shoreline flooding,
disease, starvation, and war. I/O psychologists around the world are beginning to
contribute in their research and practice to knowledge about how to promote and
encourage environmentally practices in the workplace (Aquinas & Glavas, 2013; Dilchert
& Ones, 2012; Ones & Dilchert, 2012).

The changes that are yet to come will undoubtedly lead to new ways of thinking about
organizations and how to manage them. These new theories are in the future, but
industrial and organizational psychologists are hard at work conducting research and
devising interventions to meet these challenges. The open systems/contingency views of
organizations are even more relevant in light of events that have occurred since 1990.
However, the field of I/O psychology specifically and the organizational sciences more
generally are a work in progress and who knows what new paradigms will emerge as
academics and practitioners attempt to deal with the changes that are occurring in the
workplace.

Points to ponder

1. What were some of the historical events and societal trends that this chapter has left
out? How do you think they might have shaped the way organizations and work were
understood and managed?
2. Search the websites for some organizations. Look for examples of the three major
theoretical orientations to organizations and the management of people.
3. Provide examples from your own experiences of good and bad bureaucracy. In the
good examples how did bureaucracy help or provide benefits? In the bad examples how
did bureaucracy go wrong?
4. There is continuing debate on whether a human relations approach to managing people
is worthwhile or wrong-headed. What do you think? How can a human relations benefit
employees and the organizations in which they work? How can a human relations
approach go wrong and perhaps even harm employees?
5. What are the implications of an open-systems/contingency approach for the
management of people at work? What are the demands of this approach on managers?
What are the implications for the selection and training of managers?
5. Identify some of the changes that you believe will occur in the next 100 years and their
impact on how we understand and manage human behavior at work.
Conclusions

This chapter has reviewed how I/O psychology grew from a handful of academic

scientists at the beginning of the 20th century into the international, multifaceted
profession it is today. This chapter has examined the stresses and strains that have
accompanied this growth. As we have seen I/O psychology did not evolve in isolation. To
understand I/O one needs to consider the historical, cultural, and professional contexts in
which the field was embedded. The early development of I/O psychology paralleled the
emergence of the classical/scientific management and human relations approaches to
organizations and management. As discussed in this chapter, modern thinking is
dominated by the open-systems view, which recognizes the critical importance of the
social-political-cultural-economic-technical environment in which organizations operate.
The most notable characteristic of I/O psychology today is the commitment to a scientist/
practitioner philosophy that has helped define I/O’s niche in the overall picture of modern
psychology. The turbulence and change that the 21st century promises to bring will
undoubtedly lead to further changes in the field, but this core commitment to the
scientist/practitioner model is likely to remain as a guiding principle.

CHAPTER 3: I/O PSYCHOLOGY AS A SCIENCE

Introduction

In the first chapter, I/O psychology was described as a discipline concerned with the
behavior of people in work settings, and the scientific method was described as the means
of acquiring knowledge of behavior at work. Most of the knowledge generated and
applied by I/O psychologists is based on a philosophy shared by all sciences. The only
difference is that the research methods used in I/O psychology are adapted to the special
problems inherent in studying human beings. This chapter explores the essential
components of the scientific method and its applications to studying human behavior in
organizations. Although the readers may or may not ever conduct a study, at some time
they will consume research on human psychology in the workplace. They will read
articles and watch TV shows that offer advice on how to pick a job, interview, deal with
the boss, and in other ways get by in the world of work. They may even encounter one of
the growing number of consultants who use the old “science has found” routine to peddle
their wares. This chapter will help in distinguishing between truth and the many half-
truths and shams by examining the differences between scientific and non-scientific
research and the standards against which scientific research is judged.

The importance of the scientific perspective cannot be overstated. For one, the scientific
method is the one constant in I/O psychology. As is the case with all areas of science,
knowledge of the topic is always changing as new research is conducted, but the science
that generates the knowledge remains the same. Moreover, unless one understands the
rationale and conventions underlying the scientific approach and the research strategies
used by I/O psychologists, one can hardly judge the merit of principles and explanations
presented in this text. Adopting a scientific perspective is not an easy undertaking. People
often have difficulty accepting scientific explanations for human behavior that are
inconsistent with their own intuitions. They are inclined to trust their “gut feelings” over
scientific evidence and to react accordingly, often with unhappy results. To benefit from
what I/O psychology has to offer in the search for explanations and solutions, one must
consider the possibility that intuitions are frequently wrong and must weigh evidence
dispassionately. One must, in short, understand the scientific approach. For all these
reasons, this is probably the most important chapter in the entire book.

Distinguishing Science from Non-Science

A basic message of this text is that understanding and predicting human behavior in
organizations requires scientific research. It is important to realize, however, that there
are many alternative sources of knowledge, such as personal experience, observations of
others, authority, logic, intuition, and analogy. To gain some understanding of what the
scientific approach to understanding organizations entails, the chapter begins by
examining what it is not. To supplement what is said here, take a look at this fairly
general discussion of the differences between science and non-science:
http://www.youtube.com/watch?v=PrlI9oErJJg

http://www.youtube.com/watch?v=PrlI9oErJJg

Characteristics of non-scientific knowledge

Perhaps the most common source of knowledge and the most influential is personal
experience. Managers acquire a sense of what works and does not work from
participating in the real world. There is a long history of corporate leaders who have
offered advice on how to best run an organization. Several of these were mentioned in the
last chapter (e.g., Henri Fayol, Chester Barnard). More contemporary examples of this
genre of executives include John Mackey (Whole Foods), Tony Hsieh (Zappos), Michael
Dell (Dell), David Packert (Hewlett Packard), Richard Branson (Virgin Airlines), and
Donald Trump (Trump Organization), all of whom have confidently drawn from their
experiences to proclaim the right ways to manage.

Reliance on the personal experience of others presents another source of knowledge —
authority. One may accept an individual’s opinions as the truth out of fear, love, or
respect for that individual, because that individual’s advice was followed in the past, or
because of that individual’s past successes. Knowledge of how organizations are run are
also derived from or dictated by religious authorities. Some managers may take the
simple Christian dictum “Do unto others as you would have them do unto you” as a
principle that guides their daily activities. Others might rely on the teachings of Buddha,
Mohammed, Tao, or the Talmud for knowledge of organizations.
Closely related to authority is intuition. Managers often take a course of action because it
“feels” right to them. Intuition also can come in the form of mysticism, altered states
(drug or naturally induced), religious experiences, or inspiration. In all cases, the person
arrives at some “truth” through “gut feelings” rather than through logic or factual
information.

Very different from intuition is knowledge that is logically deduced from basic
assumptions. Similar to a mathematician who starts with certain axioms and through
rules of logic forms deductions, persons may start with stated or unstated assumptions
about organizations from which they logically derive what they accept as the truth. For
example, a manager may start with a universal belief (Theory X?) that workers desire
only money and are lazy. It would logically follow from this assumption that managers
should closely supervise workers, that they should motivate them with monetary
incentives, and that they should not seek their opinions or ideas.

One can also Knowledge based on analogy such as “A good manager is to an
organization” as “A quarterback is to a football team.” A perusal of books on
organizations and management reveals a plethora of others, including the manager as a
Samurai warrior and as a Mafia boss.

Finally, there is the knowledge that comes from the values and norms that members of a
society share and that constitute their culture. Members of an organization may come to
see those beliefs that they hold in common as fact rather than mere preference. Take, for
example, norms for how one should dress. There is no logical reason that an accountant
should wear a suit to work as opposed to overalls, but social norms dictate that the former
is usually more acceptable. Similarly, what one accepts as self-evident truths about

organizations reflects to a significant degree the accustomed ways of doing things.
Differences in social norms are particularly important in comparing different countries. In
Japan, for example, the individual is considered more subordinate to the collective effort
of the organization than in the United States, where there is a strong belief in
individualism. Consequently, the wisdom of using groups and consensus decision-
making is more apparent in Japan than in the United States, where distrust in groups and
reliance on the lone manager as hero prevails.

Please don’t conclude from this discussion that all knowledge based on nonscience is
invalid. There is a place for nonscientific approaches, but there is a lot of pseudoscience
out there, especially in the pronouncements of so-called experts on how to manage people
in the workplace. For an example, check out this link. The “Great Randi” makes a living
debunking a variety of hucksters, including the ones illustrated in the video. What he
talks about in this video is not uncommon in the selection of employees, believe or not.

Characteristics of scientific knowledge

All of the above are common ways in which people seek and acquire knowledge, but
what are the defining characteristics of scientific knowledge? Unlike the above sources
of knowledge, science is empirical, objective, precise, concerned with understanding
general principles, probabilistic, and logical.

Science is empirical. Perhaps the most important attribute of science is that it relies on
observation of events rather than relying solely on logic, authority, or intuition. Take for
instance the issue of what workers want the most in a job. Rather than assuming or
deducing some answer to this question, a scientific approach could be taken by
surveying, observing, or in some other way gathering data on employees’ needs.

https://secure.flickr.com/photos/45796762@N03/4773329769

Science is objective. Science does not rely on personal experience or intuition but
generates “objective” knowledge. By objective it is meant that different scientists can
arrive at the same conclusions using the same methods. One approach to ensuring
objectivity is to use standardized methods of measurement such as questionnaires,
stopwatches, and other devices that are kept the same across settings and researchers.

Objectivity doesn’t always require special equipment, however. A more basic
requirement is that the researcher carefully defines the terms and the procedures used in
gathering data. What if higher performance is observed when supervisory styles are
supportive and participative? Simply stating what was observed is not enough.
Objectivity requires stating the procedural details that allow others to repeat the
observations. This could include describing how supervisory style and performance were
measured, the type of employees used, sampling procedures, and the statistical techniques
that were used. In comparison, intuition and personal experience are often intensely
personal and incapable of replication. For instance, Donald Trump and other executives
may have profound insights, but their knowledge derives from their unique experiences.

Science is concerned with general understanding. Industrial and organizational
psychologists often attempt to understand what accounts for organizational events. Their
ultimate concern, however, is not with what caused a specific event at a specific time
(e.g., Why did WalMart perform so well last year or why did Fred, a manager at High
Tech, Inc, fail in motivating his employees?) as much as the general principles that apply
across a wide variety of events (e.g., What are the factors that distinguish consistently
successful firms from unsuccessful firms and successful managers from unsuccessful
managers?).

Science is logical. Scientists are logical in at least two senses. First, they derive
hypotheses from general principles using deductive logic. They then test these hypotheses
through research and draw inferences using inductive logic. They use the tools of
mathematics and logic, but unlike pure rationalism, scientists always go out into the
world to observe and test the hypotheses that they logically derive

Science is precise. Industrial and organizational psychologists attempt to define their
terms and collect their data in as precise a manner as possible. Consequently,
measurement is a particularly important activity performed by I/O psychologists. It is not
enough to simply distinguish satisfied from dissatisfied employees, or good performers
from bad. I/O psychologists develop measures that can distinguish with precision among
degrees of satisfaction and performance.

Science is probabilistic. A hallmark of science is that it never really arrives at truth with
total confidence. In testing hypotheses, scientists never prove, but only fail to disprove.
This is perhaps the aspect of science that is most frustrating to the nonscientists and to
managers in particular. Managers often want unequivocal answers, and they want them
fast. The scientist is prone to state conclusions in terms of probabilities, conditions, and
hypotheses. Scientists of all types, including I/O psychologists can sound like Tevye in
the play Fiddler on the Roof who is forever agonizing over decisions as he recites “On

the other hand.” http://www.youtube.com/watch?v=zm6KapIDLvg

Similar to Tevye, those who are responsible for actually managing organizations and
must live with the decisions, have a hard time with this aspect of science. They want
answers, not probabilities. Managers seek “one-handed” psychologists, who rather than
equivocating will declare what is universally and forever true. In many of the debates on
the issues of our day (e.g., climate change), the public has a difficult time accepting the
probabilistic nature of science. Yet, by demanding that science identify what is forever
true they are also stating that they do not need research and that the question is closed.
Such an approach is in the realm of religion and faith, but it is not science. This approach
also undercuts science and prevents the accumulation of scientific knowledge and
ultimately the most effective applications of that knowledge. There is a basic divide
between science and practice that is hard to close because of the natural desire for
certainty. A good scientist learns to live with and even enjoy ambiguity. I/O
psychologists are no different than most good scientists who realize that science is a
never-ending process and that what is accepted today as the truth may be questioned
tomorrow. In this sense, science is much different than other sources of knowledge that
are too often marked by dogma, rigidity, and a compete unwillingness to revise in the
face of facts.

The Scientific Method

In attempting to acquire knowledge concerning behavior in the workplace, I/O
psychologists use the scientific method. The scientific method is a process that starts with
research questions and then recycles back to additional research based on the findings:

Step 1. Start with a research question, e.g., What is the relationship of job satisfaction and
performance?

Step 2: Examine theory relevant to this question, e.g., Herzberg’s two-factor theory of
motivation (we’ll discuss this in the motivation chapter).

Step 3. From the theory, derive a hypothesis in which variables are defined and the
expected relationship stated, e.g., Workers who report higher job satisfaction are more
productive than those reporting lower job satisfaction.

Step 4. Operationalize the variables in the hypothesis by making them concrete and
measurable and/or capable of being manipulated in the lab, e.g., one could measure job
satisfaction using the Minnesota Satisfaction Questionnaire (MSQ) or performance by
recording the number of sales made by salespersons. One manipulates leadership style by
having leaders follow one of two scripts. The leader either allows subordinates to
participate in decisions or acts as an autocrat and tells subordinates what to do and how to
do it.

Step 5. Collect data to test the hypothesis, e.g., A sample of salespeople are given a
questionnaire to measure their job satisfaction. The company provides monthly sales
records for each employee.

Step 6. Analyze the data, e.g., submit the data to a correlational analysis using SPSS… a
statistical program, to compute the correlation between job satisfaction and performance.

Step 7. Draw a conclusion about whether our hypothesis is supported or not, e.g.,
examine the probability of finding this level of correlation in a sample drawn from a
population in which there is no correlation. This is the null hypothesis. Assume that a
significance level of .05 is taken as the criterion for rejecting or failing to reject the null
hypothesis. If the probability of finding the correlation in a sample when the population
(or true) correlation is zero is greater than .05, one fails to reject the null hypothesis and
concludes that the correlation is too small to be of much statistical significance. If the
probability of finding the correlation in a sample when the population (or true)
correlation is zero is less than .05, one can reject the null hypothesis and concludes that
the correlation is large enough to justify serious consideration. Note that although a p < .05 is the traditional value chosen in testing statistical significance, this is an arbitrary value. For a discussion of p values, go to: http://www.youtube.com/watch?v=HTZ8YKgD0MI Step 8. Reexamine our theory in light of the test of the hypothesis, possibly revise, and continue with more research. For an overview of the scientific method as used in psychology, take a look at http://www.youtube.com/watch?v=SyWaxwGSWmQ http://www.youtube.com/watch?v=HTZ8YKgD0MI http://www.youtube.com/watch?v=SyWaxwGSWmQ TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 116 The goals of scientific research The three basic goals of science are description, prediction, and explanation (Kaplan, 1964). The following sections consider how I/O psychologists attempt to achieve each of these goals. Several important terms that are central to scientific research are first defined, followed by descriptions of some of the statistical tools that are used in the practice of science in I/O. Some basic terms I/O psychologists, like all scientists, operate at different levels of abstraction (Kerlinger, 1973). At the most concrete level they observe and record the events of organizational life. For example, if measuring the job satisfaction of an employee, one might give that employee a 1 – 7-point scale and have that person indicate on the scale just how happy they are with the job she occupies. The employee answers with a 6, which on the scale is at the high (satisfied) end of the scale. This is the most concrete level and consists of an actual observation of an event. Scientists seldom are content to stay with only observable data, however, and invariably evoke more abstract notions such as variables, constructs, and theories in their attempt to describe, predict, and explain human behavior in organizations. In our example, the number chosen on a job satisfaction scale is an indicator of a more abstract concept or construct, job satisfaction. A variable is an attribute or property that can assume different numerical values. In some cases, the variables are inherently quantitative in that numbers can be assigned to represent different levels or magnitudes. For instance, the number of correct responses on a vocabulary test indicates different levels of verbal ability or responses to a job satisfaction scale indicate degrees of satisfaction. Other variables such as gender, race, or social class are inherently qualitative in that there is no meaningful way to distinguish among different magnitudes or levels of the variable. One could assign, even to these variables, numerical values for the purposes of research (e.g., male = 1, female = 2), although the values in these cases reflect presence or absence of a property rather than "low" or "high." In other words, as will be discussed later in this chapter, the variable is measured on a nominal scale of measurement. A construct is a particular type of variable that is crucial to the explanation of human behavior in organizations and is inferred from observations of related events. According to Binning and Barrett (1989, p. 479), "Psychological constructs are labels for clusters of co-varying behaviors. In this way, a virtually infinite number of behaviors is reduced to a system of fewer labels, which simplifies and economizes the exchange of information and facilitates the process of discovering behavioral regularities." The reader will encounter a variety of psychological constructs in future chapters, and in each case the construct is associated with a domain of behaviors that the scientist believes are interrelated. Thus, quantitative ability is used to summarize and explain the interrelationships among tests of addition, subtraction, multiplication, and division. Verbal ability is used to summarize and explain interrelationships among tests of reading, spelling, and vocabulary. Job performance is used to summarize and explain TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 117 interrelationships among individual aspects of employee performance (e.g., quantity, quality, creativity). Job satisfaction is used to summarize and explain the interrelationships among expressed attitudes toward supervision, compensation, the work itself, coworkers, and other facets of the work. A theory is a set of constructs and the interrelationships assumed to exist among these constructs (Selltiz, Wrightsman, & Cook, 1976, p. 16). As an example, consider a theory in which employees with high levels of verbal and quantitative ability are hypothesized to perform at higher levels on the job than employees with low levels of these abilities. The theory in this case is defined by the set of constructs (verbal ability, quantitative ability, job performance) and their interrelationships (positive relationships between the abilities and performance). Subsequent chapters will describe many of the theories that I/O psychologists have used to describe behavior in organizations. As will be seen, theory is often the source of the questions that guide research. In the deductive approach, the researcher starts with a theory, deduces from this theory a specific statement of how variables related to the constructs are related, and then gathers data to test the hypothesized relationships. Whether the theory is seen as accurate or not depends on whether the statements deduced from the theory are supported by the data. An opposite strategy of research is the so-called inductive approach. The researcher starts by gathering data and then derives a theory from the observed relationships. However, even in the inductive approach the researcher is likely to have some prior notion of what constructs are important and their possible relationships. Whichever strategy is adopted, the ultimate objective of research in I/O psychology is to arrive at theory that can be used to describe, explain, and predict behavior. For an online tutorial on hypotheses, theories, variables and constructs go to the following links. Realize that these are based on another text...so when they refer to chapter 3, it's not chapter 3 in this text). http://www.youtube.com/watch?v=L4ZOtY9g3o8 http://www.youtube.com/watch?v=6DDMNUJUmpk Statistics Although description, prediction, and explanation are goals that differ in important respects, all three involve gathering data on relationships among variables. Statistics are a vital tool used in summarizing these relationships and in estimating the odds that they reflect something more than mere chance. Obviously it is impossible to cover in this short chapter the entire field of statistics. Nevertheless, it is important to review a few of the basic statistical methods that researchers use. Distributions. Inferential and descriptive statistics are based on assumptions about how the values obtained for a variable are distributed. An examination of how things are distributed in the world reveals an amazing symmetry for almost everything. For example, the seasons change predictably, and the right side of the body looks very similar to the left side. http://www.youtube.com/watch?v=L4ZOtY9g3o8 http://www.youtube.com/watch?v=6DDMNUJUmpk TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 118 Statistical principles are based on the assumption that this regularity holds for many different variables. Take, for instance, weight and height. Some people weigh very little, such as a little baby, and others weigh a lot, such as fat Uncle Henry. Likewise, pygmies are very short and professional basketball players are very tall. Most people, however, fall somewhere between these two extremes in height and weight. If a large number of people were asked their weight and the number of people at different weights were graphed, the frequency distribution would look resemble a bell curve (see figure 3.1). This is called the normal distribution and is the distribution that is assumed for most statistical procedures. As figure 3.1 illustrates, most people fall in the large "hump" in the middle of the distribution with relatively few people found in the tails of the distribution. Figure 3.1: Areas under the Normal Distribution Many of the variables that are the focus of research in I/O psychology (e.g., intelligence test scores) are distributed in this manner, but it is not uncommon for distributions of variables to deviate from normality in the manner illustrated in figure 3.2. For instance, the salaries of employees, the rate at which employees in different organizations leave the firm (turnover), and the number of thefts committed by each employee are more likely to take the form of positively skewed distributions in which there are few high values on the variable and many more low values. On the other hand, the distribution of supervisor TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 119 ratings of employee performance is often negatively skewed in that supervisors give relatively few very low ratings and many more moderate to high ratings. Figure 3.2: Positive and Negative Skews Descriptive statistics. Statisticians have been developed several numerical indices to describe frequency distributions in terms of their central tendency and variability. These include the mean, median, mode, variance, and standard deviation. Correlation is a statistic used to describe the relationships among variables. For some Khan academy short lectures on descriptive statistics go to the following links: https://www.khanacademy.org/math/probability/descriptive- statistics/central_tendency/v/statistics-intro--mean--median-and-mode https://www.khanacademy.org/math/probability/descriptive- statistics/central_tendency/v/exploring-mean-and-median-module https://www.khanacademy.org/math/probability/descriptive- statistics/central_tendency/v/statistics--sample-vs--population-mean 1. Descriptive measures of central tendency. The mean is the arithmetic average of all scores in a frequency distribution, whereas the median is the middle score, indicating the point at which 50% are below that value and 50% are above that value. The mode is simply the most frequently occurring score. The mean, median, and mode are collectively referred to as measures of central tendency. In a normal distribution, the mean, median and mode have the same value. In skewed distributions, the extreme scores pull the mean away from the center of the distribution, and the mean is no longer a good measure of central tendency (see figures 3.3). Consequently, the median is a better measure of central tendency for skewed distributions. The mode is much less frequently used than the median and mean but is a more meaningful description of central tendency in cases where https://www.khanacademy.org/math/probability/descriptive-statistics/central_tendency/v/statistics-intro--mean--median-and-mode https://www.khanacademy.org/math/probability/descriptive-statistics/central_tendency/v/statistics-intro--mean--median-and-mode https://www.khanacademy.org/math/probability/descriptive-statistics/central_tendency/v/exploring-mean-and-median-module https://www.khanacademy.org/math/probability/descriptive-statistics/central_tendency/v/exploring-mean-and-median-module https://www.khanacademy.org/math/probability/descriptive-statistics/central_tendency/v/statistics--sample-vs--population-mean https://www.khanacademy.org/math/probability/descriptive-statistics/central_tendency/v/statistics--sample-vs--population-mean TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 120 the variable is qualitative in nature (e.g., religious preference, race, gender). Now go to the following link in the Rice Virtual Statistics Lab and review a description of the mean, median, and mode. The reader may wish to tackle the exercises within the link. http://onlinestatbook.com/2/summarizing_distributions/measures.html 2. Descriptive measures of variability. Among the numerical indices that are used to describe the spread of values in a distribution are the range, variance, and standard deviation. The simplest of these is the range. This is computed by taking the difference between the smallest value and the largest value in the distribution. More informative is the variance (symbolized as the Greek letter sigma, σ2 (i.e., sigma squared). This measures the extent that values deviate from the mean of the distribution and is calculated using the following formula: where ∑ indicates summation, Xi stands for the values found for the variable in the population of data, N is the number of Xis observed in the population, and µ (symbolized with the Greek letter Mu or μ) is the mean of the Xis in the population. The square root of the variance is the standard deviation (symbolized as σ or the Greek letter sigma) and is a more interpretable measure of variability. Below is an example of the computation of the variance and the standard deviation: Given ten data points in the population: x i - 10 = 5,10,15, 20, 25, 30, 35, 40, 45, 50. Here’s how one would calculate the standard deviation. Step 1: The first step is to compute the mean of the population (µ): Mean of the population = (5+10+15+20+25+30+35+40+45+50) divided by 10 = 275 / 10 = 27.5 Step 2: The deviations of each value from the mean, the squared deviations, and the sum of the squared deviates are calculated. Step 3: The total found in the last column of the table and n= 10 are plugged into the formula for variance of the population, σ2 = ∑(xi− µ) divided by N = 2062.5 / 10 = 206.25 http://onlinestatbook.com/2/summarizing_distributions/measures.html TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 121 And the standard deviation (σ) of the population = square root of 206.25 = 14.36 The examples assume that the ten values constitute the entire population of values. If the ten data points represent a sample from a larger population, the variance and standard deviation are computed as: S2 = ∑(xi−Mean of the sample) divided by (N – 1) = 2062.5 / 9 = 229.17 And SD = square root of 229.17, or 15.1384 Another illustration of variances in several distributions is provided in figure 3.4. Distribution A has less variance around the mean, than B, which, in turn, has less variance than C. For a simple discussion and illustration of the standard deviation take a look at: http://www.youtube.com/watch?v=hUaua15QzK4 For Khan academy lectures on the variance of the population and variance of a sample go to these links: https://www.khanacademy.org/math/probability/descriptive-statistics/old-stats- videos/v/statistics--variance-of-a-population https://www.khanacademy.org/math/probability/descriptive-statistics/old-stats- videos/v/statistics--sample-variance http://www.youtube.com/watch?v=E4HAYd0QnRc http://www.youtube.com/watch?v=hUaua15QzK4 https://www.khanacademy.org/math/probability/descriptive-statistics/old-stats-videos/v/statistics--variance-of-a-population https://www.khanacademy.org/math/probability/descriptive-statistics/old-stats-videos/v/statistics--variance-of-a-population https://www.khanacademy.org/math/probability/descriptive-statistics/old-stats-videos/v/statistics--sample-variance https://www.khanacademy.org/math/probability/descriptive-statistics/old-stats-videos/v/statistics--sample-variance http://www.youtube.com/watch?v=E4HAYd0QnRc TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 122 Figure 3.3: Comparisons of Means, Standard Deviations, and Variance in Two Distributions TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 123 Table 3.1: Hypothetical Data Used in Figure 3.3 Figure 3.4: Comparisons of Three Distributions having the Same Mean but Different Variances around the Mean 3. Computing standard scores. Using the standard deviation, researchers compute standard scores to translate raw scores on a variable into standard deviation units. This allows comparisons across variables having different distributions. One type of standard score is the z score. This is computed by taking the difference between an individual’s score on a variable and the mean score for that variable and then dividing by the standard deviation: For example, assume that a comparison of how Ralph did on the first test in this course to how he did on the second test or how Ralph did on each test compared to Fred. One could simply compare the raw scores on the tests but comparisons of raw scores are difficult to interpret without taking into account the distribution of scores and where each student stands relative to other students in the class. By dividing the difference between each test TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 124 score and the mean of the tests and then dividing by the standard deviation one is converting the raw score into standard deviation units. The Z-score indicates how many standard deviations the student is above or below the mean. The Z-score indicates not only how well the student performed on a test but also allows a determination of the proportion of those taking the test that scored lower and higher than the student. Z-scores also allow a clearer interpretation of differences among scores on tests. Let’s say that on test 1, Ralph made a 90, Fred scored an 84, and the mean for the whole class was 85. On test 2, Ralph made a 96, Fred scored 85 and the mean for the class was 90. So how did Ralph do relative to Fred? And how did Ralph do relative to the mean of the class? It all depends on the standard deviation of the test scores on each test. Assume that the standard deviation on test 1 was 10 and the standard deviation on test 2 was 30. On test 1, the z score for Ralph is +.5 and the z score for Fred is -.10. On test 2, the z score for Ralph is +.2 and the z score for Fred is -.16. The absolute difference between Ralph’s score on test 1 and the class average was only slightly smaller (+5) than the difference between his score on test 2 and the class average (+6). However, Ralph did much better relative to the class average on test 1 than test 2 (.5 vs. .2) when the differences were expressed in standard deviation units. Although the absolute difference in scores between Ralph and Fred on test 1 was smaller (6) than on test 2 (11), the difference in z scores between them on test 1 was actually larger (.6) than on test 2 (.36). Many of the major aptitude and achievement tests (SATs, ACTs, GREs, MCATs, LSATs) are expressed in terms of standardized scores to provide more interpretable scores. For further discussion of z scores and the normal distribution go to: http://www.khanacademy.org/math/probability/statistics- inferential/normal_distribution/v/ck12-org-normal-distribution-problems--z-score 4. Correlation. Researchers are usually interested in describing and drawing inferences for the relationships among two or more variables. First, consider bivariate correlations as described in the bivariate frequency distributions in the figures below. The most frequently used statistic in describing relationships such as those presented in these figures is the bivariate correlation coefficient. Several potential relationships can exist between two variables. Some variables are positively related in that as values for one increase, the values on the other variable also increase in magnitude. An example is the relation between cognitive ability and job performance. As ability increases, performance tends to increase. Other sets of measurements are negatively related to each other. That is, as one set of measurements increases, the other set decreases. Ratings of job-related stress and job satisfaction are usually negatively related. As people report higher stress ratings, they also report lower satisfaction ratings; conversely, as people report lower stress ratings, they also report higher satisfaction ratings. Still other variables have little or no relationship in that there is no systematic tendency for one to increase or decrease with changes in the other. http://www.khanacademy.org/math/probability/statistics-inferential/normal_distribution/v/ck12-org-normal-distribution-problems--z-score http://www.khanacademy.org/math/probability/statistics-inferential/normal_distribution/v/ck12-org-normal-distribution-problems--z-score TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 125 Table 3.2: Hypothetical Data Used in Calculations of Correlations in Figures 3.5, 3.6, and 3.7. To illustrate imagine a study attempting to identify the variables that predict the grade point averages (GPA) of students at a university. The criterion variable in this case is GPA and thinking about this issue leads the researcher to three possible predictors: shoe size, hours spent studying each week, and hours spent partying. Twenty-five students are surveyed about their GPA, shoe size, hours spent partying, and hours spent studying. The data in table 3.2 describe the findings (these are purely hypothetical!). The correlations are calculated between each of the predictor variables (hours partying, hours studying, and shoe size) and GPA. Each of the correlations is reported in figures 3.5 – 3.7. Shoe size has virtually no relation to GPA in the scatterplot in figure 3.5 and the correlation of -.006. The line drawn through the scatterplot is the line that best fits the scatterplot. As is the case whenever the correlation is close to zero, the best prediction of TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 126 GPA from any of the shoe sizes is the overall average of the GPA’s. Shoe size, not surprisingly, is useless as a predictor of grades. Figure 3.5: Example of a Zero Correlation Hours spent partying and hours spent studying are both related to GPA, but the nature of the relation differs for the two variables. Hours spent studying is positively related to GPA with a correlation of .779. As the number of hours spent increases, the GPA also increases. By contrast, hours spent partying is negatively related to GPA with a correlation of -.742. The more hours spent partying, the lower the GPA of the student. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 127 Figure 3.6: Example of a Positive Correlation Figure 3.7: Example of a Negative Correlation To summarize some of the characteristics of a correlation: 1. Statistic that describes the relationship between two variables. 2. Also called the Pearson Product-Moment correlation coefficient. 3. The number that represents a correlation is the correlation coefficient. 4. The correlation coefficient ranges between +1.00 and - 1.00. 5. In a positive correlation, one variable increases as the other increases and decreases as the other decreases. 6. A negative correlation is an inverse relationship in which as one variable increases, the other decreases, and vice versa. 7. A correlation of 0 reflects no relationship between two variables. 8. A correlation does not necessarily indicate causality between the two variables. 9. The correlation coefficient is basically a standard score in which the covariation between two variables is expressed in standard deviation units. 10. The typical correlation found in psychological research (and in other areas of research) is in the -.30s to +.30s. Although far from perfect relationships, correlations of this magnitude can reflect important relationships. For a tutorial on the correlation coefficient go to: http://www.youtube.com/watch?v=BXXtkYOqAfM It is useful to summarize the strength and direction of relationships such as those depicted in the figures with single numbers. The Pearson Product-Moment correlation coefficient (denoted with "r") is such a number and provides both the direction (positive or negative) and magnitude (weak, moderate, or strong) of the relationship between two sets of measurements. This formula for computing this coefficient is found below http://www.youtube.com/watch?v=BXXtkYOqAfM TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 128 where r is the correlation coefficient, ∑ indicates summation, the X and Y stand for the values found for two variables, the X and Y bars are the means of the Xs and Ys, n is the size of the sample taken of the X and Y values, and the Sx is the standard deviation of the x's and Sy is the standard deviation of the y’s for a sample drawn from a larger population. The purpose of a correlation is to determine the direction and strength of the association of two sets of measurements or variables. The coefficient ranges from -1 to +1; these extreme values indicate perfect (very high) negative and positive relationships, respectively. As the value of the coefficient converges toward zero from either direction, the coefficient becomes weaker in magnitude. For example, r = -.30 or +.40 are both weaker in magnitude than r = 1.0 or +1.0. A correlation of zero indicates a complete absence of a relationship. For the variables of interest to I/O psychologists, perfect correlations are rarely found while correlations between +.3 and -.3 are frequently reported. The Pearson Product-Moment coefficient can only assess linear relationships, not nonlinear relationships. An example of a nonlinear relationship is if the Y variable increases with increases in the X variable up to moderate levels of X and then decreases at the higher values of the X variable. In such cases, other computational techniques are used to calculate the degree of nonlinear relationship between the two variables. For a short video that describes the computation of the correlation coefficient check out: http://www.youtube.com/watch?v=pOKWxNTterg In cases where I/O psychologists are interested in describing the relationship between more than two variables they must use multivariate statistical methods. The most common situation that is encountered is where there is interest in the relationship between two or more predictor variables and some criterion variable. For instance, one might want to know the relationship of a battery of two tests with a criterion measure of performance on the job. One could in this case compute a multiple correlation. This is a statistic indicating the combined relationship of the predictors with the criterion. How does one judge the strength of a correlation coefficient? Theoretically a bivariate correlation coefficient can range from +1.00 to -1.00. What size correlation is required to conclude that there is a strong, moderate, or small relation? In a much cited discussion, Cohen (1992) concluded that a small correlation is .10 or lower, a moderate correlation is between .10 and .30, and a correlation of .30 or larger reflects a large effect. but in practice the correlations found in research are typically in the +.40 to -.40 range. There are critics of these criteria (Bosco, Aguinis, Singh, Field & Pierce, 2015; Rosenthal & Rubin, 1982). Rather than imposing absolute criteria for what is large, medium, or small, a more useful and accurate approach is to assess the size of a correlation coefficient in the context of the correlation coefficients typically found in research on a http://www.youtube.com/watch?v=pOKWxNTterg TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 129 topic. Bosco, et al (2015) examined 147,328 correlation coefficients published in two major I/O journals and established benchmarks for assessing the size of correlations across several areas of research in I/O. The correlations reviewed were for the variables listed in table 3.3. Table 3.3: Variables Included in the Bosco, et al (2015) Meta-Analysis The results of the meta-analysis are summarized in table 3.4. Across all variables, the median correlation coefficient was .16. In other words, 50% of the correlations were below .16 and the remaining 50% were.16 or higher. Only 20% were .36 or larger. The size of the correlation varies considerably across topics. In studies examining the correlations between attitudes, a median correlation of .28 was found. The correlation between objective characteristics of people and movement (i.e., turnover, job choice) was only .07. The median correlation of knowledge, skills and abilities (e.g., general mental ability) and performance was .21. As readers progress through the various topics in this text, they will want to refer back to table 3.4 in attempting to assess how the size of the correlations reported compare to what is typically found. One should not necessarily recalibrate a correlation and declare it large just because it is above the median of the correlations reported in table 3.4. One could just as easily conclude that the entire set of effect sizes in these areas of I/O are moderate to small. This is a sad and discouraging conclusion that would cast doubt on the entire field. The authors wisely caution against this conclusion. It is dangerous to assume that small correlations are necessarily unimportant or that large correlations are necessarily important. A small correlation may indicate an effect that while tiny at one point in time can escalate over time. Job dissatisfaction and turnover tends to be only .13 judging by the median in table 3.4. Yet, dissatisfaction with job could put in motion ruminations about the job that eventually have much larger effects on a variety of behaviors including withdrawal. Small effects can serve as a piece of the puzzle and while small in and of itself, can combine with other variables and have much larger effects. The correlation between biases against a group (e.g., minorities) and evaluation of the willingness to hire them is typically small. But in choosing among TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 130 applicants for a job where there are a lot of applicants and very few positions, a very small correlation could yield a stronger bias in the actual choice. The message is that one should not necessarily reject a finding because of a small correlation. Indeed, a very large correlation (e.g., r = .80 or higher) is probably suspect. It probably either reflects an obvious and uninteresting relation or is the consequence of biases such as common method bias. One should not impose on a given finding the label of small, moderate, or large based on an arbitrary set of criteria. Effects need to be evaluated in the context of typical findings on a topic, practical issues, the theory surrounding the effects, and factors associated with an area of research that might prevent an accurate estimation of effects (e.g., the quality of measures that are available). In short, it is a judgment call. After taking into account all these considerations, it is still possible to arrive at the conclusion that there is a weak relation between variables. Sometimes a small effect is just that… a small effect. k = number of correlations; N = median sample size; 20th = 20 percent of correlations are lower than this value; 50th = 50 percent of correlations are lower than this value; 80th = 80 percent of correlations are lower than this value; Table 3.4: Percentile Distribution for Correlations between Different Types of Variables in Bosco et al (2015) Meta-Analysis Inferential statistics. Researchers do not stop at descriptions of samples, but also draw conclusions for the populations from which the samples were drawn. Inferential statistics are used in performing this second task. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 131 1. Testing the significance of a difference. Scientists usually begin their research with an interest in a population of persons and situations. Assume that a researcher hypothesizes that older workers are more productive than younger workers. The researcher draws a sample of workers 30 years or older and another sample of younger workers under 30 years in age. A mean performance of 4.8 is found for the older sample and a mean of 3.0 is found for the younger workers. Can the researcher now conclude that the hypothesis is correct and older workers are indeed better performers? To answer this question one would need to conduct a statistical test of the significance of the difference in the means to allow a determination of whether this difference is statistically significant or if it is attributable to chance. Inferential statistics always involve taking a hypothesis and comparing it to some alternative hypothesis. In this case the alternative hypothesis is that no difference exists in the mean performances of younger and older employees. This alternative is called the null hypothesis. The statistical test indicates the likelihood of finding the observed difference between the samples if the null hypothesis were true (i.e., the means of the populations from which the samples were drawn were actually the same). In concluding whether a difference is large enough to deserve out serious attention, it is common practice to expect that the probability of a difference is less than 5 in a 100 or even lower (less than 1 in a 100). Assume that the mean performance of older employees is larger than the mean physical ability of younger employees and that the probability of finding this difference under the null hypothesis is less than .05 (p < .05). This is equivalent to saying that the chance is less than 5 in a hundred of drawing samples that differ this much if older employees and younger employees actually did not differ in performance. Such a difference is referred to as a statistically significant difference. There is nothing absolute about requiring a probability level of .05 or lower. One could decide to require another probability level, say .10 or .25. Nevertheless, requiring a .05 or lower level reflects the general reluctance of scientists to accept differences as real unless the possibility of being a chance event is low. In addition to testing the significance of differences in means, inferential statistics also are used to test the statistical significance of correlations. For example, one might measure age and performance in a sample and compute the correlation between these two TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 132 variables. If a correlation of .779 is found between hours spent studying and GPA, as indicated in figure 3.6, the next step is to estimate the likelihood of obtaining this level of correlation given that the relationship in the population from which the sample is drawn is some other value. The most common approach is to estimate the likelihood of the correlation obtained in the sample being larger than an assumed population correlation of 0 (i.e., no relation or a null relation). The statistical test indicates the likelihood of obtaining a correlation of the magnitude found in the sample if the correlation in the population is actually 0. It is also possible to compute the statistical significance of differences in correlations computed on different samples. Perhaps the correlation of .779 depicted in figure 3.6 was found for students at university A and a correlation of .500 was found at university B. A test for the statistical significance of the difference between the two correlations could be computed that would indicate the probability of finding the difference observed if the difference was actually 0. 2. Confidence intervals. Another approach to drawing inferences about the size of a correlation or differences between means is the confidence interval. Although statistical hypothesis testing is similar in many respects to confidence intervals, many researchers prefer stating the confidence interval over a conventional statistical test of significance. A confidence interval is a range of scores around a mean score showing the precision of the mean score. The reader has no doubt heard opinion polls on TV in which there is a report of the percentage saying they would vote for each of several candidates as well as a margin of error for the projected vote. In stating a margin of error, the pollster is using confidence intervals. For instance, if a poll states that 48 percent favor candidate A, 40 percent candidate B, and the margin of error is plus or minus 5 percentage points. Assuming that a 95 percent confidence interval is being used here, one can interpret this as meaning that there is a 95 percent probability that the percentage actually favoring candidate A is between 43 and 53. The flip side is that there is a five percent chance that fewer than 43 percent or more than 53 percent would favor candidate A. A similar interval is constructed around the 40% stating they would vote for candidate B shows that there is a 95% chance that between 35% and 45%. The overlap between the confidence intervals for candidates A and B suggest that the election is too close to call. Given the error in the sampling of opinion, there is a degree of uncertainty about whether which candidate will prevail. This inference depends on the error of measurement that is used in constructing the interval. In the example a 95% confidence interval was used. Assume that an 80% confidence interval is used and that the error measurement is plus or minus 2%. In this case there the probability is 80% that between 50 and 46 favor candidate A and that between 38 and 42 percent favor candidate B. Using this confidence interval, there is no overlap in the two intervals. If one is comfortable with this level of error, then one could infer that candidate A will prevail over candidate B in popular opinion. The chances of being wrong in one’s prediction of no difference in the popularity of the two candidates is 5% using the 95% confidence interval and 20% using the 80% confidence interval. It is not entirely correct to state the chance of obtaining a specific percentage of support based on the confidence interval. A more precise statement is that if 100 samples of opinion were taken then the percentage calculated in 95 of these samples would fall within the interval set for the 95% confidence interval and 80 of these samples would fall within the interval set for the 80% confidence interval. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 133 As suggested by the example, a confidence interval indicates the possible range around the estimate and in so doing gives a sense of how much uncertainty there is estimating a measured value. Wide confidence intervals constructed around an estimate leave a lot of uncertainty about the true value of what is measured whereas narrow confidence intervals allow more certainty in the estimate. To construct a confidence interval the researcher first needs to calculate the standard error of the statistic and then multiplying the standard error by a constant. This constant is a value corresponding to a point on the normal curve associated with the level of confidence (or certainty) that one is willing to live with in making inferences about the statistic. Three common confidence intervals are for the 90%, 95% and 99% levels. The constants on the normal distribution are 2.576 for the 99%, 1.96 for the 95% confidence interval, and 1.645 for the 90% confidence intervals. Assuming a normal distribution 99% of the population scores should fall in the interval ranging from +2.576 above the mean of the population and -2.576 below the mean. 95% of the scores should fall between +1.96 above the mean and -1.96 below the mean. 90% of the scores should fall between +1.645 and -1.645. To construct a confidence interval, a researcher takes one of these constants and multiplies it by the standard error of measurement for the statistic that is estimated. The upper end of the interval is then calculated by adding the product to the mean. The lower end of the interval is calculated by subtracting the product from the mean. The z scores taken from the normal distribution are always the same, but the calculation of the standard error of measurement varies from one statistic to another. Readers of this textbook will frequently encounter meta-analyses that report Pearson bivariate correlations. Statistical significance is reported when the confidence interval for the mean correlation (weighted by sample size) excludes the zero point. The confidence interval varies across meta-analyses but typically is at 95% level. To calculate the confidence interval for various statistics, try your hand at these online calculators: For the mean: http://www.danielsoper.com/statcalc3/calc.aspx?id=96 http://www.select-statistics.co.uk/confidence-interval-calculator-population-mean For the proportion: https://www.mccallum-layton.co.uk/tools/statistic-calculators/confidence-interval-for- proportions-calculator/ For the correlation coefficient: http://www.vassarstats.net/rho.html Although this chapter will not go into any more detail on hypothesis testing and inferential statistics, the reader who wishes to know more is advised to answer the exercises at the following link in the Rice University Virtual Statistics Lab. http://onlinestatbook.com/2/logic_of_hypothesis_testing/ch9_exercises.html http://www.danielsoper.com/statcalc3/calc.aspx?id=96 https://www.mccallum-layton.co.uk/tools/statistic-calculators/confidence-interval-for-proportions-calculator/ https://www.mccallum-layton.co.uk/tools/statistic-calculators/confidence-interval-for-proportions-calculator/ http://www.vassarstats.net/rho.html http://onlinestatbook.com/2/logic_of_hypothesis_testing/ch9_exercises.html TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 134 Measurement of Variables One of the goals of science is to describe the variables that are the focus of the research in a way that others can repeat the research. The rules by which a person, object, or event is assigned a value on a variable is referred to as measurement. The specific instruments or devices used to assign the values are referred to as scales or measures. For an online discussion of alternative scales of measurement go to the following link. Realize that the video refers to another textbook...so the specific references to chapter numbers are not from your text: http://www.youtube.com/watch?v=FYcWzC4hxBs Levels of measurement. There are four levels of measurement with scales at each level identified as nominal, ordinal, interval, and ratio. A nominal scale simply names a group of persons based on the attribute and does not reflect a quantity or magnitude of the attribute. The numbers on the jerseys of football players reflect a nominal scale in which individual players are given unique numbers that signify positions. If a player has number 78 and another player has 82, nothing can be said about much more one player has of something than the other. On the other hand, ratio scales do reflect the amount or magnitude of an attribute, and there is an absolute zero point on the scale reflecting the absence of the attribute. Weight is measured on a ratio scale. With this type of scale, individuals can be directly compared on how much heavier one person is compared to the other. Thus, a person who weighs 100 pounds is twice as heavy as a person who weighs 50 pounds. In between the nominal and ratio scales are ordinal and interval scales. Ordinal scales indicate the rank order of persons measured on the scale but do not allow statements about how far apart they are on the scale. The only conclusion that can be made is that one person has more or less of the attribute measured than the other. An interval scale indicates the distance separating the people measured on the attribute. Consequently, a number on the scale allows conclusions about the rank order of the people and gives some indication of how far apart they are on the attribute. Because there is no absolute zero point on the scale, however, an interval scale does not allow conclusions as do ratio scales about how much more one person has of an attribute than another (e.g., has twice as much as the other). An example of an interval scale is a thermometer. If the temperature on one day is 70 degrees Fahrenheit and on another day it is 35 degrees Fahrenheit, it can be said that one day is hotter than the other. It can also be concluded that the difference between these two days in temperature is equivalent to the difference between a 70-degree day and 105- degree day. It cannot be concluded, however, that a 70-degree day is twice as hot as a 35- degree day. The type of statistical analysis that is most appropriate depends on the level of measurement. It does not, for instance, make sense to compute a mean or standard deviation on nominal data. If this still isn’t clear, try out the following video for further discussion: http://www.youtube.com/watch?v=hZxnzfnt5v8 http://www.youtube.com/watch?v=FYcWzC4hxBs http://www.youtube.com/watch?v=hZxnzfnt5v8 TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 135 The reader also may want to go to the following link in the Rice Virtual Statistics lab for additional discussion of levels of measurement and do the exercises associated with the discussions. http://onlinestatbook.com/2/introduction/levels_of_measurement.html http://onlinestatbook.com/2/introduction/measurement_demo.html Reliability of measurement. As discussed previously, a science is precise. Various indicators of precision have been developed to guide scientists in their determination of how precisely attributes are measured on a scale and how much confidence one can place in the data collected with the scale. This is a short clip from a series on astronomy that is very relevant to psychology. As seen here, the theory of error of measurement that is so important to research in I/O psychology and other areas of psychology actually came from the physical sciences. http://www.youtube.com/watch?v=e0bk9BY2nWc The most important indicator of the precision of measurement in the social sciences is reliability. The concept of reliability refers to the extent that a measure is relatively free of random error and is consistent in the numbers assigned to objects or events. There are four approaches to estimating the reliability of a measure: test-retest, internal consistency, parallel or equivalent forms, and inter-rater reliability. The test-retest method focuses on the measure's stability over time. For instance, a researcher could estimate the reliability of a job satisfaction scale by giving the scale to a group of employees and checking to see whether the employees maintain the same relative order on satisfaction at each testing. If on retest, the most satisfied employees are among the least satisfied whereas the least satisfied are among the most satisfied, then the measure is possibly unreliable. Another approach to assessing reliability is to examine the internal consistency of the measure. Often a single psychological scale consists of several components (items, trials, samples, observations, etc.), and these components are combined on the assumption that the values yielded across components are consistent. One might measure job satisfaction, for instance, by taking the mean response to each of ten questions about satisfaction with pay, supervisor, etc. If there were large inconsistencies across items in which high satisfaction with one facet is accompanied by dissatisfaction with another facet, then combining the separate items would yield an unreliable measure. Related to internal consistency, but less commonly used, is the method of parallel or equivalent forms. In this case one constructs equivalent forms of a measure and examines the correlation of scores on the two forms. To the extent that the scores on the two forms are highly related, the measures are reliable. It is crucial in using this method that the two forms are actually equivalent in that they have equal means, variances, and the same correlations with other measures. http://onlinestatbook.com/2/introduction/levels_of_measurement.html http://onlinestatbook.com/2/introduction/measurement_demo.html http://www.youtube.com/watch?v=e0bk9BY2nWc TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 136 The fourth approach to assessing reliability is interrater reliability. For instance, to measure the performance of employees a researcher might ask three supervisors to each rate the performance of the employees and then averaging the three supervisory judgments. A reliable rating of employees on this performance measure is one in which the supervisors are similar in the relative ratings of the three employees. Supervisor 1 gives the highest rating to employee A, the next highest to employee B, and the lowest to employee C. Similarly, supervisor 2 gives the highest rating to employee A, the next highest to employee B, and the lowest to employee C. Note that supervisors might assign different points on the rating scale to the three employees but the ranking of the three employees on the scale is the same. Interrater reliability is the extent of consistency of ratings. Interrater agreement is whether the supervisors assign the same numerical ratings to the employees. One could find that the supervisors give the same relative ratings to the employees but that they use a different portion of the scale and assign different absolute ratings. Although a reliable measure is relatively free of random error, it could still suffer from systematic error. Thus, each of several supervisors might consistently err in a lenient or severe direction in their ratings of the performance of employees, but their ratings could still be reliable in that the ordering of the employees is consistent across the three supervisors. Similarly, in a test-retest situation, it is not uncommon to find that test scores of employees improve as the result of practice. A high degree of reliability is still possible, however, if the employees all improve together and they maintain a similar rank ordering of scores across the two administrations of the test. For further discussion of the concept of reliability go to: http://www.youtube.com/watch?v=kDW0Y5msB0A Validity of measures. The second major basis for evaluating a measure is validity. The validity of a measurement basically refers to whether the number obtained truly reflects what the user intends to measure. If the intent is to measure job satisfaction with a series of questions, do these questions in fact reflect job satisfaction? If the intent is to develop a battery of tests to measure competence to perform the job, do my tests really tap competence? Reliability and validity are related, but it is crucial to understand the differences between the two concepts. A reliable measure provides consistent readings but is not necessarily valid. A palm reader might provide very reliable readings of future job success but the measure in this case is totally lacking in validity as a measure of performance potential. On the other hand, a valid measurement is necessarily reliable. A measure that yields wildly discrepant readings from one time to the next is unlikely to provide what it is intended to provide. In general, reliability is a necessary but not sufficient condition for validity, with reliability setting the upper bound to the level of validity that one can expect to find in a measure. For further discussion of the concept of validity go to: http://ec.libsyn.com/p/3/7/0/3700d5413d1b2fc1/validity2.mp4?d13a76d516d9dec20c3d2 http://www.youtube.com/watch?v=kDW0Y5msB0A http://ec.libsyn.com/p/3/7/0/3700d5413d1b2fc1/validity2.mp4?d13a76d516d9dec20c3d276ce028ed5089ab1ce3dae902ea1d01cd8f37d3c1581b97&c_id=4271019 TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 137 76ce028ed5089ab1ce3dae902ea1d01cd8f37d3c1581b97&c_id=4271019 1. Criterion related validity of measures. Traditionally three types of validity have been presented. Criterion related validity refers to whether a measure allows correct prediction of a criterion variable. For instance, a common concern in I/O psychology is whether tests of abilities predict performance on the job. One way of computing criterion related validity is to use a predictive design in which the ability tests are given to job applicants. The correlation coefficient is later computed to show the relationship of ability scores on the tests of those who are hired and the later job performance of these same employees (often measured by supervisor ratings). Another approach is to use the concurrent design in which the ability tests are administered to a sample of current employees and the correlation of their test scores and their performance evaluations is computed. In both cases, the tests are said to have criterion-related validity to the extent that they allow the prediction of job performance. The correlation coefficient of predictor and criterion is known as the validity coefficient of the test. 2. Content validity of measures. A second approach to validating a measure is to examine its content validity. A measure is content valid insofar as it actually provides a good sample of the domain of behaviors that it is intended to measure. Take, for instance, a typical exam in a course. The test is content valid to the extent that it provides a good sample of material from the lectures, classroom discussions, text, and other materials that students are expected to know. If all the questions were taken from one chapter to the neglect of other chapters, then the test is obviously low in content validity. Likewise, a test of achievement motivation is content valid to the extent that it provides a representative sample of achievement-related behaviors. Unlike criterion-related validation, there is no single quantitative index of content-validity, although there have been attempts to propose such an index (Lawshe, 1975). The evaluation of content validity requires that the domain that is sampled is carefully specified and rests largely on the judgment of experts that the sampling from this domain is unbiased. 3. Construct validity of measures. This is the most general of the various approaches to validity and refers to whether scores on a measure reflect the construct that it is purported to measure. The primary question in criterion-related validity is "Does the measure predict?” The primary question in content validity is "What does the measure contain?” In construct validity, the essential question is "What does the measure really measure?” Construct validation goes hand-in-hand with theory development and testing (Binning & Barrett, 1989). a. Steps in evaluating construct validity. According to Nunnally (1978), the construct validation of a measure requires three steps: *Specify the domain of observables related to the construct. This is similar in many to what was discussed under content validity. If the researcher intends to measure a construct in developing this measure, then he or she needs to define the construct and outline the various observable behaviors that might constitute the domain of the construct. In developing a measure of employee performance, researchers would want to http://ec.libsyn.com/p/3/7/0/3700d5413d1b2fc1/validity2.mp4?d13a76d516d9dec20c3d276ce028ed5089ab1ce3dae902ea1d01cd8f37d3c1581b97&c_id=4271019 TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 138 give careful consideration to the various behaviors that are included in the performance domain. While they might exclude behaviors such as ingratiating the boss, they would want to include other behaviors such as providing good customer service. If the researchers developed a job satisfaction measure, that they might define the domain as including attitudes about pay, fellow employees, the work itself, company policies, the work environment, and the like. In developing a measure of anxiety, they might specify the domain of anxiety related behaviors as including such responses as fear of the dark, worrying about what might happen, anxiety about catching diseases, panic attacks, freezing on tests, etc. * Examine the relations among observables that are in the domain. The next step is to determine whether the many observables specified in step 1 hang together well. For instance, one might have persons rate the extent to which they experience various fears and anxieties specified in step 1. One would then examine the extent to which the correlations among these responses are positive. If persons who state that they frequently have panic attacks and freeze on tests also have anxieties about catching diseases and fear the dark, then the total score on this test is more likely to reflect a single construct of anxiety. Note that this is essentially the same as evaluating reliability of a measure with the internal consistency approach. If responses to the various items are not highly related, then this not only suggests that the measure is not a reliable measure but that it lacks validity as a measure of the construct. *Examine the relations among different constructs. Our theories should dictate how the construct underlying our measure relates to other constructs. Validating the measure becomes a process of empirically verifying that our measure of the construct is related to measures of the other constructs in a manner that is consistent with the theory. From a theory of anxiety researchers might predict that highly anxious persons experience more job stress when faced with high job demands. They might then examine the correlation of the measure of anxiety with measures of job stress under low and high levels of job demands. If the hypothesis is supported by the data, then not only is the theory supported, but the construct validity of the measure is also supported. A basic assumption behind construct validation efforts is that there are good theories and valid measures of the other constructs. For instance, if researchers fail to find support for the above hypothesis, they cannot conclude support for construct validity of the anxiety measure unless they are confident in the theory and the validity of the measures of job demands and stress. From the discussion so far it is apparent that construct validation is a continuing process rather than a one-shot affair. As more hypotheses are tested, and a measure is shown to relate to other valid measures in a manner that is consistent with theory, researchers are more confident that the scale measures the construct that it is purported to measure. Given that the number of hypotheses that can be posed for a construct is infinite, the process is never really completed. A problem that exists in attempting to show construct validity through examining the relationships of measures of various constructs is that the use of the same method of measurement can result in artificially high levels of correlation. Common method bias TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 139 refers to the fact that some degree of correlation is likely among similar methods of measurement, even if the measures tap different unrelated constructs. b. The multiconstruct multimethod matrix (MCMM) approach to construct validity. An important strategy of construct validation that attempts to take into account common method bias is the multiconstruct multimethod matrix (MCMM) approach. When originally proposed, this construct validation strategy was called the multitrait multimethod (MTMM) approach (Campbell & Fiske, 1959). “Trait” essentially referred to any psychological construct, not just traits such as ability or personality. In the interest of greater clarity, this chapter will use construct instead of trait. In determining construct validity with a MCMM strategy one would gather data using the measures that are of primary interest and compute correlations among them. ***correlations among the scores obtained with other measures of the same construct (the same construct/different method matrix – also called the validity diagonal), ***correlations among scores obtained with measures of different constructs that use the same approach (the different construct/same method matrix), ***and correlations among scores on different constructs obtained with different methods (the different construct/different method matrix) Assume that a researcher wants to assess the construct validity of two measures used in assessing the extent to which members of a group effectively work together in their meetings. There are four constructs identified for meeting effectiveness: participation (do members fully participate in the discussions), listening (do members listen to each other), organization (does the organization use an agenda and is it organized in its activities), and conflict management (do members effectively manage disagreements). Two different methods are used to assess these four constructs: self-report (members report on the effectiveness of the group meetings for each construct) and external observations (an observer reports on the effectiveness of the group meetings for each construct). Data are gathered on groups using the two methods of evaluating the four constructs, generating eight scores. The correlations among all eight of the scores are then computed and a multiconstruct multimethod matrix is constructed. Assume that the researcher generates the following matrix of correlations. What can the researcher conclude from the pattern of correlations for the construct validity of the measures? The multiconstruct multimethod approach involves four sets of analyses: ***Compute the correlation between the measure and other established measures using the same method. To do this check to see if the correlation is positive between the method of measuring the construct and a well-established measure using the same method to measure the same trait. For instance, one might show that the self-report measures of group effectiveness are positively related to the observer ratings of group effectiveness for each of the four constructs. In the hypothetical data presented in figure 3.8, there is a statistically significant, positive correlation between the observer and self- TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 140 report measures for each group effectiveness construct (see the same trait/different method matrix or the validity diagonal). This is confirmation of convergent validity, i.e., different methods of the same construct yield similar results. The researcher might want to stop at this point and declare the measures construct valid. That would be a mistake. In addition to convergent validity, the discriminant validity of the measures needs to be evaluated. The assessment of discriminant validity requires three additional steps, described in steps 2, 3, and 4. *** For each construct, compare the correlation in the validity diagonal (i.e., the same construct/different method matrix) to the correlation between that construct and the different methods of measuring the other constructs. The correlation in the validity diagonal should be larger than the correlations found for the construct and the other constructs when the method of measuring the construct differs from the method of measuring the other constructs. In figure 3.8, the correlation obtained between self- ratings of organization and observer ratings of organization is .75. This is larger than the correlations between self-ratings of organization and observer ratings of listening, conflict management, and participation (.15, .17, and .20, respectively). It is also larger than the correlations of self-ratings of organization and observer ratings of listening, conflict management, and participation (.11, .15, and .22, respectively). This is one indication of the discriminant validity of the organization measures. ****For each construct, compare the correlation in the validity diagonal (i.e., the same construct/different method matrix) to the other correlations between that construct and the other constructs when the same method is used to measure them. The correlation in the validity diagonal should be larger than the correlations found for the construct and the other constructs when the method used to measure the variables are the same. In figure 3.8, the .75 correlation in the validity diagonal is larger than the correlations obtained between self-ratings of organization and self-ratings of listening, conflict management, and participation (.13, .16, and .32, respectively). Also, it is larger than the correlations between observer ratings of organization and observer ratings for listening, conflict management, and participation (.15, .17, and .20). *** For each construct, compare the patterns of correlations obtained in every triangle within the correlation matrix. The same pattern should be found between the construct and the other constructs regardless of whether the same or a different method is used to measure them. The data in figure 3.8 support this criterion of discriminant validity. The pattern of correlations of organization with the other three dimensions is the same for all four of the triangles of correlations associated with organization. The correlation between organization and the other constructs is largest for participation, followed by conflict management, and listening. This rank order of the size of the correlation is the same regardless of whether the same method is used to measure organization and the other constructs or different methods are used. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 141 Figure 3.8: Illustration of the Use of the Multiconstruct Multimethod Approach to Assessing Construct Validity Using Fictional Data The conclusion that the measures possess construct validity is warranted if the correlations in the matrix meet the four criteria. In other words, the scores obtained with different methods in measuring the same construct not only converge but there is evidence that the methods provide a basis for discrimination that is not confounded with the methods that are used in measuring the constructs. It is important to note that these are concocted data created for the purpose of providing a clear example. In the made up data in the figure, the evidence supports the construct validity of all the measures. But things are seldom that neat in actual MCMM studies, however. Also, simply eyeballing the correlations is not enough to make sense of the findings. Researchers now use sophisticated statistical methods to pinpoint the amount of variance associated with method and construct and to provide confirmatory tests of the construct validity of measures (Höfling, Schermelleh-Engel & Moosbrugger, 2009; Oort, 2009; Schmitt & Stults, 1986; Woehr, Putka, & Bowler, 2012). For a more in-depth discussion of the MCMM (i.e., MTMM) validation strategy check out this link: http://www.socialresearchmethods.net/kb/mtmmmat.php c. Factor analysis. One question frequently raised in assessing construct validity is whether the items constituting a multi-item scale is measuring one construct or whether the scale is multidimensional and reflects multiple constructs. Two of the statistical procedures used to assess the dimensionality of a measure are factor analysis and principal components analysis. This chapter will not delve into the differences between these approaches or the technical issues, but will instead provide a general overview of some of the features common to both. For the sake of simplicity, the chapter will refer to both as factor analysis. http://www.socialresearchmethods.net/kb/mtmmmat.php TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 142 Factor analysis is a statistical procedure that explores whether a set of observed variables can be reduced to a smaller set of unobserved variables. Assume, for example, a researcher uses a 10 items questionnaire that measures leadership style. The questionnaire is administered to a large sample and based on the responses to the items, the researcher can observe variations among respondents on all ten items. The researcher is curious as to whether each item in the questionnaire must be scored separately or whether it is legitimate to combine item responses to form a smaller set of scores. This is an issue of construct validity because the basic question is whether there are as many underlying factors (also called latent variables and constructs) as there are items or whether there is a smaller number of factors with sets of items serving as indicators of each factor. Factor analysis provides an answer to this question by examining the patterns of responses among the items that constitute the measure. Similar patterns of responses for a set of items provides support for the existence of factors that provide the justification for a simplification in the scoring. In other words, finding that responses to some items are highly intercorrelated is an indication that there is an unobserved variable (i.e., a factor, latent variable, or construct) that accounts for the high intercorrelations. Not only does a factor analysis identify the factors but also provides a statistical indicator of the correlation of the responses to each item in the scale and each factor. This correlation is the factor loading of an item on a factor. One might ask people to indicate their income, education, and occupational status, and subject the responses to a factor analysis to determine whether each question needs to be considered separately or whether the researcher is justified in combining the responses to provide an indication of socioeconomic status. In this example, the underlying factor (i.e., construct or latent variable is socioeconomic status. Finding evidence for a factor does not necessarily mean that researchers should no longer examine the component items. It is often the case that a factor analysis provides support for an underlying factor and at the same time provides support for continued attention to the individual items that are indicators of the factor. A later chapter will review factor analytic research on mental ability testing that reveals a general factor of mental ability that underlies almost all mental ability tests (the so-called “g” factor). At the same time, the factor analytic research shows support for lower level factors in the form of specific mental abilities. The research on personality tests provides little evidence for a general factor of personality but does support the existence of five general factors (also called the Big Five). Underlying each of these five factors are sub-factors that represent clusters of items loading on the higher order factor. For instance, one of the Big Five personality factors is conscientiousness which encompasses sub-factors in the form of orderliness, integrity, and achievement orientation. The research provides support for exploring personality at both the level of the higher order factor (e.g., conscientiousness) as well as the sub-factors (e.g., orderliness). The identification of factors and the items that are considered components of each factor is often a judgment call requiring subjectivity. An example is a study conducted by Breevaart, Bakker, Demerouti and Hetland (2011). They were concerned with the construct of employee engagement which is often defined as a relationship between employees and their work in which they are totally absorbed and enthusiastic in what TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 143 they do and the organization in which they are involved. This is a controversial topic and a later chapter will discuss whether this is a worthwhile construct. Breevaart et al (2011) developed a measure of employee engagement (the State Work Engagement scale) and conducted a factor analysis that explored the dimensionality of responses to the questions. Table 3.5, provides the nine items and the intercorrelations among the items. Try to identify the clusters of correlations that identify the underlying factors. Readers who think that they are all highly intercorrelated and one factor underlies the responses are correct. The researchers in this case concluded that there is justification for concluding that there is an overall factor of worker engagement and that all the items could be combined to form a score on this general factor. Readers who concluded that there are subsets of items that constitute clusters are also correct. Using factor analysis procedures, the authors concluded that items 1, 2, and 3 load on a factor of vigor, items 4, 5, and 6 load on a factor of dedication, and items 7, 8, and 9 load on a factor of absorption. The authors in this study conducted what is called a confirmatory factor analysis. In this procedure the researchers pose alternative hypotheses for the factors and the items loading on each factor and then test the relative fit of the results to each hypothesized factor structure. This is contrasted with an exploratory factor analysis that poses no a priori hypotheses for the factor structure. The authors conclude from their confirmatory analysis that the strongest support was for three factors but that a one factor solution was also justified. 4. Validity is unitary. Although criterion related, content, and construct validity are often discussed as if they were three distinct types of validity, validity is unitary rather than consisting of distinct types (Binning & Barrett, 1989; Guion, 1980; Landy, 1986). One of the three validation approaches may seem most relevant in the attempt to support the validity of the measure, but all three approaches are used in the process of validating the measure. Table 3.5: Correlation Matrix Used in Confirmatory Analysis of Worker Engagement Measure Consider the typical personnel selection situation in which there is interest in inferring from a predictor measure the position of employees on the construct of performance in the job. A measure is chosen that is used to select among applicants (e.g., a test of cognitive ability) on the basis of a hypothesis that the construct underlying the measure TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 144 (i.e., cognitive ability) is related to the performance domain. To validate the predictor measure, data are collected on how employees score on the predictor and how they score on a measure of a criterion of job performance. This criterion-related approach yields a validity coefficient that is used to justify the use of the predictor. Although the criterion- related approach is probably the most important approach in the employee selection context, the hypothesis that scores on the predictor are related to performance is also supported with the content and construct validity approaches as well. Using the construct validity approach, one could assess whether the predictor measure is a good measure of the underlying predictor domain and the criterion measure is a good measure of the underlying criterion domain. The construct validity approach is also used in assessing whether the accumulated evidence on the criterion construct and the predictor construct suggests a positive relationship. For example, one would need to review the research that has assessed the relations of cognitive ability tests to measures relevant to the performance domain. Finally, one could use a content validity approach to determine whether the predictor measure provides a representative sample of the performance domain. While it is not possible to assess with these other approaches the ability of a measure to predict performance, the evidence gathered through content and construct validation is relevant for justifying the validity of the inference that the predictor measure is related to performance domain. 5. The use of self-report measures – threat to validity? In I/O, as well as other areas of psychology, people are studied using observations by others (e.g., the supervisory reports on how much effort a subordinate exerts on the job), self-reports of the people who are studied (e.g., the subordinate reports on his or her own effort on the job), or by means of non-human, objective means (the amount of time spent working is measured with a clock). Observations and self-reports are the most frequently used approaches in I/O and typically employ rating scales in which the respondent identifies a point on a numerical scale that reflects amount, intensity, or other attribute. The use of rating scales to collect data varies in the time frame of the measure and the private vs. external nature of what is rated. With regard to the time frame, reports are retrospective in soliciting past behaviors and feelings (e.g., amount of assistance given to coworkers over the last year, typical level of performance), current in asking for ratings of something occurring at the time of the rating (e.g., feelings at this moment now), and future-oriented in asking for something that will occur in the future (e.g., intentions to leave the organization, expectation of future success). One can also distinguish between ratings that focus on the visible and overt and measures that focus on the private and subjective. Examples of the former are reports of the number of absences from work, theft, specific acts of kindness to fellow employees, number of hours spent working at home on job-related tasks, and aggressive behaviors. Examples of the latter are reports of job satisfaction, feelings of anger at work, ego-involvement in the work performed, and liking for coworkers. With the emergence of the behaviorism school of thought in the early years (see chapter 2), researchers were advised to gather data on observable behavior rather than measuring subjective private events or relying on self- reports. This advice is misguided. Self-reports are frequently used out of necessity in the TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 145 research on human behavior in organizations and at work. Most organizations are unwilling to allow observers to watch employees so as to obtain objective records of their behavior. Even when they are willing to allow the use of observation, there are serious issues of privacy and confidentiality. More importantly, one could argue that some constructs such as job satisfaction, emotion, stress, and many attitudes and beliefs are inherently subjective and require a self-report. Some critics have expressed concern that I/O research is vulnerable to biases and errors due to the dominant use of ratings and especially self-reports. The most obvious problem is the failure in memory associated with asking people to report on events in the distant past (e.g., Andersen & Mikkesen, 2008). Also, self-reports by members of a group about group process may reflect more of group members' personal theories of how effective groups behave than what actually happened in their group (Staw, 1975; Gladstein, 1984). Staw (1975) demonstrated this in an experiment in which groups worked on a task and then were given false feedback that they were in the upper or lower 20% on the task. Those receiving success feedback described their groups as more cohesive, more egalitarian, more open in communications, and less prone to conflict than those told they had failed. Gladstein (1984) reported similar findings from a field setting. Another potential problem in using self-report measures occurs in the translation of what is reported and the rating on the measurement scale. For example, imagine that a rater is asked to report on the frequency with which a person engages in helping of others. The rater is to make this report on a 7-point scale in which 1 is anchored with “Very Little” and 7 is anchored with “A lot”. One rater may consider 2 or 3 episodes of helping to constitute “A lot” whereas another may consider the same number to constitute a level of helping that reflects the middle of the rating scale (i.e., a 3 or 4 rating). When the target of the ratings is subjective (e.g., rate the general helpfulness of the person rated), the translation of what is observed to a specific number on the scale becomes even more problematic. When changes in ratings are used as the dependent variable, translation difficulties can create problems for the measurement of change by confounding three types of change: alpha, beta, and gamma change (Golembiewski, Billingsley, & Yeager, 1976). Alpha change is a shift in the numerical rating that is a real change in the target of the intervention. Beta change is a recalibration of the scale. Gamma change is the redefinition of the construct underlying the scale. As an illustration, assume that employees are participate in a teambuilding program intended to increase the openness of group members in communicating with each other. To evaluate the effectiveness of the program, members of the group are asked to rate on a scale of 1 to 7 how open the group is in its communications, with 1 labeled as "very closed" and 7 every open." Assume that before the team-building intervention the average response was 6. After the intervention, the average response is 4. Should one conclude that the team-building intervention failed and actually caused more closed communications within the group? Perhaps not. The shift from 6 to 4 could reflect beta change in the form of a redefinition of the scale on which your group evaluated openness. Team members might, after going through the team-building intervention, decide that the 6 they gave to openness of communication prior to the training was actually a 2. Before the team-building members perceived the TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 146 group as moderately high on open communications, but now that they have seen the extent to which people can achieve open communication, they perceive the previous state of the group's communications as pretty dismal. A third type of change, gamma change, reflects a shift due to redefinition of the construct underlying the scale. Thus, before the team-building intervention, the group may have conceived of openness in communications as simply providing information when asked. The group was seen as open (a 6 on the scale) in terms of this definition of the construct. However, the team- building intervention leads team members to redefine open communications as including actively giving feedback, expressing emotions, confronting the boss, and taking emotional risks. In reference to this redefined construct, group communications after the intervention are now seen as more open than before the intervention. The implication of beta and gamma change is that measuring change through self-reports is often inaccurate. In our example, the intervention actually succeeded in increasing openness, but the numerical changes in the rating scales suggested a decline in openness. Other potential problems with self-report include biases associated with social desirability (i.e., the tendency to report what appears socially favorable rather what is social unfavorable), context (i.e., the tendency for ratings to be influenced by the other issues, people, and events that are rated at the same time), halo (i.e., the tendency to use the same point on the scale in all ratings and fail to distinguish among dimensions rated), and the tendency to show lenient, negative, or middle or road ratings. These rating problems are explored at length in a later discussion of performance evaluation, but it is important to recognize that they are issues for all types of ratings, not just appraisal ratings. When a researcher relies on only self-reports or other reports in the measurement of the variables that are investigated in a study, common method bias is a potentially serious threat to the interpretation of the findings. Imagine a questionnaire that asks the respondents to report on their own job performance, job satisfaction, and relationships with supervisors. As respondents answer the various questions, they might well tend to use the same part of the response scales. Also, the feelings expressed on one part of the questionnaire (e.g., self-perceived high levels of performance) might spill over into answers given to other parts of the questionnaire (e.g., expressions of job satisfaction and descriptions of good relations with the supervisor). A likely consequence is a high level of correlation among the three measures. To the extent that these biases are at work, it is erroneous to conclude that the correlations found among performance, supervision, and performance indicate the actual relations among these constructs. Some correction is needed to take into account the common method bias associated with the use of self- report to measure all the variables. So what can one conclude from all these criticisms of self-reports? I/O psychologists, and psychologists in other specialty areas, will continue to use self-reports out of necessity despite the potential problems. The main advice is to be aware of the problems and the implications for interpretation of the findings. As readers proceed through this text, they will see many studies in which there is a reliance on ratings and self-reports. Rather than necessarily rejecting the findings, however, readers should raise questions. Was it TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 147 appropriate to use the approach in light of the construct that was measured? Could the research have used an alternative approach? And what can a researcher do to minimize errors and biases in ratings? Meta-analysis This is a good place in the chapter to review an important methodology that readers will encounter throughout this text. Meta-analysis consists of statistical tools for combining the results of independent studies. Industrial and organizational psychologists were among the leaders in developing and promoting these techniques in the 1980s. Prior to this time, those attempting to draw conclusions from the research tended to rely on qualitative reviews in which they presented a box score of how many times a statistically significant effect was observed in the research. No attention was paid typically to variations across studies in the sample sizes employed, the reliability of measures, and other factors that might account for the size of effect observed. For instance, a review of the criterion-related validity of cognitive ability tests might simply enumerate the number of studies reporting statistically significant correlations or compute the mean of the validity coefficients while ignoring the fact that some studies involved samples of only 20 or 30 people whereas other samples involved samples of thousands. These qualitative reviews also tended to ignore the fact that the studies differed in the reliability with which the criteria were measured as well as the extent to which there was a full range of scores on the predictor and criterion measures. A sounder basis for drawing conclusions is to combine the specific effect sizes (e.g., correlations, differences in means) found in the body of research, taking into account variations in sample size, reliability of measures, and restriction in range. Meta-analysis is a method for conducting quantitative reviews that provides a much better estimate of the strength of effects found in a body of research. Although meta-analyses are quite common in the published research, the techniques used are not without controversy. The quality of the conclusions is only as good as the quality of the research that generates the effects that are aggregated. Also, there is continuing debate over when it is appropriate to use some of the corrections for statistical artifacts. Nevertheless, meta-analysis is a powerful method of making sense of the myriad of studies conducted on a variety of topics in I/O psychology. The reader will encounter this methodology throughout the text. Basic steps involved in conducting a meta-analysis There are different methods of meta-analysis but in I/O psychology these basic steps are usually followed: The research (both published and unpublished) is surveyed and the statistics summarizing the relationships found in the tests of hypotheses recorded. The various statistics are converted to a common statistic (i.e., the target statistic) so that the analyst can combine the results, For instance, in a test of the hypothesis that intelligence predicts performance, the correlation coefficient computed between intelligence and performance is drawn from each study. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 148 1. The mean of the target statistic across studies is computed. To account for the varying sample sizes used in different studies, the target statistic is weighted by the sample size. The weighted mean of the target statistic is the uncorrected effect. 2. Sometimes there is a correction for statistical artifacts. These artifacts are factors in the research that can lead to underestimates of the true effect. The most common of the artifacts are unreliability of the measures used in measuring the IV and DV and restriction of range on the IV and DV (e.g., in calculating the correlation between intelligence and performance, perhaps the research tended to only use employees who had been hired based on high intelligence, thus restricting the range of the intelligence scores to the higher scorers). There are formulae for correcting the effect sizes calculated in a meta-analysis for these artifacts that are not discussed here. The effect sizes obtained after applying these formulae are called corrected effect sizes and are usually larger to some degree than the uncorrected effect sizes. This is because the statistical artifacts lead to an underestimate of the underlying true relationship or difference. Artifacts are basically sources of noise that prevent a clear estimation of the signal that is the target of the research. The three most common artifacts that lead to underestimation of the effects are sampling error, unreliability of measures, and restriction in range on measures. 3. Assess the importance of the corrected and uncorrected effect sizes produced in the meta-analysis based on the variation found across studies in these effect sizes. If there is a wide variation in effect sizes across studies, there is less confidence that the effect is important or substantial enough to support our hypotheses that there is a positive relation between the independent and dependent variables. 4. Compare the variation found across studies in uncorrected effect sizes to the variation found across studies in corrected effect sizes. This comparison allows a determination of whether a search for moderators of the effects is warranted. For instance, if corrections for artifacts account for most of the variation in effect sizes, then it is warranted to conclude that there are few or perhaps no moderators of the effects. If considerable variation remains, then the next step is to search for variables that moderate the effects. Corrections for statistical artifacts in meta-analysis The most controversial aspect of meta-analysis is the correction for statistical artifacts in step 3. The three types of artifacts that are most commonly identified are sampling error, unreliability in measures, and restriction of range on measures. Corrections for sampling error. Sampling error has the largest influence on the variation in estimates of effect sizes. Consequently, removing sampling error is an important step in estimating the true effect size. The number of participants involved in individual studies and the number of studies are the two factors contributing to sampling error. The larger the number of participants and the larger the number of studies included in the review, the more likely the estimate of the effects will approximate the true effects. Unexpected events (e. g, layoffs), economic downturns, changes in technology, variations TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 149 in job duties, leadership differences, and many other factors are associated with individual studies and can influence the estimate of the effect. Across a large number of studies, one can assume that these factors are random and lead to underestimates in some studies and overestimates in others. Assuming that they are random, they are likely to balance out. The mean of the estimates of the validity coefficients across studies and the standard deviation of the effects allows the calculation of the extent to which sampling error is at work in the research. Corrections for restriction of range. In some types of research, restriction of range on variables can lead to underestimations of the effect sizes. For instance, in assessing the criterion related validity of a test used as a predictor, the employer might only have data on the criterion for those hired. Thus, only those applicants who score higher on the test and those who perform well enough to keep their jobs are sampled in computing the correlation between the test scores and performance. This restriction in range can lead to lower correlations than would be found if a full range of scores on the predictor and the criterion were available. Fortunately, there are statistical corrections that can be applied to estimate what the effect size would have been without restriction of range. Here are two links to interactive demonstrations of the effects of restriction of range on the correlation between two variables. http://onlinestatbook.com/2/describing_bivariate_data/restriction_demo.html#video http://cnx.org/contents/161d1fe5-0299-4188-852a-6469bb9c1953@4/Restriction-of- Range-Simulatio Corrections for unreliability in measurement of variables. To the extent that the variables used in calculation of the effect sizes are unreliable, the effect size computed in any one study or in an aggregate of studies will underestimate the true effect size. There are formulae that researchers can use to estimate the effect size that would be achieved with higher levels of reliability on the variables studied. These are known as corrections for attenuation. These formulae only tell what might be, and do not correct for lousy measures. Issues in artifact corrections. Of the three corrections, the correction for sampling error is the least controversial. There is no disagreement among scholars about the need to correct for sampling error when aggregating effects from several studies. Moreover, the data needed to make this correction is always almost available. All that is needed is the sample size for each study, the number of studies conducted, and the observed effect sizes. Although in principle most scholars agree that one should also correct for restriction in range and unreliability of measures, there is much more controversy surrounding these corrections. The primary source of controversy is that not all studies report data on the reliability of measures and restriction of range. Consequently, corrections often use estimated values for reliability and restriction of range based on the aggregate of the studies that do report these data. Another source of disagreement is in what type of reliability or restriction of range estimate to use. For instance, some scholars would argue that one should never use the correlation of assessments of different raters as an estimate of reliability in these http://onlinestatbook.com/2/describing_bivariate_data/restriction_demo.html#video http://cnx.org/contents/161d1fe5-0299-4188-852a-6469bb9c1953@4/Restriction-of-Range-Simulatio http://cnx.org/contents/161d1fe5-0299-4188-852a-6469bb9c1953@4/Restriction-of-Range-Simulatio TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 150 corrections (Murphy & DeShon, 2000). The correction for sampling error is the most important of the corrections and typically the increases in validity coefficients that occur after application of corrections for unreliability and restriction in range are small. Perhaps the best advice is to correct for all three when the reliability and restriction of range are known in each individual study. When they are not known and it is necessary to rely on estimates of reliability and restriction of range, approach these corrections with caution and try out different approaches. Most importantly, consumers of the research should always exercise a healthy skepticism when corrections for unreliability and restriction lead to large increases in the estimates of effect sizes relative to the uncorrected validities. Terms to keep in mind. This text will not go into all the controversies surrounding the use of meta-analysis. However, throughout this text, the results of meta-analyses are summarized. When correlation coefficients are aggregated in the meta-analysis, the summaries will report the following: k: this represents the number of effect sizes (e.g., correlations, standardized differences between means) summarized in the meta-analysis. An issue in meta-analysis is whether multiple effect sizes in single studies should be included as separate effects in the meta- analysis. In the meta-analyses summarized in this text, no attempt is made to distinguish how many independent effect sizes are represented in the total number of effect sizes. The reader should have less confidence in findings the smaller number of effects included in the meta-analysis. For instance, a meta-analysis reviewing 10 or fewer effects provides a much less confident basis for conclusions than a meta-analysis reviewing 30 to 50 effects. N: this is the total number of participants in the research reviewed. As in the case of k, the larger the number of participants sampled across studies included in the meta- analysis, the more confidence the reader can have in the results. ruc: this is an uncorrected mean correlation coefficient weighted by sample size. It is computed by multiplying each correlation observed by the size of the sample used in computing the correlation, summing these products, and then dividing by the total number of participants across all samples in the research. Thus, if one sample of 30 yields a correlation of .15, a second sample of 100 yields a correlation of .40, and a third sample of 50 yields a sample of .30, the computation of the weighted mean correlation is (.15 X 30) + (.40 X 100) + (.30 X 50) = 4.5 + 40 + 15 = 59.5. One would divide this amount by 180 (30 + 100 + 50) to yield a weighted mean correlation of .33. The ruc is not corrected for statistical artifacts. rc: this is the correlation corrected for statistical artifacts. Some meta-analyses are corrected for all three of the statistical artifacts (sampling error, unreliability, restriction of range) whereas others correct for only one or two of these artifacts. The summaries of the results reported in this text do not distinguish among studies on the basis of the specific correction that was used. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 151 *: In this text the * is used to signify those mean effects for which a confidence or credibility interval computed in the meta-analysis included the mean effect and excluded zero. In most cases this is determined by constructing a 95% confidence interval around the uncorrected weight mean correlation. If 0 is included in the interval, the effect is considered relatively unimportant. If 0 is not included in the interval, the effect is considered relatively important. Another way of stating this is that when 0 is included in the confidence interval it is statistically nonsignificant and when 0 is excluded, it statistically significant. The reader should observe extreme caution in interpreting what are identified as the significant and nonsignificant effects. As noted earlier in this chapter, the computation of statistical significance is considered by a growing number of researchers as an antiquated practice that should be replaced with other approaches such as Bayesian statistics. Statistical significance is not reported for rc. The corrected r purportedly represents the “true” relation. Consequently, sample statistics do not apply. Designing Research for Explanation Developing good measures is a crucial objective in scientific research, but it is only a means to the ultimate ends of explanation and prediction. An important goal in applying scientific method to research on human behavior is to explain the relationships that are found among variables. For any single relationship (e.g., participative leaders have employees with higher productivity than nonparticipative leaders), alternative explanations are almost always possible. It is not possible to eliminate all of the possibilities, but through properly conducted research, the number of alternative explanations can be dramatically reduced. A research design is essentially a plan for how to treat variables that can influence results so as to rule out alternative interpretations. In the design of a study, researchers can hold constant, eliminate, manipulate, randomize, match, or measure variables. To the extent that a research design does not incorporate one or more of these strategies, the differences observed are subject to error in interpretation. In investigating a phenomenon, the researcher is confronted with a huge number of possible variables and interrelationships. The art of designing research is largely a matter of deciding how to treat the variables that can potentially influence the phenomenon under investigation so as to rule out alternative interpretations of the relations observed. For an interesting podcast covering in broad outline research design go to: http://ec.libsyn.com/p/8/8/4/884710c5673b4dac/TPF_045_BasicResearch_020508.m4v? d13a76d516d9dec20c3d276ce028ed5089ab1ce3dae902ea1d01cd8f35d6c05bca98&c_id= 1316583 Hold constant or eliminate variables Some variables are controlled by either holding the values of the variables constant or eliminating them. In studying leader participation and employee satisfaction the researcher might hold constant the task on which the employees are working or the sex and age of the employees so as to obtain a clearer look at the relationship of primary concern. Similarly, in studying the effects of noise on performance the researcher would http://ec.libsyn.com/p/8/8/4/884710c5673b4dac/TPF_045_BasicResearch_020508.m4v?d13a76d516d9dec20c3d276ce028ed5089ab1ce3dae902ea1d01cd8f35d6c05bca98&c_id=1316583 http://ec.libsyn.com/p/8/8/4/884710c5673b4dac/TPF_045_BasicResearch_020508.m4v?d13a76d516d9dec20c3d276ce028ed5089ab1ce3dae902ea1d01cd8f35d6c05bca98&c_id=1316583 http://ec.libsyn.com/p/8/8/4/884710c5673b4dac/TPF_045_BasicResearch_020508.m4v?d13a76d516d9dec20c3d276ce028ed5089ab1ce3dae902ea1d01cd8f35d6c05bca98&c_id=1316583 TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 152 want to conduct the experiment in a setting where it is possible to control the level of noise while eliminating or holding constant other factors that could influence the results. Manipulate the variable Other variables are manipulated so that they assume certain values. In research on leadership and employee satisfaction, some researchers have arranged things so that some employees are assigned a participatory leader and others are assigned a nonparticipatory leader. The variable that is manipulated in a study to determine their effect on some other variable is called an independent variable (IV) because it is independent of all other potential influences in the situation. The best way to determine what is causing what is to directly manipulate the variables that are identified as the causal factors. Of course, some variables are not amenable to manipulation, as will be discussed later. Measure the variable Researchers can measure variables as they occur without attempting to directly manipulate them. For example, if they were conducting a study of leadership, they might want to observe leaders and measure the type of leadership they exhibit as it naturally occurs. The variable that is hypothesized to cause changes in other variables is called the independent variables. The variables that the independent variable is hypothesized to influence are called dependent variables because the underlying theory hypothesizes that the values of the measured variables depend on the values of the independent variable. The readers should note that some psychologists would restricted the use of the terms independent and dependent variable to experiments where the independent variable is manipulated. This text will take a broader approach and use the terms to describe variables in non-experimental research as well. The crucial consideration is whether there is a theory identifying some variables as causes and others as outcomes. Randomly assign research participants to conditions Random is sometimes associated with something "bad" or arbitrary, but in research it is the most powerful means of dealing with all the unmeasurable, unmanipulated, and uncontrollable variables that are potential sources of error (see figure 3.9). TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 153 Figure 3.9: Controlling for Variables with Random Assignment By random it is simply meant that various events have an equal chance of occurring. If interested in comparing employees under participative and nonparticipative leaders, one could allow each employee to pick the leader he or she preferred and then compare the performance of those who picked a participative leader with those who picked a nonparticipative leader. Unfortunately, this is nonrandom in the sense that there is an unequal chance of each employee working with one or the other leader. The employees who strongly prefer making their own decisions or who have prior experience on the task might prefer and choose, for instance, the participative leader, whereas those with low need to make their own decisions or who have no prior experience might prefer and choose the nonparticipative leader. To ensure that each employee has an equal chance of working with each type of leader, one would want to use a randomization procedure such as flipping an unbiased coin for each employee (heads, he works for the autocratic and tails, he works for the democratic leader). An even more sophisticated approach is to use a table of random numbers to assign employees to each of conditions. Randomization protects against systematic errors confounding the comparison of two or more groups, but random error will still exist within each of the groups compared. For example, in an experiment one might randomly assign some employees to participatory leaders and others to nonparticipatory leaders. To the extent that the samples are large, this random assignment to conditions will equate the two groups on age and experience of the employees. However, employees within each group may differ considerably among themselves on both age and experience. To explore further the distinctions between random selection and random assignment go to the following tutorial: http://www.randomizer.org/tutorial.htm http://www.randomizer.org/tutorial.htm TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 154 Match on the variable A possible means of reducing the within group variation mentioned above is to measure the variable and then make sure that matching sets of persons are selected as subjects. In an experiment in which subjects are exposed to either a participatory or nonparticipatory leader, one could form pairs of employees who are matched on a set of variables (e.g., experience and age). For each pair, one of the employees is then assigned on a random basis to either a participative or nonparticipative leader. The visual below illustrates a common practice in research in which an experimental group is compared to a control group and differences that might exist between the experimental and control groups prior to the experiment are controlled. Let’s assume that the “treatment” in this case is a participatory leader who is assigned to the subjects who are in the participation condition and the control group is an authoritarian leader who is assigned to the subjects assigned to the control group. Also assume that the hypothesis is that subjects given the participatory leader are more satisfied with the tasks they perform than subjects given the authoritarian leader. The problem is that there is uncertainty about whether there were differences between the participants in task satisfaction prior to the research. If the hypothesis is confirmed, perhaps it is the result of the fact that happier people are in the participatory leader group than in the control group where there is an authoritarian leader. If the sample of participants is large enough and participants are randomly assigned to each group, there is much more confidence in concluding that there are no pre-test biases of this sort. To make sure, the researcher could use a matched subject design (see figure 3.10) in which participants are given a satisfaction measure prior to the assignment of the leaders. The researcher assigns subjects to the conditions so that the two groups are the same on level of satisfaction. For instance, the researcher could make sure an equal number of those responding to each point on the satisfaction scale are assigned to the treatment and control groups. Different Types of Research I/O psychologists use more than one type of research in testing hypotheses about behavior at work. The various approaches to research are distinguishable on the basis of how the independent variables are treated (are they measured or manipulated), whether participants are randomly or non-randomly assigned to conditions, whether it is obvious to participants that they are in a research study or not, and the setting in which the research is conducted. This section distinguishes among research studies that are experimental vs nonexperimental, lab vs. field, and obtrusive vs. unobtrusive. Experimental vs. nonexperimental The essential attribute of an experiment is that participants are assigned at random to various levels of the independent variables. In some cases, an experiment involves an independent variable that is under the control of the experimenter. In other cases, natural experiments are conducted in which the independent variable is a naturally occurring event rather than a manipulated variable. A researcher who knows that an organization TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 155 plans to change over to a new technology, for example, might persuade the organization to compare the effects on morale and productivity of those given the new technology and those who are not given the technology. As long as there is random assignment of employees to the new technology and the old technology groups, this would qualify as an experiment. Natural experiments in the field of I/O psychology are relatively rare. In most of the experiments discussed in this text, the experimenter not only randomly assigned participants to various conditions of the experiment but also manipulated the independent variable. In non-experimental research the persons studied are assigned on a nonrandom basis to conditions. One variety of nonexperimental research is the correlational study. Here the variables are measured as they naturally occur and the researchers have no direct control over what happens with either variable. Theory or commonsense may suggest that one variable causes the other. Consequently, one variable is often considered to be the independent variable (the X variable) and the other the dependent variable (the Y variable). Despite the identification of independent and dependent variables one cannot conclude from correlational research that there is a causal relation between them There are alternative explanations for a correlation: 1. The X could cause the Y. 2. The Y could cause the X. 3. A third variable (a Z) might cause both X and Y without any direct causal relation between the X and Y. These third variables are often identified as confounding variables in that they can lead to erroneous conclusions. For example, theory may suggest that leader participativeness causes employee productivity. In a correlational study, researchers could measure the participativeness of a leader and then see how it relates to the productivity of employees without any direct intervention to change the leader's behavior. Finding a positive correlation between the two does not allow one to conclude that participativeness is a cause of higher productivity. Although it is possible that participativeness causes higher productivity, it is also possible that higher productivity causes the leader to allow more participation. It is also possible that there are other variables that cause both participativeness and productivity and are responsible for the positive correlation. For instance, perhaps the organization tends to assign more participative leaders to groups of employees with higher productivity and less participative leaders to those groups that have lower productivity. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 156 Figure 3.10: Controlling for Extraneous Variables with Matching It is important to distinguish between correlational analyses and correlational research designs. A correlation coefficient is a description of the relation of the independent and dependent variables. It is a type of statistical analysis that could be used in both correlational and experimental studies to describe the relation found between independent and dependent variables. Take, for example, the correlational research design in which the participativeness of leaders and employee productivity are measured and the naturally occurring relation between the two is observed. One could conduct a correlational analysis to describe the relation. One could take another approach such as identifying some leaders as highly participative and others as less participative, computing the mean productivity of employees having each type of leader, and then testing the statistical significance of the difference between the means. In either type of analysis, the researcher is unable to draw causal inferences because of the correlational nature of the research design. Also consider an experimental design in which the participativeness of the leader is manipulated and employees are randomly assigned to work for a leader who is either high or low on participativeness. In describing the results of this experiment one could compute a correlation between independent and dependent variables. The researcher could assign a value of 1 to low participation leaders and a value of 2 to the high participation leaders and then compute the correlation between the independent and TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 157 dependent variables. A correlation coefficient would describe the relation and because it is an experimental design, one could also draw inferences for the causal relation between the independent and dependent variables. A quasi-experiment is somewhere between a correlational study and a true experiment and could involve an independent variable that is either manipulated or naturally occurring (Cook & Campbell, 1979). The one important difference between a quasi- experiment and a true experiment is that the assignment of participants is nonrandom in the former and random in the latter. Evaluations of training programs often must rely on nonrandom assignment procedures. Employees participating in a program may volunteer for this program whereas those in the no-training control are chosen from among those who do not volunteer. This type of quasi-experiment is called a nonequivalent control group study, and despite the lack of random assignment, can still provide valuable information on the trained and non-trained groups. Procedures exist that investigators can use to statistically control for differences between the two groups. Also, through measuring the dependent variable (e.g., measure of training effectiveness) many times (in other words using repeated or multiple measures of the dependent variable over time) and examining the trends before and after the training intervention, the researcher can eliminate some of the alternative interpretations of differences between the trained and untrained groups. If one simply conducted a one trial pre-test and a one-trial post-test without a control group and found a big increase in productivity from pre-test to post-test, this increase could simply reflect the fact that productivity was rising over time and continued to increase after the intervention. The upward trend occurring before the intervention argues against concluding that the intervention caused the increases observed after the intervention. Taking multiple measures of productivity and examining the trend over time could allow an elimination of this alternative explanation. Laboratory vs. field research In addition to designing research so as to eliminate alternative explanations, one must decide where to conduct the research. The research setting is another important basis for distinguishing among varieties of research. Laboratory research is conducted in settings created for the purpose of research. Laboratories are usually designed for the explicit purpose of controlling and eliminating extraneous variables and allowing an uncontaminated manipulation of the independent variables. In field research the settings are for some non-research purpose such as work or education. If researchers enter an organization and have employees complete a survey in their work places on their attitudes toward their supervisors, this would qualify as a field study. If they had workers come to a special room set aside for the purposes of research outside the work setting and respond to the same questionnaire, the research is conducted in a laboratory and is a laboratory study. Obtrusive vs. unobtrusive research TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 158 Some research is done in a manner that makes it obvious to participants that they are the objects of study. Research where awareness of being a subject of the research is high is called obtrusive, whereas research where awareness is relatively low is unobtrusive. The setting of the research determines to a large extent its obtrusiveness. Almost all laboratory research is obtrusive, but field research varies considerably along this dimension. The manner in which variables are measured is another important factor. Self-reports in which participants in the research are asked to provide introspective accounts of their satisfaction are highly obtrusive whereas using hidden observers who record signs of satisfaction and dissatisfaction among the rank-and-file is relatively unobtrusive. Finally, the method by which variables are manipulated is still another source of obtrusiveness. Typically, laboratory experimenters try to create as much impact as possible in as covert a way as possible to avoid the suspicion associated with a highly obvious manipulation. Factors Influencing the Validity of Explanations The discussion so far should make clear that there are a variety of approaches one can use in conducting research. Although all are scientific, they each have their strengths and drawbacks. How does one evaluate the merits of an investigation? Cook and Campbell (1979) set forth three primary means of evaluating research design: ***Internal validity. The basic issue here is whether variations in X variable cause the variations observed in the Y variable? To the extent that a research design allows one to infer from findings that X causes Y, the design is internally valid. ***Construct validity: The basic issue here is whether the variables measured and manipulated are what the researchers think they are. In other words, do the measure or manipulation of the independent variable (the X variable) and the measure of the dependent variable (the Y variable) really reflect the underlying construct identified in the hypothesis? *** External validity: The issue here is whether the relation found between the independent and dependent variable with one set of participants, settings, or procedures are found with other participants, settings, or procedures? The following link provides a discussion of some threats to validity. http://www.youtube.com/watch?v=zIpDPrbRiBo Another link that provides a lecture on these three types of validity in evaluating research design as well as a fourth type, statistical conclusion validity. http://www.youtube.com/watch?v=FEhSFdjwiv8 Research in I/O often is concerned with whether one variable causes another. A study is internally valid if one can conclude from an observed relationship between an independent variable X and a dependent variable Y, that X causes Y. Nonexperimental http://www.youtube.com/watch?v=zIpDPrbRiBo http://www.youtube.com/watch?v=FEhSFdjwiv8 TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 159 research is a good starting point, but if substantial knowledge about causal relationships is desired it is necessary to conduct experiments. Here are some of the major threats to the internal validity of nonexperimental research and experimental designs that can help eliminate these threats as alternative explanations. Correlational findings are usually subject to multiple interpretations. A fairly strong correlation exists between the act of getting married and the occurrence of pregnancy. Obviously, it is incorrect to conclude that the marriage ceremony itself caused pregnancy. The causal factor is a third variable - sexual intercourse. Although correlational research is probably the least powerful approach to explanation, a correlational study can set the stage for more rigorous explorations of causality. Also, even though one cannot use a correlational finding as definitive proof of a causal relationship exists between two variables, commonsense can allow the elimination of some alternatives. Finding that age of employees is positively correlated with performance on the job does not justify concluding that age actually caused performance. It is also safe to conclude that performance did not cause age! For further discussion of correlation and causality go to the Khan Academy lecture at the following link: https://www.khanacademy.org/math/probability/regression/regression- correlation/v/correlation-and-causality Threats to the internal validity of a research design Another nonexperiment that is frequently found in organizations is the simple pretest- posttest design. The dependent variable is measured prior to the manipulation of the independent variable and again after the manipulation. The effects of the manipulation are then evaluated in terms of changes in the dependent measure. Suppose, for example, researchers evaluate the effects of a new incentive program in which employees are rewarded individually for their performance on the productivity of the employees. In a pretest-posttest design, they measure the productivity of the employees prior to the changeover to the new incentive program and then again afterwards. What if they found an increase in productivity following the introduction of the incentives? Should they conclude that increases in productivity were the result of the incentives? This type of simple design, although quite common, is vulnerable to all sorts of threats to internal validity. Here are some of the more common ones: 1. History: These are events occurring at the time of the manipulation or measurement of the variables that could account for the relation of the independent and dependent variables. Take, for example, a quasi-experimental design in which the productivity of a sample of employees is measured prior to a training program and again after the training program. How does one know that the changes in the posttest or performance were not caused by something that happened to employees in between the time the pretest and posttest? For example, employees might have heard a rumor of an impending layoff if the company did not improve its profit position. As a consequence of fear of losing their https://www.khanacademy.org/math/probability/regression/regression-correlation/v/correlation-and-causality https://www.khanacademy.org/math/probability/regression/regression-correlation/v/correlation-and-causality TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 160 jobs, employees might have improved their productivity. It is often the case that many things are going on at the same time of major interventions in organizations, and these other things can lessen the confidence in concluding that the independent variable caused the dependent measure. 2. Maturation. Still another factor that can create problems are the growth and learning that can occur in between the pretest and posttest. If an incentive program is imposed on workers who are in the process of learning a new job, then gains could be expected regardless of the incentives. 3. Testing. When persons are pretested in an obtrusive manner, changes observed at the posttest could reflect the pretest more than the manipulation of the independent variable. 4. Instrumentation. This threat to internal validity results from changes in the measurement process. What if, for instance, productivity was tracked more carefully after the intervention than before? Increases in productivity might simply reflect the careful attention now being given to productivity than the actual implementation of the incentive program. Other examples of changes resulting from instrumentation are observers becoming fatigued and mechanical instruments losing calibration. One can deal with the above threats to internal validity by providing a control group that does not receive the manipulation of the independent variable. In other words, one could use a pretest-posttest control group design. In the above example, this might mean that the experimental group receives the incentive program while a control group does not. The crucial issue is how to determine who is assigned to the experimental group and the control. If the assignment was something other than random then the following factors could still threaten internal validity. 5. Selection bias. This is a common threat to validity when participants are assigned to the control group on a nonrandom basis. If employees were allowed to volunteer for the new incentive program any differences between the experimental and control groups on the dependent measure could simply reflect the fact that the employees who volunteered differed from those who did not. Employees who were committed to the organization and involved in their jobs may have been more likely to volunteer whereas those employees who were low in these respects did not. Improvements in productivity observed for the incentive program could result from the greater motivation of the employees who volunteered, rather than the specific manipulation. 6. Mortality. The term "mortality" as a threat to internal validity was taken from animal research where subjects literally died in between the pretest and posttest. In most organizational research, it is unlikely that enough employees die to influence the results, but it is not uncommon for employees to show differential dropout rates during the course of an experiment. Consequently, differences observed on the dependent measure could reflect who is left more than it reflects the effects of the independent variable. 7. Awareness of being a research participant. In a complex organization, experiments are difficult to carry off because participants are often aware that some employees are TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 161 receiving something and others are not. This awareness can evoke feelings and perceptions that affect the results. The control group might feel neglected and consequently show a decline in productivity. Conversely, the group receiving an intervention might feel special and demonstrate an increase in productivity as a result of these feelings. Differences can also result from groups feeling some degree of rivalry. A group that sees itself as the control group that is to receive "brand X" could compensate for this stigma by increasing performance. In all these cases awareness of being in a study is the causal factor rather than the variables manipulated in the experiment. This is basically what was called the Hawthorne effect in the previous chapter. The surest way to protect against most of the threats to internal validity is to randomly assign persons to the experimental and control groups. One can have even more confidence in concluding that the independent variable caused observed differences on the dependent variable if, in addition to random assignment, pretests and posttests are taken at multiple points in time. The problems that arise when participants know they are the objects of research are not easily avoided through random assignment or multiple measurements. Indeed, some critics of traditional research methods (e.g., Argyris, 1968) have argued that this awareness of being a subject is a fatal flaw that invalidates much of the psychological research in the field and the lab. The solution according to Chris Argyris is to allow subjects to participate fully in the research rather than keeping them ignorant of the research design. Most psychologists, including I/O psychologists, have been reluctant to discard rigorous research and have seen the alternatives to randomization and control as introducing even more serious problems. For a lecture on internal validity and the threats to internal validity, take a look at the following link: http://www.youtube.com/watch?v=F7kjR30tEAc Internal validity is concerned with whether the operationalization of the independent variable is actually the cause of the changes observed on the dependent variable. The construct validity of the independent variables in an experiment, on the other hand, refers to whether the manipulation of the independent variable actually reflects the underlying construct that it is intended to reflect. Take, as an example, an investigator who attempts to examine the effects of anxiety on motivation by manipulating anxiety through verbal abuse by a supervisor. The research measures motivation by examining the quality of the subject's performance on a task. One can question the construct validity of both the independent and dependent variables in this example (not to mention the ethics of the research). Does the act of verbally abusing subjects arouse anxiety or other emotions such as hostility or frustration? Likewise, do differences in quality of performance reflect differences in motivation or differences in skill, knowledge, or ability? Threats to the construct validity of a research design Randomization and control groups cannot protect against threats to the construct validity of the independent variable. The only recourse is to repeat the experiment with other manipulations and measures of the variables to see if the same results are found. If the http://www.youtube.com/watch?v=F7kjR30tEAc TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 162 independent variable is anxiety, one might manipulate anxiety by conducting one experiment in which research participants are threatened with physical pain in the form of electric shocks, another experiment in which stressful movies are shown, and still another in which subjects are threatened with failure on a task. This program of research would be unethical and unlikely to be approved by most institutional research boards, but it would all some assessment of construct validity. If the same results were obtained from all three experiments, then there is more confidence that the manipulations were actually tapping the anxiety construct. Self-reports of motivation, observing how much effort is exerted, and supervisor reports of motivation are all potential measures of motivation in addition to quality of performance. Once again, if similar results emerge across different indicators of motivation, the researcher is on stronger ground in concluding that it was motivation that was affected by the manipulation of anxiety. For further discussion of threats to construct validity see: http://www.youtube.com/watch?v=O4vIsTxpVgc Threats to external validity of a research design If a relationship is found between two variables in a study then an additional issue is whether the finding is capable of being repeated with different settings, subjects, and procedures. A continuing debate has been whether findings in the laboratory with college students can generalize to actual work settings (Dipboye, 1990). Criticisms of research in lab settings. Critics of laboratory research claim that the artificial nature of the laboratory setting makes participants acutely aware that they are subjects in research. This awareness, in turn, can cause them to behave in ways that are limited to the laboratory. For instance, subjects in laboratory experiments often try very hard to do what they think the experimenter wants them to do. The power of the laboratory was made very clear to the author in an experiment in which he participated as an undergraduate. The experiment required that he swallow a sensing device on a string so that the experimenter could measure the acidity level in his stomach. Electric shocks were administered to his arms and legs when he failed to respond quickly enough to a randomly occurring signal. In addition to all this, he had to drink a glass of bicarbonate soda every 30 minutes and was not allowed to go to the bathroom. The author submitted to all of this for over eight hours. Some critics would argue that the passive acceptance of experimental demands shown by the author, and frequently observed among subjects in laboratory experiments, is unrepresentative of how people behave in the real world. Another criticism of laboratory research is that it provides an unrepresentative sampling of stimuli, settings, and subjects. While the manipulated variables in a lab experiment http://www.youtube.com/watch?v=O4vIsTxpVgc TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 163 were the only stimuli presented, events in the real world usually must compete for our attention. One consequence of this is that the effects observed in the laboratory are larger than typically found in the external world. Some researchers have tested the hypothesis that performance appraisals are biased against women by giving subjects an essay that is described as coming from either a man or a woman. Some researchers have shown that the same essay is rated as poorer if the author is described as a woman than if the author is described as a man. Is bias against women in a lab situation such as this likely to generalize to performance appraisal situations in organizations? Critics of this type of laboratory research have argued that in the messy world of the organization, those evaluating performance are presented with lots of other factors in addition to the sex of the ratee. In a lab experiment such as the one described here, the manipulation of sex is so obvious that it is hard to miss. In a similar vein, Murphy and Balzer (1986) have speculated that effects found in laboratory experiments are stronger than the effects found in organizations because raters in the lab do not have as difficult a time distinguishing the "signal" from the "noise" as they do in the real world. The biggest complaint against the external validity of laboratory research is that the subjects are usually college students. Those readers who are college undergraduates may believe that students are representative of "normal" humans. But critics such as Sears (1986) have expressed their doubts. He claims that the typical American undergraduate is less likely than the general public to have a fully formulated sense of self, but more likely to show large shifts in self-esteem, feelings of insecurity and depression, egocentricism, and need for peer approval. College students, according to Sears, are not even representative of other young people in that they have higher cognitive abilities, but are more compliant to authority and have less stable peer relationships. The result of using students as subjects, according to Sears, has resulted in a view of human behavior as '......lone, bland compliant wimps who specialize in paper-and-pencil tests' (p. 527). Consistent with this claim, Gordon, Slade and Schmitt (1986) found 12 studies that appeared to show major differences between students and non-students and concluded that researchers should not use college students if the intention is to draw conclusions about behavior in work organizations. In defense of lab research. With all these complaints about the validity of the laboratory as a setting for research one would think that laboratory research would cease. IO psychologists continue to go to the laboratory for the reason that they cannot study some of the most important phenomena in IO psychology in the field and must go to the lab. For example, it is usually difficult and often impossible for IO psychologists to go into an organization and manipulate the leadership styles of supervisors. Consequently, laboratory experiments are needed to provide controlled investigations of leadership dynamics. Despite the need for laboratory research, can one conclude anything about "the real world or work" from using college sophomores in highly contrived settings? The laboratory has had its defenders. One argument is that being representative of the real world is not an important consideration when attempting to examine basic psychological processes. Indeed, in many laboratory experiments the investigator is more interested in determining TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 164 whether something "can" occur, not the frequency or strength with which it typically occurs in the external world. Other defenders of the laboratory have argued that the lab does not have to look like the real world to yield results that are generalizable. The more important issue for Berkowitz and Donnerstein (1982) is that the subjects in the lab interpret the lab situation in the same manner they would interpret the field setting. Others have argued that essential similarities or boundary variables determine whether lab findings are generalizable, not general similarity. Assume that the essential variable influencing whether subjects act as they would in the field is whether they are held accountable for their behavior. If the reasoning is correct, the laboratory need not be similar in all respects to the field but only with regard to this essential variable, i.e., make participants accountable for their actions. Another defense of the laboratory is that research in these settings is usually focused on general processes and theory. A researcher may need to create a novel or even ridiculous situation to flush out properly the phenomenon under investigation. Through such research, it is possible to validate a theory or model of the phenomenon that is then generalizable across a variety of situations. Whereas one can always doubt the generalizability of the findings of any individual, specific study, one can have much more confidence in the generalizability of a theory that is tested and developed through a program of experimental and nonexperimental research in both the lab and the field. The above arguments are only a few of the pros and cons of lab research. The best approach is perhaps to use more than one method, realizing that each is limited in some respects and that all research involves tradeoffs. In field research the researcher attains more realism but at the cost of precision and control. In lab research the researcher has a lot of control but at the cost of realism. By testing hypotheses in both the lab and the field, the researcher can determine whether findings in the lab setting are transportable to the field setting. There are a few areas of research where enough work has been done in both settings to allow these types of comparisons. A recent book edited by Edwin Locke (1986) reviewed findings in the laboratory and field from several areas of research and provided a highly optimistic view of the generalizability of laboratory research. The most reliable effects uncovered in the lab appeared generalizable to field settings. Looking for statistical interactions in evaluations of external validity. One could frame all of the above discussion about lab vs. field in terms of statistical interactions. Imagine a researcher who tests the hypothesis that a leader who allows employees to participate in decision making will have more productive employees than a leader who is authoritarian and directs the employees (see figure 3.11). The researcher conducts an experiment in the lab in which he randomly assigns to groups either a participative or an authoritarian leader. He finds that participative leaders have more productive employees than authoritarian leaders. He then replicates the experiment in the field. Employees are assigned to either a participative or authoritarian leader and the effects of the leadership styles on employee productivity are measured. Consider the following alternative sets of results (see figure 3.11). The real value of this relation, however, lies in the ability to predict how successful a TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 165 student will be in college based upon his or her performance on the SAT. In fact, that is how cutoff scores are frequently determined for acceptance into colleges or universities, and why prestigious schools have high cutoff scores. The mass of predictive data concerning SAT scores and college grades indicates that if a student does not perform at a certain level on the SAT, then the likelihood of success in college course work drops. As discussed earlier in the section on statistical procedures, it is possible to use more than one predictor in regression analyses. For example, in predicting first year college grade point averages, universities often consider the high school grade point average, letters of recommendation and SAT scores. The use of multiple predictors can enhance our ability to predict the outcome (in this case, college grades), and is referred to as multiple regression. Chapter 10 returns to this procedure in a discussion of how multiple regression is used to determine the optimal combination of tests to use in predicting job performance. It is important to recognize that achieving good prediction does not necessarily require an explanation of why variables are related. Why students with higher SATs achieve higher grades is still open to debate, but this has not stopped the use of SATs in college admissions. On the other hand, a relation between two variables may have a good explanation, even though it is not possible to predict one variable from the other. For an online tutorial on first, correlation, and then, regression check out: correlation: http://www.youtube.com/watch?v=BXXtkYOqAfM regression: http://www.youtube.com/watch?v=xojW6OEDfC4 http://www.youtube.com/watch?v=BXXtkYOqAfM http://www.youtube.com/watch?v=BXXtkYOqAfM http://www.youtube.com/watch?v=BXXtkYOqAfM http://www.youtube.com/watch?v=xojW6OEDfC4 TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 166 Figure 3.11: Hypothetical Results Showing Potential Effects of Leader Style and Research Setting. Ethics in I/O Psychology Research In discussing how best to conduct research it is easy to lose sight of the fact that human beings are the focus of the research. I/O psychologists along with all other psychologists must adhere to ethical guidelines in conducting their research (see chapter 2). Here is a summary of some of the major attributes of these guidelines. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 167 The I/O psychologist carefully considers prior to the research the possible risks involved and takes safeguards to protect the rights of the human participants. Perhaps the most important safeguard is that the "investigator establishes a clear and fair agreement with the research participants, prior to their participation that clarifies the obligations and responsibilities of each." This means that the investigator must inform the participants prior to the research of those aspects of the research that might influence their decision to participate. Moreover, the investigator should answer the questions of participants. Sometimes research requires that investigators keep participants in the dark on various aspects of the methodology. Investigators are under no obligation ethically to divulge everything to everyone prior to the research, but they must take care to protect subjects from physical or psychological harm. Most importantly, investigators must obtain the informed consent of participants by informing them of the potential risks and then giving them the opportunity to decline to participate or to withdraw during their participation. After the research is completed, the investigators are obligated to make a full disclosure of what was done and why. Investigators should treat any information collected on individual participants as confidential and should not share this information with others without the permission of the participants. When research is conducted in a field setting with actual employees the ethical responsibilities become even more complex. Not only is the investigator held to the above responsibilities, but there are also responsibilities to the organization and the management of that organization. An I/O psychologist is obligated to inform the organization of what he or she is doing and must ask their permission to conduct the research. A particularly sticky issue can arise when the I/O psychologist is in the situation of being a consultant to the organization. If an employer does wish to disclose the results of research, the psychologist must tell participants prior to the research who will and will not receive the results. Moreover, a prior agreement is struck with management of the organization that participation is voluntary and that each employee has the right to decline to participate or to withdraw at any time without being punished. Also, each employee participating in the research must consent before the researcher can use the information that the employee has provided in the research. If the organization does not agree to these procedures, then the I/O psychologist should look for an organization that will. Psychologists who violate the ethical standards can be reported to ethics committees at the state and national level. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 168 Academic Research vs. Practical Realities Some I/O psychologists work in industry, some in consulting firms, and some in universities. The demands placed on them differ, and as a consequence, their views of how to best conduct research can differ considerably. Psychologists in industry are under pressure to come up with answers to problems. The psychologist in academia is expected to conduct careful, programmatic research and to avoid rash conclusions. The former may view the latter as ivory tower eggheads with little conception of the real world. The latter may view the former as flimflam artists willing to sell any method that can make a buck regardless of the scientific evidence. Such stereotypes are quite unfortunate, and while they do occur, they are not widespread. Indeed, one of the major strengths of I/O psychology is the diverse mixture of practitioners and academics that contribute to the storehouse of knowledge. Nevertheless, there are differences between the ideal scientific model and the way investigators often must conduct research in organizations. Virginia Boehm (1980) has provided a thoughtful comparison of the two approaches. In contrast to the academic model of research, research conducted within an organization can follow a much different course (see figure 3.12). In the first place, the topic is often not chosen by the investigator but is prompted by an organizational problem, which the investigator is asked (or ordered) to solve. After some analysis, a solution is proposed (a training program, a new performance appraisal system, etc. and a study is designed to evaluate the solution. The investigator often has to persuade the organization that the research is worthwhile and even after conducting the study must sell the organization on the benefits of the proposal. Investigators often have to go to Herculean efforts to maintain rigorous experimental control when conducting research in an organization and the effort more often than not is frustrated by management's desire to get on with it and come up with a solution. The solution is not to give up on standards of scientific research, but instead to open the method to creative solutions. Well worth remembering is P. W. Bridgman's dictum that "the scientist has no other method than doing his damnedest" (Kaplan, 1964, p. 27). TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 169 Figure 3.12: The Process of Conducting Research in Organizations Points to ponder 1. A large proportion of the public is skeptical of scientific research and instead rely on the nonscientific sources of knowledge. Take, for example, the germ theory of disease. By the late 1800s this theory was well established but many people still rejected the theory and the scientific research validating it. Even physicians in the1800s rejected germ theory and ridiculed the idea of having to wash their hands before surgery. Why do people find it so hard to accept scientific research and instead adhere to notions that ultimately prove false? 2. Search the internet, magazines, newspapers, TV, and the radio for media reports of research on human behavior. How would you evaluate what was done and was claimed TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 170 based on what you have read in this chapter? 3. Randomness (either in the form of random selection or random assignment) is an important concept in scientific research. Provide examples of its importance from the discussion of statistics, measurement, and research strategies. 4. What are some of the research questions pertaining to human behavior at work and in organizations that you think can only be researched in the field using correlational methods? Why can’t true experiments be conducted on these topics? On other hand, identify questions pertaining to human behavior at work and in organizations that can be the focus of true experiments? 5. Must a science accomplish all three goals: prediction, explanation, and control? Can you provide examples from established sciences in which one or two were accomplished but not all three? 6. Some have argued that generalizability is at the heart of both construct validation and the assessment of reliability. How is generalizability an important concern for both topics? 7. Why is reliability a necessary precondition for criterion-related validity but not vice versa? Can you provide examples? 8. Compare and contrast the concepts of correlation and prediction. 9. Explain why could argue that internal external, external validity, and reliability are all subsumed under the notion of construct validity. 10. Compare and contrast the steps one must take in rigorously applying the scientific method with research that often occurs in organizational settings. Is it ever worthwhile to conduct research even though you cannot adhere to all the scientific standards? Why or why not? Conclusions This chapter admittedly makes for difficult reading and is probably not the favorite topic for many of the readers of this text. It is only natural to search for answers, and a message of this chapter is that science is the best way to seek these answers. But the primary lesson of this chapter is that the journey of a scientist is never really completed. Obviously those responsible for making decisions have to act and cannot engage in a lifelong search. They must make a decision based on the best available information even though this information may be incomplete or even flawed. There is a tension between the science of I/O psychology and the practice of I/O. The trick is living with the tension and maintaining a creative dialogue between the two. The present chapter reviews a variety of techniques and tools that can be used in scientific research. All of these techniques and tools are limited in some way. To build a valid body of knowledge in I/O the research needs to employ multiple methods so that weaknesses of an approach can be balanced with the strengths of other approaches. In the end, one can only hope that the scientist does his or her best in the pursuit of knowledge. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 171 CHAPTER 4: WORK MOTIVATON TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 172 Introduction A lack of motivation is a common explanation for poor performance in organizations. Take, for example, this posting from a supervisor of 60 employees providing cardholder account services: “I have never seen a group of employees like this…… I organized two Town Hall meetings. In each meeting I laid out a number of new practices that will make our work faster and more accurate. The employees sat in their chairs and shot me evil looks. No one spoke up. I am beyond frustrated…. My take is that the employees are lazy. They don’t want to work.” (R. Liz, “Can anything motivate my lazy employees?”, June 28, 2015, Forbes, http://www.forbes.com/sites/lizryan/2015/06/28/can-anything-motivate- my-lazy-employees/#246999a54da0) A lot of the concern about lack of motivation is aimed at younger workers as seen in the comments by the chef at a leading restaurant in Great Britain. “The 38-year-old TV chef – who has created thousands of UK restaurant jobs in the last five years – reckons most of our youngsters are not up to a 46-hour week. Jamie said: ‘I think our European immigrant friends are much stronger, much tougher. If we didn’t have any, all of my restaurants would close tomorrow’”(http://www.mirror.co.uk/news/uk-news/jamie-oliver- blasts-young-britons-2231664). These sentiments are echoed by the owner of restaurant in the United States: “many younger workers do not accept that it takes long, concerted effort to build a career. ‘They’ve been blessed with parents and grandparents laying the foundation to give them a better life,’ he said. ‘But that hunger is not really in them. But the desire for success is. They want to make money but don't want to put in the required hours or effort’ “(I. Shapira, April 3, 2010, Millennials accused of lax work ethic say it's not all about 9-to-5, Washington Post). A frequent accusation is that a lack of motivation to work hard on the job is a driving force behind the economic success or failure of entire nations. Take, for example, the European debt crisis. One commentator wrote “people in prosperous northern European countries “believe in a simple moral formula: effort should lead to reward as often as possible … self-control should be rewarded while laziness and self-indulgence should not.” European countries such as Italy, Spain, and Greece lack this work ethic and as a consequence have failing economies. (http://www.nytimes.com/2011/12/02/opinion/brooks-the-spirit-of-enterprise.html?_r=0) What do you think of these observations? Have you ever experienced co-workers or subordinates who seemed lacking in work motivation? Do you think younger workers lack a work-ethic and that immigrants are more motivated to work hard? Do you agree that differences in the economic success of countries are the result of differences in the motivation of the citizenry? In observing people at work on specific tasks, it is obvious that some people work harder than others. But why? Explanations that use work motivation to account for why things go right or wrong typically rely on vague http://www.forbes.com/sites/lizryan/2015/06/28/can-anything-motivate-my-lazy-employees/#246999a54da0 http://www.forbes.com/sites/lizryan/2015/06/28/can-anything-motivate-my-lazy-employees/#246999a54da0 http://www.mirror.co.uk/news/uk-news/jamie-oliver-blasts-young-britons-2231664 http://www.mirror.co.uk/news/uk-news/jamie-oliver-blasts-young-britons-2231664 http://www.nytimes.com/2011/12/02/opinion/brooks-the-spirit-of-enterprise.html?_r=0 TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 173 commonsense notions about what it means to be a hard worker or a lazy worker. They build on stereotypes based on nationality, age, and race, but provide little insight and few solutions. To improve on these commonsense notions, psychologists have attempted to clarify the behavioral, cognitive, and emotional processes involved in work motivation. The present chapter focuses on the theory and applications that are the products of this work. Key behavioral indicators of motivation So what is motivation? Similar to most psychological constructs, you cannot directly observe motivation but instead must infer it from behavior. Take a look at this short collection of clips from the movie Rocky. In case you do not remember, Rocky Balboa is a fighter who is over the hill, and apparently at the end of his career, but is chosen to fight the world heavy weight champion. Based on the clips, what are the behavioral indicators of Rocky’s motivation? (http://www.youtube.com/watch?v=45FfHlzRqTI&feature=related) Psychologists have defined motivation in terms of three behavioral characteristics. 1) Direction. The decisions that are made among alternative courses of action are one indicator of motivation. In the movie Rocky must decide whether he will seriously try to fight the champion or just walk through the rounds and collect his money. He chooses to make it a serious fight. Highly motivated behavior is always characterized by some decision on the part of the person to go in one direction and to not go in other directions. The fact that Rocky is in the gym or running in the streets as opposed to drinking in a bar or going to the movies provides additional evidence that he is highly motivated to achieve his goal...going the rounds with the heavyweight champion of the world. 2) Effort and vigor. Another indicator of motivation and the one we most often associate with motivation is effort exhibited in the pursuit of the goal. Rocky shows vigor and intensity in his training for the fight. On physical activities this is shown in the flex of muscles, the strain in facial features, sweat, and grunts and groans. It is also shown in the time spent on the activity, the amount of workload, and speed of working. Effort is also apparent on cognitive tasks. A person involved in solving a difficult task may not sweat, but you can probably see the strain, vigor, and intensity of the efforts to solve a problem. 3) Persistence. A third indicator is the persistence in pursuing the goal. Rocky does not give up, in the face of the pain, the frustration, and the doubts of others that he can succeed. Rocky keeps at his workouts to the point of possible exhaustion despite the difficulty of the tasks and setbacks. There are other characteristics of motivated behavior, but psychologists attempting to measure how motivated a person is to work on a task or achieve a goal have used these as the primary indicators. http://www.youtube.com/watch?v=45FfHlzRqTI&feature=related http://www.youtube.com/watch?v=45FfHlzRqTI&feature=related http://www.youtube.com/watch?v=45FfHlzRqTI&feature=related TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 174 Performance vs. motivation A few more terms deserve some attention before diving into the topic of work motivation. One distinction that the reader should keep in mind is that between performance and motivation. The level of performance on the task is often used as an indicator of motivation, but motivation is not the same as performance. Although there is usually a correlation between motivation and performance, just because a person is highly motivated does not mean that the individual performs at a high level and low motivation does not necessarily translate into low performance. As described in this formula, performance is a function of not only motivation but also the ability of the individual and the knowledge of the person. Performance = Ability X Motivation X Knowledge The multiplicative nature of the combination above reflects the fact that if any one of these three determinants is zero, it can cancel out the other two. For instance, even with an extraordinarily high level of motivation, persons totally lacking the ability and knowledge to do the task are likely to perform poorly. It is more appropriate to look for the more direct and immediate signs of motivation in the form of choice, effort, and persistence rather than an indirect and more distal indicator such as performance. Having made clear this distinction, the reader will see research in this chapter that measured performance as the primary indicator of motivation. If the task is so simple that everyone possesses about the same ability and knowledge, this is not a problem. The more complex the task, the less one can rely on performance as an indicator of motivation. On a complex task differences in knowledge and ability, in addition to motivation, are causes of performance. The internal focus of motivation theory Another important consideration to keep in mind is that motivation refers to forces within the person. This chapter reviews a variety of psychological constructs that motivation theorists propose as mediators of the effects of external events such as the work environment and outcomes in the form of direction, effort, and persistence of goal- directed behavior. A personal need, such as a need to achieve or gain power, could explain why a person chooses to work overtime. An expectation of reward may keep a salesperson making cold calls despite repeated failures. A belief about what is fair and unfair may motivate a person to restore justice in a situation. A sense of self-efficacy and competence may spawn vigorous work activity in the absence of any obvious reward for these actions. Each motivational construct presumes that there is something inside the person that accounts for behavior. The interaction of the environment with internal forces An understanding of motivation requires more than internal forces. One must also consider the environment in which the person is pursuing the goal. Specifically, TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 175 understanding motivation requires a consideration of how the environment interacts with internal to reinforce or redirect that goal directed behavior. A person who has a need to achieve may reevaluate his goals to go to medical school after receiving poor grades in biology and chemistry. A worker with a high need for social affiliation and approval may decide to seek friends outside the work place and to become less of a team player after other work group members criticize him and give him the cold shoulder. A salesperson may, after repeated failures to make a sale, lower her expectations and spend less effort making cold calls. In each example, the interaction of personal factors with environmental factors constitutes a motivational system (Steers & Porter, 1991). This chapter takes a systems orientation by examining targeted personal characteristics, such as achievement needs, environmental characteristics, such as incentives provided by the organization, and combinations of both personal and environmental characteristics, such as how people with different achievement needs are affected by organizational incentive systems. Seven practical questions This chapter organizes the discussion of work motivation around seven practical questions a person would ask in attempting to motivate workers to pursue a goal and to pursue this goal with vigor and persistence. 1. What are the goals of employees and the characteristics of these goals? 2. What are the consequences of employee efforts to achieve these goals? 3. What do employees expect to occur if they try to achieve the goals? Will they succeed in performing the task (self-efficacy expectations) and what are the outcomes they expect from succeeding on the task (outcome expectancies). 4. Do the consequences of attempts to achieve goals fulfill important needs and are they valued? 5. Do employees perceive that the efforts of management to motivate them are fair and just? 6. Is the work environment relevant to the self-concepts and self-evaluations of employees? 7. Is the behavior shown by employees in the pursuit of goals intrinsically or extrinsically motivated? These questions constitute the steps we would take in attempting to increase the motivation of workers. The first step is to set a goal for the workers to pursue. The second step is to provide consequences in the form of rewards and punishments for actions taken in the pursuit of the goals. The third step is to ensure that the workers have high expectations for accomplishing these goals. The fourth step is to examine the needs of the workers and to ensure that hard work in the work environment pays off in fulfilling important needs. The fifth consideration is whether workers believe that they are treated fairly and justly. A sixth consideration is whether employees believe that the pursuit of task goals can allow them to fulfill valued self-ideals and boost their self-esteem. The final consideration is whether the direction, effort, and persistence of the behavior shown by workers in pursuit of goals reflect intrinsic or extrinsic motivation. Each section will TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 176 review the theoretical framework for the question that is posed and the empirical research that has tested the theory. The implications of both theory and research for application in organizations are also discussed. The last topic covers recent efforts to combine the various motivation theories into larger, integrative theoretical frameworks, or meta- theories. What are Employees’ Goals? It is meaningless to ask “Are employees motivated?” Of course they are motivated. The only unmotivated human is a dead human. A more meaningful question to ask is “What goals are employees motivated to achieve?” Especially important is the question of whether their person goals coincide with the goals of management and the organization. One must begin with goals because motivation is defined as goal-oriented behavior and the characteristics of the goals that people pursue are crucial determinants of the vigor and persistence they show in pursuit of organizational objectives. An example of a practical intervention that builds on setting specific, attainable goals to achieve some desired outcome is Weight Watchers. Its members set weekly goals (for example, a two-pound weight loss) that are attainable, instead of simply attempting to lose weight. These weekly goals are also much less psychologically distressing than the thought of having to lose 40 pounds to achieve long-term goal. Similarly, time management experts recommend the setting of specific, attainable goals for each project. Large or long-term projects divided into smaller projects, each with its own goals. An example is writing a term paper for a class. Because many students consider writing a term paper a difficult task, they often postpone starting the paper until just before it is due. The consequences are frequently a poorly conceived and written paper that receives a low grade. Time management experts suggest that dividing the task of writing the paper into discrete, manageable portions and set specific goals for their attainment. First, early in the semester, plan to go to the library to research the topic; set specific time limits for this first phase. Then, draw up an outline that describes the various sections of the paper. Again, allow only a certain amount of time for accomplishing each section. Finally, following the outline, write one section of the paper at a time, reviewing and editing the old sections when writing the new or later ones. This last phase probably consumes the most time. A helpful secondary goal would be, for example, to work on the paper at least one hour every two days. Writing the term paper in the manner just described, will produce a carefully conceived and written product and little, if any, stress. Why goals improve performance: Mediators of goal effects A variety of practical interventions ranging from Alcoholics Anonymous, Weight Watchers, Time Management, and Management by Objective (MBO) have achieved impressive results using the goal-setting process. Goal setting really works! Not surprisingly, the potential usefulness of goal-setting techniques also attracted the attention of industrial organizational psychologists such as Edwin Locke (1968), who proposed goal setting as a theory of work motivation. In goal-setting theory, conscious intentions motivate the person to do and act according to plans. Events in the external TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 177 environment, such as the successful completion of a college course or a potential promotion for superior work performance, trigger a cognitive process in which the person evaluates the events against his or her values. Values in this context refer to what the individual wants to attain. If the person's values are not satisfied by the external event, he or she will experience a negative emotion. The outcome of this emotional experience is a conscious intention directed toward the completion of a goal. Suppose Meredith’s supervisor criticizes her poor job performance. Because Meredith places a high value on work, she is very unhappy with her negative evaluation. Consequently, she sets a goal to improve her job performance before her next six-month review. For Locke, the intentions and thoughts of the person are the primary mediators in the process by which goals influence behavior. Perhaps the most inspiring illustration of the use of goals to motivate was President John Kennedy's speech at Rice University in 1962 in which he set as a specific, difficult goal getting a man on the moon within the decade (http://www.youtube.com/watch?v=g25G1M4EXrQ). There are four primary mechanisms or mediators through which goals impact performance, according to Locke, and we can see each mechanism at work in Kennedy’s speech and the subsequent efforts to get a man to the moon. First, goals direct the attention and effort on the tasks that will accomplish the goal. There were a lot of other potential goals that could have captured the attention of the U. S. public at the time of Kennedy’s speech, but in setting a goal to put men on the moon, he focused the attention of the public on this one objective. The reader has probably experienced this when a deadline for a project is coming. The completion of this project becomes the sole focus of attention as the deadline approaches. Second, goals energize leading to greater physical and cognitive effort on the tasks related to accomplishing the goals. President Kennedy’s speech was clearly inspiring in its vision and it triggered efforts of educators, legislators, parents, and others to accomplish the objective. Third, goals lead to persistence on the tasks. When a goal is set, individuals are more likely to keep working at reaching that goal even when they fail. Without a goal, they may quickly give up. The space program was faced with many setbacks including a disastrous event on the launch pad that led to the death of three astronauts. With the goal of putting men on the moon within the decade, the efforts continued even in the face of the difficulties. Fourth, goals stimulate the person to think about strategies for achieving the goals. The presence of goal triggers cognitive processes in which the individual thinks about ways to reach the target with the least effort and this, in turn, leads to higher performance than an absence of a goal. The clear, difficult goal set by Kennedy stimulated a huge amount of intellectual activity as NASA’s engineers, scientists, and astronauts attempted to figure out strategies for achieving the objective. What goal characteristics influence motivation and performance? http://www.youtube.com/watch?v=g25G1M4EXrQ TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 178 The simple elegance of goal-setting theory has captured the attention of many organizational researchers over the past twenty years; so much, in fact, that goal setting is currently among the most widely researched and accepted theories of work motivation. The research has identified several characteristics of goals that provide leverage to the employer attempting to motivate employees. Goal difficulty and specificity. More than 50 years of research and hundreds of experiments (Latham, Mitchell, & Dossett, 1978; Latham & Steele, 1983; Locke, 1968; Locke, Shaw, Saari, & Latham, 1981) have supported the following basic propositions stated in goal-setting theory: Specific, difficult goals lead to higher performance than no goals, easy or moderate goals, or "do your best" goals. Approximately 90% of all laboratory and field studies have confirmed this prediction. Before the goal can lead to optimal results, a person must not only develop conscious intentions of achieving the goal and translate those intentions into behavior, but he or she must also set difficult and specific goals. Without these concrete guideposts, people do not have enough structure to perform optimally. There are very few universal laws that apply to human behavior, however, and as we will see there are limits even to this strongly supported proposition. Knowledge of results (feedback). A goal provides a target that the person can use to gauge the success of his or her efforts. Feedback on how close the person has come to meeting the goal (goal discrepancy feedback) energizes the person and provides information on task activities that needed to bring the performance in line with the goal. Although goals without feedback may still increase motivation, the addition of feedback to the mix has a large and beneficial effect on performance. This is especially true on complex tasks where the effect of goals + feedback more than doubles the effect found for goals alone (Neubert, 1998). Contemporary researchers have attempted to examine in some detail the efficacy of different types of feedback. Researchers in one experiment explored the differential effects of two types of feedback (Vance & Colella, 1990). Goal discrepancy feedback informed the participants of how they were performing relative to the goal that they had been assigned. Performance discrepancy feedback informed them of how their performance had changed either in a negative or positive direction relative to their performance on the previous trial. The experiment was designed so that participants received increasingly difficult goals and, as a consequence, increasingly negative goal discrepancy feedback. Although the participants were given both goal discrepancy and performance discrepancy feedback after each task trial, the authors found that the participants increasingly focused their attention on the performance discrepancy feedback as they experienced increasingly negative goal discrepancy feedback over the trials. These findings bring attention to the fact that there are usually multiple sources of feedback in a task environment and feedback on goal discrepancy is only one of these. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 179 One cannot assume that the focus of attention will remain on the goals that were set for the employee but may instead shift to other standards such as past performance. These two studies illustrate an exciting trend in goal-setting research that may enhance its usefulness in organizations. Participation in goal setting. Another important characteristic is the extent to which the employee participates in the setting of a goal. The readers may remember occasions in which they set a goal for themselves and other occasions in which someone else set the goal for them. In the latter instances, the person setting the goal may have sought their input or may have simply told them the goal without any attempt to explain the rationale. How did the response in the full participation situations differ from the response in situations in which the goals were assigned without participation? A human relations approach to management suggests that participation enhances the commitment of the person to the goal and leads to more effort than assigning the goal without participation. Contrary to this assertion, research has shown that the relationship of participation is more complex and depends on other factors. In one series of experiments examining the role of participation, the researchers found that participants whose supervisor assigned a goal exerted less effort on the task than participants who participated in the goal setting only when the supervisor simply told the employee the goal without explaining it (Latham, Erez & Locke, 1988). When there was an explanation for the goal, there were no differences between people who participated and those who did not on the effort they exerted. This is not the end of the story on the influence of participation but these findings show is that there is more than one component to participation. One very important component is the extent to which supervisor justifies and explains the decision. As shown in this research, in some instances a simple explanation is sufficient to ensure that a goal stimulates a higher level of motivation. Another very important component of participation that is often related to justification is the extent to which the participation facilitates the generation of strategies for accomplishing the goal. For instance, if a supervisor and employee or a group of employees discuss the goal, they may derive various ways of best doing the tasks and this may improve performance. Of course, whether participation in goal setting benefits participation depends to a great extent on the quality of the strategic planning that is associated with the participation. In an experiment involving 16 simulated organizations, the researchers evaluated the influence of the actual time spent planning and the quality of the planning process (Smith, Locke & Barry, 1990). Planning quality was measured by the presence of such factors as the development and communication of action plans. The researchers found that more time spent planning led to higher performance only if the planning quality was high. These studies suggest that the quality and amount of planning activity are crucial mediators of the impact of participation in goal setting on performance. If participation stimulates high quality planning it is likely to benefit performance, but if it stimulates poor quality planning, it may well hurt performance. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 180 What moderates goal effects? Remember that a moderator variable is a variable that influences the relationship between two other variables without necessarily causing either of the other two variables. A mediator is a variable that “comes between” and accounts for the relationship between two variables. The mediator is an outcome of the independent variable and, in turn, causes the dependent variable. The distinction is not always entirely clear, but research has identified several variables that appear to serve as important moderators of the effect of goals on performance and motivation. The research setting: laboratory vs. field. One meta-analysis of goal-setting research showed that goal-setting performance effects were stronger in laboratory than in field studies (Tubb, 1986). One possible explanation is that participants were more willing to accept and work toward harder goals in short-term situations, such as laboratory experiments, than in field research. Like equity theory research, goal-setting research has been conducted mostly in laboratory settings with student subjects performing simple tasks, such as listing novel uses for a coat hanger or paper clip. Subjects in these experiments are typically asked to perform the task over several trials with some subjects assigned goals without participation and others participating in the setting of the goals. In the lab experiments, the duration of the task performance is very short, whereas in field settings goals are frequently set for weeks or months. Despite the difference in the strength of goal setting effects in lab and field settings the reader should not conclude that the lab research is invalid. One would expect stronger effects in the lab because of the greater control over extraneous variables. Indeed, this is the strength of the lab --- i.e., it allows a stronger test of theory as the result of greater control over variables. The lab allows us to determine whether goal setting “can” affect performance but does not allow a good basis for determining the actual strength of the effect in the field. One can always expect the effects found in the lab are not as strong in the field for the obvious reason that there is more going on in the field. Goal acceptance and commitment. For a goal to energize, focus attention, lead to persistence, and generate task strategies, Edwin Locke and his associates (Locke, Latham & Erez, 1988) propose that individuals must initially accept the goal and then become committed to achieving the goal. Acceptance and commitment to the goal are moderators of the effects of goals on performance and motivation to the extent that they result from factors other than the goal itself. It is feasible one could view them as mediators if the goal itself led to the level of acceptance and commitment observed. This chapter treats them as moderators. Among the factors that seem integral to forming a commitment to the goal are self-efficacy, goal importance, and monetary rewards: 1. Self-efficacy or the expectancy that if I try I can achieve the goal. Increasing the self- efficacy of the employee that he or she can achieve the goal is an important strategy for increasing commitment to the goal. If self-efficacy is low, the employee is likely to reject the goal or show a low commitment to accomplishing it. Supervisors can do a lot to TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 181 increase self-efficacy by communicating the expectation that the worker will achieve the goal and possibly inspiring the worker to make the attempt. 2. Public commitment. Another strategy for increasing goal commitment is to have the participant make a public commitment in which he or she states to others the goal and the intention of achieving the goal. 3. Monetary rewards: Still another way of gaining commitment is to provide monetary rewards for accomplishing the goal. An important factor to consider, however, is the difficulty of the goal. If a goal is set that is too difficult, providing monetary rewards actually can hurt performance (Lee, Locke & Phan, 1997). The employee apparently becomes less committed to the goal once he or she perceives that goal accomplishment is unlikely. It is also possible that there is a boomerang effect in which the person exerts even less effort than would have occurred without the goal and the monetary inducements. Commitment to goals moderates the effect of goal difficulty on performance according to the results of one meta-analysis (Donovan & Radosevich, 1998). In other words, performance increased with goal difficulty more for persons with high commitment to the goal than for those who had low commitment. However, the effect was quite small and accounting for only about 3 percent of the variance in performance. The authors of this meta-analysis suggested that researchers need a more unified conceptualization of goal commitment and better measures of this construct. The nature of the task. The strength of the goal-setting-performance relationship is affected by contextual factors including the complexity of task performed, the extent to which the person must coordinate with others, and the stage of learning the task. 1.Task complexity. Stronger goal effects are demonstrated with simple tasks than with complex tasks or novel tasks that allow multiple, alternative strategies. Specific, difficult goals appear to enhance performance on simple tasks but that this effect does not appear to extend to novel tasks that allow multiple, alternative strategies. In one experimental demonstration, undergraduate students performed a stock market prediction task and were given either a general goal (e.g., “do your best) or a specific, difficult goal (Earley, Connolly, & Ekegren, 1989). Participants assigned specific, difficult goals engaged in more strategy search, but the additional search time did not improve performance on this complex task. The researchers concluded that whereas specific, difficult goals do improve performance on simple tasks, they are not as beneficial on complex tasks. This was also the conclusion from a meta-analysis of goal-setting research that assessed the moderating effects of task complexity by examining whether the type of task affected the goal-setting-performance relationship (Wood, Mento, & Locke, 1987). Goal-setting effects were strongest for easy tasks (reaction time, brainstorming) and weakest for the most complex tasks (business game simulations, scientific and engineering work). The findings of this review are summarized in figure 4.1. The mean corrected effect size (or Cohen’s d) along the Y-axis is the difference in performance when the goal was specific TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 182 or difficult as opposed to easy or general. The larger the d, the stronger the typical goal effect, i.e., the more specific, difficult goals led to higher performance. As shown in the figure, the effect of goal diminished with increasing complexity of the task. Figure 4.1: The Moderating Effects of Task Complexity on the Effects of Goal Specificity and Difficulty on Performance. 2. Required cooperation. In a simple but clever laboratory experiment, people who used individual goals on a task that required cooperation performed more poorly than those who used group goals (Mitchell & Silver, 1990). In this case, the setting of individual goals led to more competition and less cooperation. Because success required cooperation, the competition led to less effective performance on this task. The researchers recommended from these findings that when the task requires cooperation, group goals are likely to encourage collaboration and are preferred to individual goals. 3. Learning vs. performance stage of task. Stronger effects of difficult, specific goals are also obtained when the person has mastered the task (Kanfer, Ackerman, Murtha, Dugdale & Nelson, 1994). Setting difficult, specific goals when the person is learning a task may hurt performance. The same thing probably occurs when difficult, specific goals are set for complex tasks. When one has not yet mastered the task, it is important to stay flexible and learn from mistakes. A very specific, difficult goal may lead to a rigidity and unwillingness to experiment. Individual differences. Not everyone reacts in the same way to goals. The impact of goal setting on behavior and performance is moderated by Individual differences among TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 183 employees, such as personality and ability characteristics. High task-ability people reliably outperform low task-ability people in response to increases in goal difficulty. Hollenbeck, Williams and Klein (1989) examined the influence of personality characteristics on commitment to difficult goals (see figure 4.2). They found that whether participants set their own goals or others assigned the goals another made little difference if the person was low on need for achievement, but self-set goals led to higher performance than assigned goals for those high on need for achievement. The finding that high need achievement persons are motivated to achieve task success by self-set goals is consistent with the preference of high need achievement persons for situations in which they have personal responsibility and control. Another individual difference variable that appears to predict success in academic domains is GRIT (Duckworth & Gross, 2014). This is the propensity to set long term goals and then exhibit perseverance and vigor in the pursuit of these goals. One could hypothesize that those higher on GRIT will show larger increases in effort, persistence, and performance in response to challenging goals than those lower on GRIT. To my knowledge this has not yet been tested. Figure 4.2: Moderating Effects of Need Achievement on Self-set vs. Assigned Goals All things considered, goal theory is a powerful framework for explaining motivation in the work place. In addition to its theoretical benefits, goal theory provides an obvious tool for motivating employees. There are moderators of the effect, however, and goal setting is more powerful in some circumstances than in others. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 184 Points to ponder 1. Why are specific, difficult goals so often effective in increasing motivation and enhancing performance? 2. What are the factors that can prevent specific, difficult goals from boosting motivation and benefitting performance? 3. In light of the research on goals, what would you do to ensure that goal setting will boost motivation and improve performance? 4. Some supervisors are reluctant to set specific goals and to openly state that these goals will serve as the standards for evaluating performance and providing rewards. Why the reluctance? 5. Can goals become a substitute for action? Under what circumstances might this occur and why? What are the Consequences of Employee Behavior? Self-set or assigned goals are the first step in motivating another person. Once the goals are in place, the next step is to make sure there are consequences for achieving or not achieving these goals. That the environment affects behavior by means of consequences is not surprising. The readers no doubt can recall many instances in which outcomes originating from the outside world influenced their behavior. For example, if the skies are dark and cloudy tomorrow, they will probably decide to take an umbrella or raincoat when departing for work or school. They do this because of past situations in which there were similar weather conditions and the consequence was a soaking due to a lack of an umbrella. In response to cloudy, threatening skies they carry an umbrella because of the rewards for carrying an umbrella in these situations and negative outcomes for not carrying one. Similarly, if a company's policies provide negative consequences for tardiness, employees are more likely to attempt to show up on time for work than if the company lacks such policies. Again, they do this because there are consequences of tardiness that they want to avoid and other consequences for being punctual that they desire. Psychologists who take a radical behavioristic approach (this approach was discussed in the history chapter the pioneers of this school of thought such as John Watson) systematically investigate how aspects of the environment influence human behavior. The radical behaviorist believes that psychologists should study only observable behavior, not feelings or affect, and that behavior is shaped solely by environmental factors. The reader is probably thinking at this moment, “wait a minute” …you just told me that motivation dealt with internal factors. Yes, a strict behavioristic approach is in some ways an anti-motivation approach because it eschews any mention of constructs such as needs, expectations, intentions, and the like. Few psychologists these days adopt a radical behavioristic approach. Nonetheless, such radical behaviorists as Watson and Skinner contributed greatly to understanding of how consequences play a role in motivation even though they generally ignored internal factors. The application of behavioral principles to shaping behavior in real situations is called Behavioral TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 185 Modification or BMod for short. Here are some of the major distinctions and learning principles that are at the core of BMod. Types of conditioning The behavioral psychologist recognizes two basic learning processes: classical conditioning and operant conditioning. Classical conditioning (Michael & Meyerson, 1962; Pavlov, 1902) is concerned with involuntary or reflexive behavior and is not directly applicable to our discussion of work motivation. However, researchers and practitioners have applied operant or voluntary conditioning quite broadly in work settings (Michael & Meyerson, 1962; Skinner, 1969). In operant (or voluntary) conditioning, rewards (and punishments) are contingent upon the subject's response or failure to respond (i.e., they immediately follow the response). If the door-bell rings, one will probably answer the door. The reason for this action is that going to the door and opening it is shaped with operant conditioning; from many prior experiences, one knows that someone is on the other side of the door. The stimulus is the ringing doorbell; the response is answering the door; and the positive consequences are greeting and visiting with a friend or acquaintance. The decision to answer the door is always totally voluntary and under control. The ringing doorbell does not automatically elicit walking to the door and opening it. Rather, the ringing doorbell sets the stage and acts as cues for the acts of going to the door and opening it. The actual response is under your control and voluntary, not involuntarily evoked as in classical conditioning. A common example of operant conditioning in a work setting is the administration of praise by a supervisor for a task well performed. The worker completes a report in a timely and thorough manner and is praised by his supervisor. This sequence of events increases the probability that the worker will decide to respond similarly when preparing future reports. Operant conditioning. In using operant conditioning, three types of outcomes or consequences are distinguishable: positive reinforcers, negative reinforcers, positive punishments, and negative punishments (see figure 4.3). Based on the research, behaviorists clearly advocate the use of positive reinforcement and warn against both negative reinforcement and both positive and negative punishment. In positive reinforcement a positive consequence, such as praise, follows the desired response and increases the frequency of the response. Realize that it is the increase in the probability of the response as the result of giving the reinforcer that defines positive reinforcement, not the desirability of the reinforcer. Check out, as a humorous interlude, this short illustration: http://www.youtube.com/watch?v=9UB0NHgK_Bc Negative reinforcement refers to an increase in the frequency of a response as the result of the removal of a negative consequence. If the readers eat with their hands rather than using a spoon and fork, a negative consequence is criticism from others. When the reader starts using the silverware, the criticism stops. The removal of the negative consequence….criticism….increases the probability of eating with silverware. Punishment, on the other hand, involves decreasing the probability of an undesired http://www.youtube.com/watch?v=9UB0NHgK_Bc http://www.youtube.com/watch?v=9UB0NHgK_Bc http://www.youtube.com/watch?v=9UB0NHgK_Bc TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 186 behavior by providing a consequence. With negative punishment, a noxious consequence is provided and is only withdrawn once the person stops doing what is being punished. With positive punishment, a negative or aversive outcome is administered until the person stops doing what is being punished. An example of positive punishment is a supervisor reprimanding a worker (or withheld a pay raise) because the worker did not produce a timely and thorough report. With negative punishment a positive outcome is withheld until the worker stops the behavior that is being punished. An example of negative punishment is a supervisor who withholds friendly interaction with the worker until the worker starts producing the report. Figure 4.3: Types of Operant Conditioning Although punishment is commonly used in all phases of life, including organizations (see Arvey & Ivancevich, 1980), behaviorists do not generally advocate its use (Skinner, 1969; Whyte, 1956). Based on the results of research on various types of reinforcement, psychologists generally perceive punishment as not only cruel but also ineffective compared with positive reinforcement. One reason for punishment's ineffectiveness is that it demonstrates to people what they should not do, but does not demonstrate what they should do. Reprimanded employees know what they did was wrong, but they may not know how to correct the behavior. Punishment can evoke negative emotion such as pain, fear, anxiety, and sadness, all of which can interfere with performance. Also, people possess beliefs about what is fair and justice in the treatment of people and the distribution of outcomes. Punishment is likely to violate these beliefs and lead to perceptions of unfairness. If you are still not clear about the distinctions among positive TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 187 and negative reinforcement and punishment, check out this video. It also provides a quiz to test your understanding: http://www.youtube.com/watch?v=wfraBsz9gX4 Schedules of reinforcement Behaviorists are concerned with different types of schedules of reinforcement and the question of which type can lead to the best performance. For example, to produce optimal performance, should the supervisor praise the employee for every good performance, or only occasionally? As the result of a huge body of research on non-humans, a lot is known about what works with rats, pigeons, dogs, cats, and other critters. Whether the findings generalize to humans is less certain. There are five basic types of reinforcement schedules. Continuous reinforcement is where reinforcement is provided each and every time the response is emitted. This schedule is always necessary at the beginning of conditioning to establish the connection between the behavior and the contingent outcomes. Once the behavior is learned we can continue with continuous reinforcement or switch to one of four types of partial reinforcement schedules. Partial schedules, which are also called intermittent reinforcement schedules, are either fixed or variable. With a fixed schedule, the reward is received after a fixed time, such as a bonus every six months, or after a fixed number of correct responses, such as after every 1,000 widgets produced. Under variable schedules, the reinforcement schedule is not obvious or predictable because it varies around some average time interval or number of responses. There are four types of partial or intermittent reinforcement schedules. 1. Fixed ratio schedules provide the reward after a specified number of responses and this number stays the same over time. An employee could receive a bonus each and every time she produces 1,000 widgets. 2. Fixed interval schedules provide a reward for the first correct response occurring after a specified amount of time and the interval of time that must pass before a correct response is rewarded stays the same. An employee could receive a bonus for the first widget produced after 40 hours on the job. 3. Variable interval schedules also provide a reward for the first correct response occurring after an interval of time has passed but the amount of time that must pass is kept unpredictable. The schedule is built around a mean amount of time that must pass, but the amount of time that must pass varies randomly across reinforcements. 4. Variable ratio schedule. A reward is given after one or more correct responses with the number of correct responses required for a reward varying around a mean number of correct responses. One example of a variable ratio schedule is a slot machine. It is programmed to pay off for the gambler after a predetermined mean number of attempts with the actual payoffs occurring on a random basis around this mean. Gamblers know that they will eventually win if they keep plugging coins into the machine, but they have no idea of how many coins it will take http://www.youtube.com/watch?v=wfraBsz9gX4 http://www.youtube.com/watch?v=tMMNkxxXVKI http://www.youtube.com/watch?v=tMMNkxxXVKI TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 188 before they get a payoff. The casino is required by law to pay off a certain percentage of the money put into the machine by all gamblers using the machine but they program the payoff so that an individual gambler never knows how many coins are needed or how much of a payoff will occur. Anyone who has visited Las Vegas and observed people playing the slots has seen the power of variable ratio schedule. Indeed, we know from the research with nonhumans that once the connection between the response and the reward is established, the variable ratio schedule produces a more permanent, sustained high level of performance than any of the other four schedules continuous reinforcement and fixed schedules, once the connection is learned (Ferster & Skinner, 1957). When training Fido to fetch, the most effective approach is to start by providing a positive reward for every occurrence of the desired behavior (i.e. continuous reinforcement). Eventually Fido gets the idea and knows that something good is going to happen if he gets the stick and returns it. If Fido continues to receive a reward every time he fetches, it is likely that Fido would slack off eventually, most likely because he is stuffed full of treats. Once the linkage is established, then the reinforcement schedule should switch to a partial schedule. If a fixed ratio schedule is used, where every fifth fetch is rewarded, Fido is likely to get excited as he approaches the fifth fetch but then an immediate slacking off (Fido might be “thinking” …I have to do this four more times so I might as well take my time). It is also possible to use a fixed interval schedule in which Fido is rewarded for the first fetch after each 10 minutes of fetching passes. Here Fido’s enthusiasm increases as he gets closer to the end of the ten minutes, but then after getting his reward, he immediately slacks off. After all, Fido might think, “why do I need to fetch the stick at all until the ten minutes is almost up?” A much better reinforcement schedule is the variable interval schedule where you reward Fido for the first fetch after a period of time but you randomly determine the time around some average amount of time. So one might choose ten minutes as the average of amount of time that must pass but the specific amount of time varies around this average. With this schedule, one would get a lot more motivated behavior on the part of Fido. Even better is the variable ratio schedule. In this case, Fido is first given a treat after fetching the stick three times, then after five times, and then after seven times. Assuming there were only three trials, the average number of fetches required for the reward is five ((3+5+7)/3). Of course, the number of trials typically is much larger and the average fetches required for a reward is computed across all these trials. Any one instance of reinforcement will vary considerably around the average. A variable ratio schedule keeps Fido guessing and on his paws. So he is observed panting with excitement after each throw and asking in doggy language, “Will I get the treat this time”? If the laboratory research generalizes to the field, supervisory praise should achieve an optimal effect if administered on a variable ratio schedule, which means that reinforcement (e.g., praise) is dispensed irregularly. If the supervisor was attempting to increase safe behavior by a worker using a variable ratio schedule of 5, he or she might praise the employee one time after seeing four examples of safe behavior, another time after seeing nine examples of safe behavior, still another time after seeing two examples TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 189 of safe behavior, and so on across all the safe behaviors observed, for an average of one praise per five safe behaviors. A variable ratio schedule establishes a connection between the desired behavior and the consequence and also keeps people on their toes (and Fido on his paws) because they never know when to expect the reinforcement. As already noted, most of the research showing that partial (or variable or intermittent) schedules achieve more sustained levels of high performance was conducted in the laboratory and much of that research with rats, monkeys, and other nonhumans. Little is known about the potential utility of using the different types of schedules with humans in organizations, and the small amount of organizational research that does exist is inconclusive (Hamner 1991). Some laboratory research shows that a variable ratio schedule is associated with higher performance than continuous reinforcement schedules (e.g., Saari & Latham, 1982). Pritchard and his colleagues (Pritchard, Hollenbeck, & DeLeo, 1980; Pritchard, Leonard, Von Bergen, & Kirk, 1976) investigated the effects of different reinforcement schedules on workers' performance on self-paced tests of job- related knowledge. Results indicated that employees who were contingently reinforced, or paid according to the number of tests they passed, performed better than those who received an hourly wage. Whether they were paid on a fixed or variable ratio schedule, however, made no difference. The missing link when applying reinforcement schedules to humans is that unlike Fido and other non-humans, humans attempt to make sense of reinforcement schedules. In the process, they form impressions about the schedule such as it is fair or unfair, achievable or impossible, and a challenge or a demeaning experience more worthy of a dog than a person. Although there are still questions about what type of schedule of reinforcement works best with workers, two general conclusions appear warranted it does appear more effective to reward employees on the basis of the number and quality of what they have done rather than on the basis of fixed time intervals. Fred Taylor noted this in his principles of Scientific Management and advocated a piece rate system of incentives in which workers were paid by the unit of work. Because the worker is paid each and every time his productivity achieves the standard, it is a continuous reinforcement schedule. The findings of research on non-humans suggests using a partial schedule in which reward only occurs after every nth occurrence in achieving the standard level of productivity with either a fixed or variable number of occurrences required for a reward. Whether the schedule is continuous, fixed, or variable, rewarding an employee in some way for what they have done rather than the mere passage of time is likely to yield better results. With a fixed or interval ratio reward system, employees can see a direct link between their performance and organizational bonuses, and the perception of this linkage is a tremendous incentive builder. Despite the recommendations of Taylor and the behaviorists, reinforcement schedules in the workplace are most often based on fixed intervals, such as paying people every two weeks, or giving a bonus every Christmas. There are some good reasons for this. As suggested in the Human Relations approach and even by some of the Scientific Management proponents, people often need the security that comes from knowing that a paycheck will arrive on a regular basis and is not totally dependent on what they do. To TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 190 the extent that productivity is affected by other factors that employees cannot control, a continuous or ratio schedule is seen as unfair and frustrating. For instance, an employee paid only for the number of units produced is not paid if the machine breaks down or some other unexpected and uncontrollable event. Lowering their pay under such circumstances could lead to less effort on the tasks, not to mention action to restore justice (e.g., a work stoppage in protest of the unfair treatment). There is also the matter of human dignity. Workers are not rats in a Skinner box! A second general conclusion, in addition to the recommendation to pay contingent on performance, is to make the consequences for meeting performance standards immediate. Whatever the schedule used in administering the positive reinforcer, negative reinforcer, or punishment, the occurrence of the behavior and the consequence is clearly linked by providing the consequence contingent on the behavior and as soon after the behavior as possible. This is a principle that is often violated such as when we punish a puppy for urinating on the floor several hours after the event. In the workplace, rewards are too often given months or even years after the behavior that is being rewarded. The lack of linkage undercuts the ability of the reward to affect future behavior. Application of BMod in the workplace Most practical applications of operant conditioning to the workplace focus not so much on the specific schedule of reinforcement as on identifying the specific behaviors that are desired and the positive reinforcers to provide to increase the likelihood of these behaviors, and then establishing a connection between the behaviors and the receipt of the reinforcers. The essential component of BMod is providing positive reinforcers contingent on the occurrence of desired behaviors. According to Komaki, Coombs, and Schepman (1991), a BMod program typically follows four steps: 1. Specify desired behavior (i.e., set a performance goal). 2. Measure the desired performance that meets this goal. 3. Provide frequent, contingent, positive consequences for meeting the goal. 4. Evaluate the effectiveness of the BMod program in improving performance on the job. Management must specify the exact behavior or behaviors that workers are expected to achieve, along with an accompanying time line. In essence, this is equivalent to setting a specific goal. If, for example, management wished to motivate its salespersons to increase their sales volume by 20% within six months, then they would need to clearly state that goal in a way that would allow a quantitative measurement of progress in achieving the goal. They would need to carefully measure sales volume to provide documentation for performance changes over time. It is very important to publicly post performance changes (in this case, sales volume) so that these changes are salient to all involved workers. Public posting provides frequent direct feedback to employees on their progress in meeting performance goals. Finally, to assess the effectiveness of the program, they would need to evaluate the performance changes quantitatively. Luthans, Paul and Baker (1981) provided a classic example of the use of BMod to TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 191 motivate employees. The researchers provided contingent rewards to sales personnel in a large retail department store to improve their work performance. Some of the behaviors identified as constituting good performance were assisting a customer within 5 seconds after he or she arrived in the sales area and keeping the shelves stocked within 70% of capacity. The experimenters observed and recorded employee behavior on the sales floor and then rewarded them with various amounts of paid time off, depending on how well the performance goals were met. Prior to this behavioral intervention specific goals were not set and the most common consequence of poor performance were threats by management. The evaluation of the experiment over a 60-day interval is illustrated in figure 4.4. The baseline period refers to a period of time during which the experimenters observed the employees prior to the intervention. After the intervention was initiated, performance increased dramatically in the experimental group and remained high even after the reward program was discontinued (the post intervention phase). Figure 4.4: Aggregate Retailing Behavior as a Consequence of Providing Contingent Reinforcement (experimental group) vs. Pre-intervention and No Reinforcement Control Group. Probably the best-known example of a successful application of a behaviorally based motivation program was the Emery Air Freight Corporation program. A focus of this program was to motivate workers to use containers properly in packaging items for shipment. The employees were shown how to properly package items and performance goals for packaging were stated in specific, quantifiable terms. Supervisors then used verbal praise as a reinforcer and performance progress was charted. Although the program was eventually discontinued, it apparently achieved some success at the time it was implemented. According to Emery, the program saved the organization $2 million in one year alone. For an account of the Emery Freight use of BMod take a look at the TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 192 following 1972 film featuring the leading proponent of behaviorism at the time, B. F. Skinner: http://www.youtube.com/watch?v=oDal--Gp2BE Implications and current trends From both the empirical research and applications in industry, it seems obvious that behavioral principles do work. Contingent reinforcement schedules, particularly fixed and variable ratio schedules, are often quite effective in motivating workers to alter their behavior. The distinction between variable and fixed schedules is not especially important when applied to humans. The important lesson of behaviorism is that management should establish the linkages between what workers do and consequences of their efforts rather than basing rewards solely on passage of time. Despite their effectiveness, behavioral programs have not been widely applied in organizations. One reason is practicality. The record keeping and paper work required to implement and maintain these types of programs are prodigious. Moreover, such programs can have a boomerang effect: If they are discontinued, performance may fall below baseline. Strict behaviorism has also been criticized for its "black box" approach to human behavior; that is, it ignores the obvious influence of cognitive and unconscious factors (Bandura, 1982, 1986). An additional concern expressed by many psychologists is that behavioral programs are unethical when they attempt to blatantly manipulate workers' behaviors (Rogers & Skinner, 1956). Social learning theory is an interesting extension of behavioral theory that recognizes that there is more than a “black box” and that people think and feel (Bandura, 1982; 1986). One of the tenets of social learning theory is that most learning results not from rewarding or punishing behavior, but from observing the behaviors of other people; this is called observational learning. For example, children learn many behaviors, such as how to cross the street, by observing older children and adults. O'Reilly and Puffer (1989) nicely demonstrated an application of social learning theory to organizational research. In both laboratory and field studies, they found that the observation of rewards and punishments affected observers' job attitudes. For example, when there was an inappropriate reward or sanction toward a fellow subject or worker, such as receiving punishment that was not deserved, motivation, satisfaction, and equity were observed to decrease. More cognitively oriented approaches to behaviorism (e. g., O’Reilly & Puffer, 1989), address one of the major criticisms of behavioral research and application, namely, the neglect of cognitive factors. The future of behaviorism in organizations clearly lies in the marriage of behavioral and cognitive principles. Manipulating the consequences of employee actions so as to reward or punish them is a powerful and proven method of changing the behavior of employees and getting them to exert more effort on the job. It's also the first question one ought to ask after determining worker goals. But this question is not enough. It is clear that an understanding of employee motivation and the use of interventions to increase motivation requires that one http://www.youtube.com/watch?v=oDal--Gp2BE TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 193 goes inside the person to understand not only the behavior and environmental consequences of behavior but also the person’s thinking and emotions. To some extent the goal-setting approach does this. The discussion now turns to three other theories that focus on the psychological processes in the form of expectations, needs, and perceptions of fairness. Points to ponder: 1. “All we need to do to ensure high levels of motivation is make sure people are monetarily rewarded for high performance.” What are the problems with this assertion? Why do pay-for-performance programs go wrong so often? 2. Skinner and other radical BMod advocates assert that we do not need to take into account human perceptions or psychological factors. What are the problems with a radical approach to BMod? How do you believe a BMod approach needs to be modified when applied to humans as opposed to critters? 3. Provide examples of the various types of operant conditioning as they might occur in the workplace. 4. The work on BMod suggests that operant conditioning in which management provides positive outcomes contingent on behaviors that are desired is more effective than management punishing behaviors that they want to discourage. What accounts for the ineffectiveness of punishment as a means of modifying behavior? 5. If punishment is so effective, why do people continue to use it so frequently and why don’t they use positive reinforcements more frequently? 6. You are irritated by a specific behavior of your friend and want to use BMod to eliminate the behavior. How would you go about doing this? What are Employees’ Expectations? Imagine the following situation. A supervisor attempts to motivate her employees to achieve higher levels of productivity in their jobs. She starts by setting specific, difficult goals as advised earlier in this chapter. She also makes sure that attempts to achieve these goals are immediately rewarded, as advised by the BMod advocates. She is disappointed to find, however, that despite the goals set and the rewards provided, employees show little effort or persistence in attempting achieve higher levels of productivity. What is holding them back and preventing them from successfully stimulating their motivation? The missing component here are factors internal to the employees that influence their motivation. The remainder of the chapter discusses these other factors, including expectancies, needs, perceptions of fairness, and the self. This section is concerned with what employees expect to occur as a consequence of their efforts. Do they expect that they can successfully achieve higher levels of productivity if they try and if they are more productive, do they expect that they will receive valued outcomes as the result of the success? Let us consider an example that is probably familiar to the readers. Students have registered for a course and are excited because the topic is one they find really interesting and their friends have recommended. They are even considering graduate study in the TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 194 area covered in the course. After attending the first class session, they are bubbling over with excitement. They savor every word the instructor utters and read the textbook chapters with care. They study many hours for the first exam because it is extremely important for them to excel in this course. Their dismay is indescribable when their grade on the first exam is a C-. Disappointed but determined, they then try harder: They listen even more closely to the instructor and study even more than before. Imagine their shock when, on exam 2, they receive another C! As exam 3 and the end of the term approach, their enthusiasm wanes. They invent excuses not to attend class and have difficulty concentrating on reading assignments. What has happened? The students began with a great expectancy of success; they believed that the appropriate effort (in the form of attending class and studying hard) would pay off in superior test performance and an A in the class. They also believed that a good grade would increase the probability of their being accepted into graduate school. As the term progressed, however, the expectancy that their effort would result in high performance decreased, and, consequently, the probability that the course grade would enhance their chances for admission to graduate school decreased. It is obvious that their motivation to expend effort on the course decreased during the term. A prominent theory of work motivation that could explain their decreased motivation is valence- instrumentality-expectancy theory, often abbreviated as VIE theory. This theory incorporates three key components in predicting and explaining the choices people make among goals and the vigor and persistence with which they pursue these goals. The components are (1) the expectancy that effort will lead to some level of performance, (2) the probability that this level of performance will lead to some valued outcome or outcomes, and (3) the valence or attractiveness of the outcomes. Basic components and predictions of VIE theory Although several versions of this theory exist, Victor Vroom (1964) drew from an earlier theory of work motivation called path-goal theory (Georgopolos, Mahoney, & Jones, 1957) to propose what has become the Valence-Instrumentality-Expectancy (VIE) theory. VIE theory is fairly complex; however, the fundamental idea behind it is simple. The theory is based on the notion that, before expending any effort to achieve some outcome, people ask themselves, "What is in this for me?" If as illustrated in the example, they cannot see the association between their effort to reach a goal and the attainment of desired outcomes for reaching that goal, they simply will not choose to pursue the goal and if they do, they show a low level of vigor and persistence. More than any other motivation theory, VIE theory has a well-developed and quantified theoretical framework. Expectancy is the perceived likelihood that the effort expended will result in some desired level of performance. The expectancy discussed in the previous example is the perceived likelihood that attending class diligently and studying hard will result in superior test performance. Expectancy is expressed in terms of a subjective or perceived probability ranging from 1 (you have no doubt that you can achieve the level of performance that you want to achieve, e.g., an A) to 0 (under no circumstances can you TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 195 achieve the level of performance that you want to achieve). The determinants of expectancy include, to name only a few possibilities, the self-esteem of the employee, past success experiences on the task, communications from others about one's chances of succeeding, and the actual situation (e.g., the difficulty of the task). All other things held constant, a person will work harder to succeed on a task the higher that individual’s expectation that effort will lead to success. To boost expectancy (see figure 4.5), one could provide success perhaps by first setting goals that are relatively easy to boost employee confidence and then gradually escalating the difficulty of the goals. One could also work to remove barriers in the situation to successfully achieving goals by providing training, resources, and designing the tasks to clarify the strategies that lead to success. These changes in the situation will often require that peers and supervisors communicate to the employee these strategies and ways of overcoming barriers. Trying to raise expectancies becomes a much more difficult task when the employee’s personality is the source of low expectancies. For instance, low expectancies that effort will lead to success are often the result of feeling generally unworthy or incompetent and increasing task expectancies may require boosting the employee’s self-esteem. The chapter covers this type of expectancy in a later discussion of the role of self-concept in work motivation. Figure 4.5: Sources of Expectations That Effort Will Lead to Good Performance Valence refers to the desirability or undesirability of some outcome (see figure 4.6). In the previous example, the positively valent or desired outcomes were an A in the course and admission to graduate school. Although the example did not include negative outcomes, VIE theory does account for negatively valued outcomes, such as dismissal or flunking out of school. Although a variety of different scales have been used, valence ratings of outcomes are often made on a scale from + 10 (very positively valent) to -10 (very negatively valent); a rating of 0 indicates no valence or preference. All other things held constant, the employee will work harder to succeed the more positive the valence of TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 196 performance success (i.e., the more attractive, pleasant, positive, etc.). But where does valence of performance success come from (see figure 4.6)? An important source of valence is what the employee perceives as the linkage between performance success and other positive outcomes. The employee asks if I achieve the goal of higher productivity will this lead to increased salary, recognition from my supervisor, promotion, a feeling of a job well done, etc. To the extent that employees expect positive outcomes to result from a successful performance, a successful performance is perceived as attractive (i.e., positively valent). The key intervention here is to make sure the employee perceives that achieving goals results in immediate, contingent rewards. This is consistent with what a BMod approach would recommend, but one often cannot rely solely on the objective reinforcement schedule to accomplish this. In addition, one often must communicate clearly to employees that success will bring rewards. In work situations managers need to consistently reward employees who do well and make visible to all employees how hard work has paid off. The discussion of the theory could stop at valences and expectancies for most practical purposes, but VIE theory goes a step further and proposes a third component, instrumentality, to explain the valence associated with a successful performance of the task. In VIE theory instrumentalities are used to explain the positive and negative valences associated with outcomes and refer to Figure 4.6: Making Sure that Employees Value Task Success correlations among outcomes. If a student in a course considers an outcome of A to have a high positive valence (let’s say a 10 on a scale of -10 to +10), what is the origin of this high valence? VIE theory answers this question by exploring how instrumental making an A is to achieving still other outcomes. For one student, an A is seen as positively instrumental to other positive outcomes such as the admiration of his fellow students and TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 197 the praise of his parents. The same student could perceive an A as negatively instrumental (i.e., getting the A will make less likely) to other negative outcomes such as flunking out of school. Vroom originally conceived of instrumentalities as correlations that range from + 1 to -1. As is the case with any correlation, a value of 0 indicates no relationship, a value of + 1 indicated a high positive relationship (where getting a higher grade is perceived as very likely to get one into graduate school), and a value of -1 indicates a high negative relationship between outcomes (where getting a higher grade is perceived as leading to a lower likelihood of getting into graduate school). A negative instrumentality between grades and getting into graduate school is unlikely but it is not unusual to find a negative instrumentality between getting higher grades and other valued outcomes. An example of a negative instrumentality is the perception that getting high grades in the course will lead others to perceive you as nerd and thereby limit your social life. Most researchers subsequent to Vroom have decided in the interest of simplicity to ignore negative instrumentalities and focus solely on positive instrumentalities. If only positive instrumentalities are considered, then the metric for instrumentality is exactly the same as expectancy (i.e., they range from 0 to 1). In motivating employees to higher levels of effort, management will want to increase expectancies that effort will lead to successful performance and the valence of success. To increase the valence of success management will want to make sure successful performance is positively instrumental to achieving other highly valued outcomes. Classical and human relations approaches often assume that the valence of outcomes is the same for all employees. Classical theorists assumed that money is positively valent to the same extent for all employees. For human relations theories, recognition from peers and supervisors and belong to a group were positively valent to the same extent for all employees. VIE theory recognizes that there are individual differences and to motivate employees we must make sure that successful performance on the job is linked to outcomes that are actually valued. According to Vroom's formulation of VIE theory, the three components interact in a multiplicative fashion such that the amount of effort exerted to accomplish a goal is the product of the expectancy that effort will achieve the goal and the valence of achieving the goal. Vroom called the product of expectancy and valence Force (F), which is the motivational force to strive to achieve a goal or a level of performance. An example of two people is provided in table 4.1. Susan and Joe are two salespeople deciding if they want to devote their time and energy to selling insurance policy A or the new insurance policy B. Both employees consider the same outcomes in making their decisions, and both consider the same behavioral options (i.e., work to sell policy A or try to sell policy B). If we apply the within-person model as originally conceived by Vroom (1964) and perform the arithmetic operations on the valences, instrumentalities, and expectancy, it is obvious that Susan will choose to exert more effort to sell the new policy B, and Joe will choose to push policy A. It is important to recognize, however, that the within-person model does not necessarily yield the same prediction as a between-person application of the model. This is discussed in the next section. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 198 TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 199 Table 4.1: Illustration of VIE Theory TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 200 Research testing VIE theory How has VIE theory held up under empirical investigation? Since Vroom published his book in 1964, I/O psychologists have conducted a lot of laboratory and field research using actual employees and students to test the theory. Most of the early studies only weakly supported the theory's major assumptions. By the early to mid-1970s researchers had grown somewhat disenchanted with VIE theory (Campbell & Pritchard, 1976; Mitchell, 1974). One criticism was that the theory was unnecessarily complex and does not reflect true decision processes (Stahl & Harrell, 1981). VIE theory assumes people are rational decision makers and will seek to maximize their outcomes. The chapter on the history of I/O discussed how decision theorists such as Herbert Simon and James March questioned the assumption that people are maximizers who seek to optimize their outcomes. A more realistic assumption is that people have limited cognitive capabilities, often are guided by unconscious motivations, and seek to satisfice rather than maximize (Locke, 1975; Pinder, 1984; Staw, 1977). Other researchers claimed that either instrumentalities, valences, or expectancies alone account for the observed motivational effects and that including the products of these components does not add to the prediction of work motivation (Jorgenson, Dunnette, & Pritchard, 1973; Lawler & Suttle, 1973; Schmitt & Son, 1981). Based on these and other criticisms, it is clear that VIE theory is not enough to predict and explain work motivation. However, a renewed respect for VIE theory has emerged from research identifying several moderator variables that define the conditions in which the predictions most likely hold. Between-subjects vs. within-subjects research designs. A lot of the previous research testing VIE theory used a between-subject research design in which the correlation was computed between performance on the job of each individual employee and the individual’s Force component. Vroom (1964) originally used the theory to explain how one person chooses among alternative courses of action rather than using it to compare how multiple people work toward one course of action. The within-subjects approach examines the process of individuals choosing among different behavioral options and is a more appropriate test of Vroom's (1964) VIE theory. For example, in the example we just discussed, Joe and Susan individually consider which of two alternative actions to pursue. Should they continue to sell policy A or should they work harder to sell policy B? Many researchers ignored the within-subjects emphasis of Vroom and instead predicted how different employees compare in their behavior with regard to single behavioral options. This between-subjects version of VIE is illustrated in figure 4.7. The focus is on sales of the new policy (policy B) and questions about expectancies, valences, and instrumentalities for that particular behavioral option. Because Joe has more motivational force (higher E x V) than Susan to sell policy B (6.40 > 4.68), a between-subjects VIE
theory would predict that he would show more motivation in attempting to sell policy B
than Susan. Research using a between-subjects approach has yielded results that are
much weaker than the results of research using a within-subjects approach (Pinder, 1984).
Below you will find an illustration of an application of the between person expectancy
model and another illustration of the within-person expectancy model (see figure 4.8).

Figure 4.7: The Between Person Expectancy Model of Motivation

Figure 4.8: The Within Person Expectancy Model of Motivation

Use of standard vs. individualized lists of outcomes. Related to the second criticism is the
questioning of the typical practice by researchers of presenting their subjects with a list of
standard outcomes (money, promotions, etc.) and asking them to rate the desirability
(valence) of each outcome. Standard lists do not necessarily capture the important
outcomes for each person. For instance, Joe (in the example) may value none of the
outcomes that he is given to rate. Subsequent research showed that when subjects were
allowed to generate their own lists of outcomes, the theory received more support (Matsui
& Ikeda, 1976). Using only desirable outcomes also seems more effective than using both

desirable and undesirable outcomes (Leon, 1981). The findings suggest that having
people try to calculate the relationship of the behavior to both negative and positive
outcomes is simply too complex of a calculation for the average person.

Use of performance rather than motivation as the dependent variable. In much of the
research testing VIE theory, the primary dependent measure is performance on the task or
job. On a simple task where performance is a direct function of effort, the researcher
could use performance as a surrogate measure of motivation. However, in the field
research with more complex tasks and jobs, it is erroneous to necessarily assume that
higher performance reflects higher motivation. As we noted earlier in this chapter,
performance is not the same as motivation and is a function of motivation, ability, and
knowledge. Indeed, if high motivation is associated with stress and anxiety and the task
is complex, it is possible that high motivation is detrimental to performance.
Consequently, much of the research testing VIE theory may have underestimated the
support for the theory by using performance rather than motivation as the dependent
variable.

Individual differences. There are differences among people in the extent to which a VIE
theory predicts their work motivation. Reviews of the research (Miller & Grush, 1988)
have shown that some individuals are more affected by internal factors, such as
expectancies, and others by environmental factors, such as social norms (influence from
friends, family, and teachers). As predicted, people were affected less by their personal
expectancies to the extent that they were focused on what other people consider
appropriate and inappropriate. To the extent that they were unconcerned with social
norms, their behavior conformed to a greater extent to the predictions of VIE theory.

Implications

The implications of valence, expectancy, and instrumentalities for practice are fairly
obvious. Pinder (1991) recommended several actions that management could take to
implement VIE theory:

First, managers should make sure that their employees expect that they can achieve
whatever standards of success are set on the tasks they perform. This requires selecting
employees who have the ability to perform these tasks, training employees so that they
have the skills and knowledge required on the tasks, and designing the work environment
so that there are adequate supplies, tools, and facilities.

Second, managers should offer rewards that employees positively value (valence).
Although management may perceive promotions and job transfers as suitable rewards for
superior performance, not all employees may share managerial perceptions of
desirability.

Third, managers should forge the link between performance and positively valent rewards
(instrumentalities). Employees must also perceive this linkage. Management assumes that
employees see the association between their performance and valued rewards, which is

often an unfounded assumption. Management must make sure to define for employees
the instrumentalities between performance and valued rewards rather than assume that
employees already perceive them in the way that management intends.

In addition to these interventions, we would add another suggestion that follows directly
from VIE theory. Management should communicate information on expectations,
instrumentalities, and valences to make sure that all these interventions are perceived in a
way that facilitates the motivation of employees. For instance, it is not enough to provide
valued rewards contingent on meeting work standards. Management should convince
employees that their efforts will pay off with valued rewards. A basic assumption of VIE
theory is that it is workers’ perceptions and beliefs that have the most immediate impact
on motivation.

Current research on VIE theory has corrected many of the faults of earlier studies and
subsequently demonstrated more support for VIE theory predictions (Pinder, 1984).
People often do make personal behavioral choices after weighing their expectancies,
valences, and instrumentalities. As discussed later in this chapter, VIE theory along with
goal theory serve as the core of an integrative approach to work motivation.

Points to ponder:

1. How are VIE theory and BMod similar in their approaches to increasing motivation?
How do they differ?
2. Compare and contrast the within-person and the between-person versions of VIE
theory. Provide examples of each one.
3. Why does the research provide more support for the within-person version of VIE
theory than the between-person version?
4. Imagine that you were to use VIE theory to predict the grades received on the next
exam in a class. How would you use VIE theory to make the prediction? What would you
measure and how? How would you use the measures to make a prediction?
5. VIE theory assumes that a situation in which expectancies are high and valences are
low is equivalent in the impact on motivation to a situation in which expectancies are low
and valences are high. Is this always the case? Why or why not?

What do Employees Need?

The discussion so far leads to three recommendations for motivating employees. First,
specific, difficult goals lead to higher levels of effort to the extent that the person is
committed to the goal. Second, providing desirable rewards contingent on the
accomplishment of goals leads to higher motivation. Third, motivation increases to the
extent that the person has high expectancies that effort will lead to the accomplishment of
goals and that the accomplishment of goals, in turn, is instrumental to achieving desirable
outcomes and avoiding negative outcomes. Although useful, these recommendations beg
the questions of just why people are more committed to some goals than others and
consider some outcomes more desirable than others. Work motivation theorists have
posed these questions and, in the attempt to determine why people seek some outcomes

and avoid others, have posited needs as central psychological constructs.

A need is basically a category of goals possessed by the individual that direct, energize,
and sustain attempts to achieve the goals defined by this category. For instance, one
might have the goal of eating a big, juicy hamburger for lunch. Driving the choice of the
goal and the attempts to find and eat a burger is a need for food. The need is more
abstract than the specific goal and is fulfilled to varying degrees by other goals falling
into the need category. If one fails to find a nice juicy hamburger, perhaps a pizza will
accomplish the same thing. The need for food is physiologically based but most of our
needs are psychologically based and part of our personality. When describing a friend,
one often uses words like aggressive, domineering, competitive, nurturing, and sociable.
These are personal descriptors or characteristics that psychological theorists believe are
related to basic human needs. For instance, aggressiveness, dominance, and
competitiveness could reflect needs for power, whereas nurturing and sociability could
reflect a need for affiliation. Everyone possesses these needs but people differ in the
importance they place on each need and the impact on behavior. For example, if Fred has
a greater need for affiliation than Bob, then Fred will exert more effort and show more
persistence in the search for other people with whom to interact than will Bob.

Figure 4.9 describes the homeostatic model of needs and the cycle of events associated
with need driven behavior. The cycle begins with a discrepancy between a goal and the
current state of the person. This discrepancy leads to frustration of an internal need
associated with the goal. The frustration of the need leads to an arousal that could occur
as anxiety, anger, and a host of other negative emotions. The person seeks to reduce the
negative feelings by reducing the discrepancy that exists between the current state and the
goal. The more motivated the person, the more the person chooses actions that achieve
the goal and the more persistent and energetic he or she is in the attempt. Once the goal is
achieved, the negative emotions subside until another goal discrepancy is encountered
and the cycle begins anew.

Two caveats are worth mentioning at this point. First, the homeostatic model describes
people as seeking to reduce tension and negative feelings associated with goal frustration.
Motivation theorists also posit cycles in which needs lead the person to increase rather
than reduce tension. Theorists such as Abraham Maslow proposed that this type of cycle
describes the effects of self-actualization on behavior. Another caveat is that the cycle not
only occurs as conscious attempts to achieve goals but also as unconscious and relatively
automatic processes. People are not always aware of their own needs and the effects of
these needs on how they behave.

Figure 4.9: A Homeostatic Model of Needs

Murray’s taxonomy of needs

Henry Murray (1938) pioneered work on need theory in the 1930s. Some theorists would
distinguish between needs and wants. What we need is supposedly what we must have
for the sake of our survival and well-being. A want is something a person would like to
have but is not absolutely necessary. As the Rolling Stones sang:

“You can’t always get what you want
But if you try sometimes you just might find
You just might find
You get what you need”

There is no doubt that some of the biological needs such as the need for food, water, and
air must be satisfied to some degree for the person to survive. However, Murray believed
that there were psychological needs that were also important. Even though one might not
die from their absence, they could have a motivational impact as large as any biological
need.

Murray proposed more than 20 human needs (see the list of needs in table 4.2).

Table 4.2: Henry Murray’s Need Taxonomy

Viscerogenic needs are associated with physiological functioning and included food,
water, sex, urination, defecation, and lactation. Psychogenic needs are mostly learned and
are organized into the following categories: materialistic needs, ambition needs, affection
needs, power needs, and information needs. Murray differed from earlier theorists in his
assertion that psychogenic needs are learned, not inherited. Additionally, he believed that

environmental cues activate psychogenic needs. For example, people with a need for
affiliation only seek companionship when other people are available. Murray believed
that all people possess the same basic needs, although in different amounts. For example,
Sally and Susan both have a need to achieve, but Sally’s need to achieve is greater than
Susan’s. Each person is unique in the relative importance of these needs. Most of the
categories of needs are further divided into those that are positive and those that are
negative. The positive needs push the person in the direction of obtaining the object
related to the need (e.g., air is a need that pushes the organism to breathe and thereby
obtain air). The negative needs push the person in the direction of creating distance from
the object or experience (e.g., harmavoidance is a need to avoid physical pain).

Murray said little about needs within the context of work motivation, but his ideas and
measurement techniques greatly influence work motivation theories. His conception of
human needs was particularly important in the theories of Abraham Maslow, Clay
Alderfer, and David McClelland.

Maslow’s need hierarchy

Abraham Maslow (1954) proposed a theory of human motivation that grew out of the
humanistic movement in psychology. By observing clients in his clinical practice,
Maslow determined that five types of needs motivate people: biological, safety or
security, social, ego, and self-actualization needs.

Maslow proposed that these needs are arranged in a hierarchy of importance: Higher
order needs are not important until lower order needs are satisfied. For example, if a
person is hungry or thirsty, he or she is not as concerned with fulfilling needs for
friendship and protection from harm until the hunger or thirst is satiated. Maslow thought
that, most people are more successful in fulfilling their lower order needs than their
higher order needs; that is, the needs for food and sex are more often satisfied than the
need for relationships with other people. However, people do not need to completely
satisfy their lower order needs before they are motivated by higher level needs. Some
threshold in the satisfaction of lower level needs is required before higher order needs are
activated. Given that this threshold is met, lower level needs can continue to motivate
people at the same time that their higher order needs become increasingly important in
influencing their behavior.

In Maslow’s theory, the physiological, security, social, and ego needs all operate
according to a homeostatic principle. If they are frustrated there is tension and negative
affect followed by attempt to reduce the frustration. Once the person achieves goals that
satisfy these needs, the tension and negative affect subsides and the person is at peace and
no longer driven. In contrast to these needs, self-actualization, or the need to realize one’s
potential in life, are never entirely fulfilled. In essence, Maslow conceived of self-
actualization as a lifelong growth process. An artist who strives for self-actualization,
according to Maslow, does not need to become a Picasso or Rembrandt, but only the best
artist he can be. This self-improvement process continues throughout a person’s lifetime
(Maslow, 1943; 1954).

Maslow proposed that once individuals reach the level of self-actualization
on the need hierarchy, they keep climbing and never arrive at a state of complete fulfillment.

Maslow never envisioned his theory as a theory of work motivation, but some very
prominent psychologists have used the need hierarchy to understand the motivation of
people at work and in organizations. Douglas McGregor (1960), in his article The Human
Side of Enterprise, is one of these theorists. One implication of the need hierarchy for
work settings is that management must determine where workers are located on the
hierarchy. If workers are primarily concerned with their safety needs at work (for
example, job security), Maslow’s theory suggests they are less likely to respond to
attempts to satisfy their esteem needs (for example, recognition or status). Because at any
one point in time individuals may differ in the relative importance of their needs on the
need hierarchy, Maslow’s theory also implies that each worker may require something
different to satisfy his or her needs. Still another implication of the need hierarchy is that,
if management can fulfill a worker’s physiological and security needs, they should next
address the ego and self-actualization needs. As a consequence, the need hierarchy has
been cited as the impetus for diverse organizational change efforts, from the creation of
pleasant work environments and employee-oriented supervision to participative
management programs and job enrichment.

Maslow’s need hierarchy was popular with many of the human relations theorists that we
discussed in module 2. McGregor (1960) introduced the concepts of Theory X and
Theory Y management philosophies. Theory X describes a task-oriented, directive
management style such as proposed in scientific management where the only concern is
with motivating employees through money and the threats of punishment. Theory Y
describes an employee-oriented, participative management style. A Theory Y manager
would attempt to determine the unsatisfied needs of his or her subordinates (in Maslow’s
terms, their location on the need hierarchy) and then attempt to satisfy those needs.
According to McGregor, organizations can only hope to obtain maximum productivity
from workers to the extent that needs are taken into account. Many of the human relations
theorists such as McGregor suggested that a Theory X approach probably worked just
fine at the turn of the century when the majority of workers were most concerned with the
bottom rung of the Maslow hierarchy and seeking to gain a livable wage. With rising
affluence, the passage of laws guaranteeing security to workers, and the increasing
education level of workers, they became more concerned with the higher level needs. The
basic proposal of McGregor and others was that management styles needed to change
from Theory X to Theory Y to fit the needs of workers in the mid-20th century and
beyond.

Despite of its intuitive appeal, the research attempting to test Maslow’s theory has found
only mixed support. Factor analyses of the five proposed dimensions of the need
hierarchy have shown these dimensions are not independent or even separable (Payne,
1970; Wahba & Bridwell, 1976). Also, little support has been found for the predictions
for the effects of satisfaction or frustration of one level of need on other levels. As you
recall, the need hierarchy theory requires that, as a need is satisfied, it becomes less
important and the next higher need assumes more importance. Experimentation with
humans in which their needs are frustrated or fulfilled are neither practical nor ethical.
Consequently, the research tests have consisted mostly of survey studies examining how
need satisfaction is related to need importance over time. The hypothesis that follows

from Maslow’s theory is that satisfaction of a need at time 1 is negatively related to rated
importance of that same need at time 2. Also, it is hypothesized that satisfaction of a need
at time 1 is positively rated to the importance of the need at the next higher level. The
longitudinal tests of Maslow have yielded little support for this prediction (Hall &
Nougaim, 1968; Lawler & Suttle, 1972; Rauschenberger, Schmitt, & Hunter, 1980).

Alderfer’s existence, relatedness, growth (ERG) theory

The research has not yielded strong support for Maslow’s need hierarchy theory. In the
attempt to modify and improve the model, Alderfer (1969) proposed the existence-
relatedness-growth theory or ERG theory. Alderfer altered the need hierarchy by
adopting a three-level classification system and abolishing the concept of a rigid
hierarchy. The three categories are: l) existence needs, which subsume Maslow’s
physiological and safety needs; 2) relatedness needs, which include Maslow’s social
needs; and 3) growth needs, which subsume Maslow’s ego needs and self-actualization.

Unlike Maslow, Alderfer believed that people could move up and down this hierarchy
and are motivated by more than one need simultaneously. For example, individuals who
fail to obtain a coveted job promotion by taking on extra assignments at work may decide
to spend more time on socializing while still striving for a future promotion. In this case a
frustration of growth needs devotes more time to fulfilling both growth and relatedness
needs. It is also possible they may decide to spend all their time attempting to fulfill
relatedness needs, an option that Alderfer calls frustration/regression. As a consequence
of thwarted attempts to achieve, the person regresses down the hierarchy and focuses on
lower needs. The few empirical tests provide mixed support for ERG theory (Alderfer,
1972; Scherf, 1974; Wanous & Zwany, 1977). Undoubtedly, the failure of Maslow’s
theory to hold up under empirical scrutiny has discouraged research on any type of needs
hierarchy. This is particularly unfortunate, given that Alderfer addressed many of the
deficiencies of Maslow’s hierarchy.

McClelland’s three need theory

David McClelland and his colleagues (Atkinson, 1964; McClelland, 1965) narrowed the
list of needs to three that they consider especially important from personal, social, and
organizational perspectives: need for Achievement (nAch), need for power (nPow), and
need for affiliation (nAff) Unlike Maslow and Alderfer, these theorists did not propose a
hierarchy of needs. Instead, they described people as possessing different profiles on the
three needs. They proposed that each constitutes a trait that is modifiable. They also
proposed that each operates according to a different set of dynamics.

Need for achievement (nAch). According to McClelland, our competitive and success-
oriented society values and reinforces the need to achieve; friends, supervisors, and the
media constantly encourage us to excel, to be “on top.” Unlike some of the earlier
theorists who viewed needs as innate and unchanging characteristics, McClelland
believes that the achievement motive is a learned need that we acquire from our parents.
At first glance, the implications of the achievement motive for personal and

organizational functioning are very impressive: The need to achieve would logically
seem to have an impact on the type of job or career you desire, the level of job or career
performance you attain, and the attitudes you hold about that job or career.

McClelland and his colleagues defined the essential dynamic at work in the need to
achieve as behavior oriented toward competition with a standard of excellence. The
method of measuring the achievement motive has relied primarily on a projective test, the
Thematic Apperception Test, or TAT. An individual taking the TAT is shown a variety of
pictures and asked to tell a story about these pictures. A need for achievement score is
calculated by determining the number of times competition with a standard of excellence
is mentioned in the stories. In figure 4.10 a high need for achievement person might tell a
story in which the two women are working night and day to find a cure for a deadly
disease and are on the brink of a great discovery. A person more oriented to power or
affiliation might focus less on task accomplishment and on the relationship between the
women, feelings of isolation and loneliness, or competition for power.

Three salient characteristics define the emphasis of high-need achievers on task
performance. First, people high in the achievement motive take personal responsibility
for solving a problem or performing a task. Second, they require concrete feedback about
their performance. Only with performance feedback can the high-need achiever improve
poor performance or bask in the glow of successful performance. Third, they set
moderately difficult goals rather than very difficult or easy goals. The rationale here is
that moderately difficult goals are challenging but still attainable.

Some research suggests that the high need achiever is most motivated there is about a
50/50 chance of success. Consequently, McClelland’s hypothesis for the effects of goal
difficulty appear to conflict at first glance with Locke’s goal-theory hypothesis that the
more difficult the goal, the higher the performance. If one considers the acceptance and
commitment to the goal, however, the two theoretical positions are really compatible. As
long as the person accepts and is committed to achieving a goal, increasingly difficult
goals appear to lead to increased effort, as predicted by Locke and associates. However, it
is often the case that commitment and acceptance eventually decline at very high levels
of goal difficulty. This tends to support the achievement motivation predictions that
motivation is at its peak in the vicinity of a .50 probability of success.

McClelland and his associates have developed a training program to enhance the
achievement motive. They claim to have increased the need to achieve in many groups of
people who participated in these programs. The training program consists of four major
components:

1) Obtain information on the achievement motive and its importance for success.
2) Set high performance goals and maintain a record of performance accomplishments.
3) Try to think of yourself as a high achiever and a winner.
4) Seek group support from other people who are (or who are trying to become) high
achievers.

The training program’s most well-known success story involved business managers in
India (McClelland, 1962). Several months after experiencing the training program, the
business managers doubled the amount of their performance-oriented, and, therefore
success-oriented, activities. These results concur with McClelland’s contention that the
achievement motive is increased through training, and that groups of people or whole
cultures can become more achievement oriented. A moment’s thought, however, reveals
that the achievement motive training program contained elements other than need
achievement, specifically goal-setting and social support. Any one or a combination of
other factors may have produced the desired outcome. Consequently, finding that the
training program works does not confirm the theory or justify the conclusion that
increased nAch alone can account for the outcome.

Look at the picture. Your task is to write a complete story about the picture you see
above. This should be an imaginative story with a beginning, middle, and an end. Try to
portray who the people might be, what they are feeling, thinking, and wishing. Try to tell
what led to the situation depicted in the picture and how everything will turn out in the
end. Be sure and examine the picture for several seconds before clicking the button
below. Once you push the button, you will have 10 minutes to write whatever story comes
into your mind. You should write continuously the entire 10 minutes. If you need more
than 10 minutes, feel free to continue writing until you are finished.
Source: http://www.utpsyc.org/TATintro/

Figure 4.10: Example from the Thematic Apperception Test

Need for affiliation (nAff). The need for Affiliation (nAff) refers to the desire for warm
interpersonal relationships and interaction with others. Evolutionary theorists suggest that
the need to affiliate with others and belong to a group is perhaps among the most primal
of needs because the survival of humans as a species required affiliation with others. In
measuring nAff with the TAT, researchers look for imagery that conveys thoughts about
being with others, loneliness, approval, and the like. For instance, in the example shown
in figure 4.10, a person with high nAff might tell a story in which the younger woman
longs for the approval of the older scientist and seeks to become her friend. People with a

http://www.utpsyc.org/TATintro/

strong nAff seek the company of others and will behave in the workplace so as to win
approval even if it hurts task performance. At the same time that a high nAff is probably
advantageous if an employee must work as part of a team, research has shown that nAff
is negatively related to success in managerial roles. One possible explanation is that
managers must take responsibility and make hard decisions despite their personal
feelings. High need for affiliation persons are more concerned that others like and accept
them and to avoid rejection might avoid making the hard but necessary decisions that are
always part of the managerial role.

Need for power (nPow). Another need with implications for work motivation is the need
for power (nPow), defined as the need to control other people. Characteristics associated
with the need for power include a desire to control other people, maintain leader-follower
relations, and influence and direct others (Cherrington, 1989). In measuring nPow in the
TAT picture presented in figure 4.10, the researcher looks for stories containing themes
that involve dominating others. For instance, the story of a high nPow person might state
that the younger woman is scheming to take over the older woman’s position. Note that
competition is part of high need for Achievement but it involves competition against
personal standards of success not competition against other people. In nPow the person
wants to compete out of a desire to beat and subjugate other people. McClelland (1970)
describes two faces of power: personal power, or striving for power for its own sake,
and social or institutional power, or striving for power for the attainment of
organizational goals. Social or institutional power facilitates organizational effectiveness
more than personal power because those individuals with a high need for social power
will work toward organizational objectives rather than personal, ego-satisfying objectives
(McClelland, 1970). Indeed, McClelland and Burnham (1976) proposed that the need for
power, particularly social power, is a far more important managerial attribute than the
need for achievement. In support of this contention, McClelland and Boyatzis (1982)
reported that a high need for power and low need for affiliation predicted managerial
success over a sixteen-year period at AT&T. By contrast, the need for achievement did
not predict managerial success as well, particularly at the higher organizational levels.
The author is not aware of any practical application of these findings, however, other than
selecting managers on the basis of their needs for power, affiliation, and achievement.

The need for personal power involves the subjugation
of others and differs from the need for socialized power.

Criticisms of McClelland’s three need theory and later modifications. How has the three-

need theory fared under the scrutiny of systematic investigation? Despite its intuitive
appeal, researchers have criticized McClelland’s work on nAch, nPow, and nAff.
Empirical evidence for the theory’s propositions is mixed (Brockhaus, 1980; Finger,
McClelland, 1976). Perhaps the weakest aspect of the theory is the reliance on a
projective test, the TAT, to measure these needs. There is some recent research using
sophisticated quantitative analyses that is more supportive (Spangler, 1992), but critics
question the TAT’s reliability and validity (Entwisle, 1972; Kraiger, Hakel & Cornelius,
1984). Intensive training of those rating TAT stories on the needs can
improve interrater reliability but this is time-consuming and costly. Critics also point out
the low correlation typically found between TAT measures of each of the three needs and
self-report questionnaires measuring the same needs. As noted in the chapter on research
methods, researchers seek evidence of convergent validity when attempting to establish
the construct validity of a measure. So, if the TAT and self-report measures are both
measuring the same need, one should find strong and positive correlations between self-
report questionnaire scores and TAT scores. Unfortunately, researchers usually report
that the correlations are low.

In later modifications of the three-need theory, McClelland and his colleagues
(McClelland, Koestner, & Weinberger, 1989) explained the low correlations by positing
that TAT and self-report questionnaires measure different facets or dimensions of the
need. Needs exist at both an implicit, unconscious level and at an explicit, conscious
level. The TAT purportedly measures the implicit dimension of the need. Self-report
questionnaire measures of nAch, nAff, and nPow measure the explicit dimension and
consist of conscious self-attributions of these needs. McClelland and colleagues
hypothesized that where there are extrinsic incentives for a behavior and the situation is
structured so that the person is aware of these contingencies, a self-report questionnaire
measure of needs relevant to the behavior will show higher correlations than an implicit
measure. Where the incentives are not obvious and the person is intrinsically motivated to
show the behavior, implicit measures such as the TAT will show higher correlations than
an explicit measure. For example, if an engineer is told that he will receive a $1000
bonus and a promotion if he is able to win a contract, his effort and persistence in the
attempt to succeed will increase to the extent that he believes that he has a high need for
achievement as reported on a self-report questionnaire. The conscious self-attribution that
“I am a high need achiever” will drive his efforts to achieve the goal. On the other hand,
if the situation is unstructured so that the engineer does not know what it takes to earn
bonuses and promotions and the situation lacks clear incentives or contingencies, his
performance is affected more by his intrinsic motivation to do well. On a day-to-day
basis, no one is standing over the engineer telling him what to do and his choices are
guided more by implicit needs that are unconscious. It is in these situations where the
TAT measure of nAch is a better predictor, according to McClelland, than a self-report
questionnaire.

There are also many situations in which the implicit and explicit needs combine to
facilitate performance and other situations in which the implicit and explicit needs
conflict and undermine performance. McClelland et al (1989) backed up their hypotheses
with supportive research, but much more research is needed before these hypotheses rise

to more than fascinating speculations and achieve the status of established theory. One
issue remains. Because they depend on the subjective judgments of those coding
responses to the TAT, this measure, along with most other projective measures, are
notoriously low in inter-rater reliability. Consequently, the low correlations found for
needs measured with the TAT and other variables may reflect on the low reliability of the
TAT more than the implicit or explicit dimensions of the need.

Practical implications of need theories

This chapter distinguishes among several need theory approaches to worker motivation,
but there are some underlying similarities among them. They all borrowed from the
original work of Henry Murray in proposing needs that they considered most important.
Although the labels attached to the needs differed, there were striking similarities among
the subsequent need theories that were derived from Murray’s work (see table 4.3).

Table 4.3: Comparison of Maslow’s Needs with the Needs
Proposed by Alderfer, Herzberg, McClelland and Hackman.

Once the needs of employees in an organization are identified, several options exist for
increasing motivation.

1. Select employees on basis of needs that fit the work. Because needs are seen as
associated with stable traits, one could evaluate applicants on the needs that we believe
are associated with motivated work behavior and hire those who are high on these needs.
Some needs are seen as relatively malleable (i.e., open to change).

2. Train employees to increase the needs that fit the work. If successful employees are
more likely to have a certain set of needs, one could attempt to increase the importance of
these needs through training or persuasion. When a sales manager gives a motivational
speech to a group of salesmen with images of financial success he or she is attempting to
increase the importance of money and thereby motivate them to pursue sales goals.

3. Change the situation to engage employee needs. Still another application of need
theory is to change the work situation. If workers already have a high need for money and
are motivated to earn more money, management could design the incentive systems in the
organization so that they can in fact earn more money by working hard. If workers have a
high need for affiliation, then management could reward successful performance with
opportunities for social interaction (e.g., parties, casual Fridays). If workers have a strong
need for self-actualization, management might reward them for working hard and
showing persistence in the pursuit of organizational goals with opportunities for self-
growth (e.g., send them to workshops, courses). In these practical interventions, it is
important to realize that the potential benefits of a strong need in increasing work
motivation and job performance are in doubt unless three other things are in place. First,
specific, moderately difficult goals must accompany the needs. Second, immediate and
contingent rewards must accompany the accomplishment of these goals. Third,
employees must expect that they can achieve the goals associated with the needs through
their efforts.

Points to ponder:

1. What is a need and does it affect human behavior in the homeostatic model?
2. Some critics of the need approach suggest that the use of the need construct is not
parsimonious in that an almost endless number of needs are proposed. Do you agree?
What is the value of a need approach to motivation and what are some of the downsides
of this approach?
3. McClelland focused on three general categories of needs. What are these needs and do
you believe they suffice in explaining motivation at work and in organizations?
4. Explain, providing examples, Maslow’s need hierarchy. How well has it been
supported in the research?
5. How did Alderfer modify Maslow’s need hierarchy in his theory?
6. Distinguish between implicit and explicit needs and the ways of assessing them.
Discuss the implications of these two types of needs for motivating employees.
7. Discuss the similarities and differences between the need approach to motivation and
the goal, VIE, and Bmod approaches.

What is Fair?

Still another piece of the motivation puzzle is the perceived fairness or justice of the work
situation. Claims that managers treat them unfairly are common complaints in
organizations and can undermine otherwise worthwhile attempts to set goals, provide
rewards, raise expectancies, and activate needs. On the other hand, fair and just treatment
of employees and distribution of outcomes can encourage workers to work with vigor and
persistence toward goals compatible with organizational objectives. This section
considers four types of fairness that I/O psychologists have identified (see figure 4.11).

Figure 4.11: Distributive, Procedural,
Interactional, and Informational Justice

Distributive fairness is concerned with whether or not employees believe that the
outcomes they receive are fairly distributed to employees. Procedural fairness is
concerned with whether the decision processes leading to these outcomes are just.
Interactional fairness is the perception of employees that they are treated with dignity and
respect. Informational fairness describes the extent to which employees believe they are
provided with adequate explanations for decisions. As we will see, all four types of
fairness are crucial determinants of the motivation to work hard and persist in the attempt
to accomplish goals (Colquitt, 2001). Example items used in measuring each type of
fairness is described in table 4.4. Sample items used for each type of justice are presented
in table 4.4.

Table 4.4: The Four Dimensions of Justice and Sample Items.

Distributive fairness

Imagine a situation in which an employee discovers that coworkers that he considered his
equals in terms of such factors as job type, tenure, and training made much higher
salaries. How would the employee react? The employee probably would respond with
surprise and anger. He might even say that “Well, because I am doing the same work as
Joe but making less money, I will simply not work as hard.” Because the employee
perceives that he and the coworkers are putting the same amount of work into their jobs
(inputs) but receiving quite different salaries (outcomes), that the situation is unfair. One
possible response to this inequity is to reduce inputs; that is, because the employee
received less, he produces less. If the employee felt that there is no resolution to the
problem, he might eventually quit his job and seek employment elsewhere.

Adams theory of distributive justice. More formally, this process is labeled a social
exchange process (Adams, 1965; Homans, 1961). The foundation for this process is that
people make investments (inputs; typically work of some sort) for which they expect
specific rewards (outcomes; often money). People typically have predetermined
expectations about how much and what types of inputs are required for certain outcomes.
They evaluate the fairness or equity of their inputs and outcomes through interactions
with relevant others in which direct comparisons between self and relevant others
determine whether the exchange is favorable. Adams’ (1963; 1965) equity theory is the
most widely used social exchange theory in organizational research, and has focused
mostly on the effects of workers’ reactions to organizational compensation decisions.

The two major components of Adams’ equity theory are inputs and outcomes (see figure
4.12). Inputs are those tangible and intangible commodities a person brings to
the exchange relationship. Examples of inputs are effort at work, training, education, and
experience. Outcomes are the products of the exchange. Examples of outcomes include
pay, bonuses, promotions, and recognition. According to Adams, people weigh the
importance of their different inputs and outcomes and sum across them. The total inputs
and total outcomes between a person (p) and a relevant other person or comparison other
(o) are compared in the form of a ratio:

A state of equity exists when the ratio of the person’s inputs and outputs is equal to the
other’s ratio of inputs and outcomes. Inequity exists when the person’s ratio is less or
greater than the other’s ratio. The following two figures visualize the process involved in
judging equity. The first, describes the process as beginning with an evaluation of one’s
own inputs and outcomes. Following this individual picks a relevant comparison other
and computes the ratio of that individual’s inputs to outcomes. The situation is perceived
as equitable if the ratio of the person is the same as the ratio of the comparison other. The
situation is perceived as inequitable if the ratio is either less than or greater than the ratio
of the other.

In the weighing of inputs and outcomes (outputs) the person could compare himself or
herself to either a specific other person or to a category of others. For example, a person

might compare the ratio of what she is putting into a job and what she is getting as
outcomes against the ratio of what she believes workers in general are putting into their
work and getting. As depicted in the next figure, the person weighs inputs and outputs
(outcomes) and then compares this ratio to the comparison other or others.

Recent discussions of the rising inequality in the United States and other countries have
focused attention on how the perception of unfairness can damage an entire society.
Some economists and other social scientists have argued that the widening gaps between
the richest in a society and the middle and lower socioeconomic classes are leading to a
variety of ills including violence, demoralization, and alienation. Based on Adams’
equity theory, one could propose that people compare their own personal situations with
society as a whole or possibly one category of persons such as the upper classes. The
waitress at Joe’s Cafe perceives that she works six days a week, 10 hours a week for less
than minimum wage while many others do a fraction of the work but earn millions. The
following links are to videos describing this issue in Asia, the UK, and the US.

http://www.youtube.com/watch?v=XjlyFmgLjtc

http://www.youtube.com/watch?v=uM_MhNEk9Ic

http://www.youtube.com/watch?v=y-lXJaEVecY

There are three especially important assumptions underlying equity theory and the equity
formulas. First, the equity formulas do not imply that the amounts of input and outcome
for the person and the other are equal, but that the relative amounts (or ratios) are equal.
For example, Sally may not work as hard as Don, but if she also receives a lower salary
than Don, then a state of equity exists. Second, the theory assumes that perceptions of
equity or inequity occur in reference to whatever comparison other or others that an
individual chooses. Employees in a company may receive very little reward for the work
they do relative to upper management, but employees who do not choose upper
management as the comparison other will not feel cheated. Third, the equity or inequity
expressed by the formulas does not imply that equity or inequity exists in any objective
sense. Sally may perceive that she does not work as hard as Don, although her supervisor
or other coworkers may disagree. Equity (or inequity) is totally perceptual and highly
subjective.

Similar to the homeostatic model presented for need theories, if a state of inequity exists
people attempt to restore equity. There is more than one way they can attempt to achieve
this. Adams (1965) described six different methods of restoring equity:

1. Alter inputs. Example: Change the quantity or quality of work. If you believe that you
are being paid less than others doing the same quantity and quality of work you might
decrease the quantity and quality of your work.
2. Alter outcomes. Example: Change the amount of compensation for work. If you
believe that you are being paid less than others doing the same quantity and quality of
work you might ask your supervisor for a raise.

http://www.youtube.com/watch?v=XjlyFmgLjtc

http://www.youtube.com/watch?v=uM_MhNEk9Ic

http://www.youtube.com/watch?v=y-lXJaEVecY

3. Leave the field. Example: Turnover is an extreme reaction to perceived inequity;
absenteeism is a less extreme reaction. If you feel that you are receiving less than is
equitable, you may quit your job and seek a more equitable situation.
4. Cognitively distort inputs or outcomes. Example: Reevaluate work so that inputs seem
less or greater than previously. Given that employees often lack information on how they
are paid relative to others, this is a very common way of restoring equity. We may simply
change our perceptions of our situation and the situation of others.
5. Try to change the inputs or outcomes of the comparison other. Example: Convince the
comparison other to change the quantity or quality of his or her work. Often there are
group norms for productivity.
6. Change the comparison other. Pick a comparison other that will make you feel better
about your ratio of inputs to outcomes. To use an extreme example, if I feel that I am
underpaid relative to inputs, in comparison to my fellow workers, I might stop thinking
about them and instead use “downward comparison” and compare myself to someone in
the workplace who is much worse off.

Research. The Adams theory of distributive justice has been tested in both laboratory
experiments and in field, correlational research. The experimental tests have manipulated
inputs and outcomes to create conditions of inequity. The effects of inequity are then
measured for the performance of very simple tasks such as the proofreading of
manuscripts. The use of simple tasks that almost everyone can perform with minimum
preparation is to ensure that differences in performance reflect motivation rather than
ability.

In addition to the experimental tests of the theory field studies have included employee
ratings of how fairly they believe important outcomes such as salary are distributed in
their organizations.

Figure 4.12: Essential Components of Adams’ Equity Theory

1. Tests of underpayment and overpayment inequity. A unique aspect of the Adams
theory is that under some circumstances, perceived inequity is hypothesized to produce
higher motivation and performance than perceived equity. In testing this nonobvious
hypotheses, researchers have used two different perceived inequity manipulations:
qualifications and circumstances. In the qualifications manipulation, subjects are told that

they are either underqualified or overqualified for the job, and that the payment they
receive will not reflect their qualifications. In the circumstances manipulation, subjects
are told that because of budgetary factors the experimenter will pay them either more or
less than promised. Regardless of the type of equity manipulation, subjects are supposed
to attempt to restore equity by either decreasing or increasing the quality or quantity of
their inputs. Psychologists have experimentally tested equity theory under two types of
compensation systems: piece-rate and hourly payment. Piece-rate refers to payment
according to the number of products produced, and hourly refers to paying the
participants for the number of hours worked. Table 4.5 summarizes the specific equity
predictions according to the type of equity manipulation and compensation system.

Table 4.5: Equity Theory Predictions of Worker
Reactions to Inequitable Payment.

Whereas the support for the overpayment predictions are mixed, support for the
underpayment predictions are much more consistent. In general, the reviews of empirical
research support equity theory predictions, particularly in the piece rate and hourly
compensation underpayment conditions (Campbell & Pritchard, 1976; Goodman &
Friedman, 1971). Not surprisingly, the findings are less consistent in supporting
counterintuitive overpayment condition predictions. Researchers have offered two
explanations for the inconsistent results in the overpayment condition (Campbell &
Pritchard, 1976; Goodman & Friedman, 1971). First, the overpayment manipulation
seems to require a threat to the participants’ self-esteem. Studies that have used the
circumstances manipulation in which they are told that budgetary restrictions prevent
them from receiving an equitable level of pay have reported less support for equity theory
predictions in the overpayment condition (Pritchard, Dunnette, & Jorgenson, 1972;
Valenzi & Andrews, 1971). On the other hand, telling participants that they lack the
qualifications for the task is a direct threat to self-esteem and seems to result in more
support for overpayment predictions. However, participants may work harder not to
restore equity (i.e., produce higher quantity and quality to justify their higher pay) but
rather to prove their capabilities and defend their self-esteem (Andrews & Valenzi, 1970;
Weiner, 1970). A second reason for the mixed results in tests of overpayment is that most
people are likely to rationalize overpayment. Workers in real organizations are rarely told
they are overpaid. A more likely outcome is that they adjust their perceptions of their

inputs and outputs so that they match. As the result of this rationalization they come to
believe they are receiving what they deserve (Locke, 1976).

To summarize the research testing the underpayment and overpayment hypotheses of the
Adams equity theory:

* The predictions of Adams equity theory are generally supported.
* The predictions for underpayment receive stronger and more consistent support than the
predictions for overpayment.
* Individual differences moderate the extent to which people respond to overpayment
inequity as predicted in the theory. One such moderator is moral maturity.

2. Other norms for distributive justice. Mowday (1991) has discussed four basic
conceptual weaknesses of Adams’ equity theory that point to directions for future theory
refinement and research. They include the need for a broader conceptualization of equity,
alternative methods of resolving equity other than changes in input, ambiguities about the
choice of a comparison other, and individual differences in equity resolution.

The “equity norm” is deeply rooted in early socialization; people especially in North
American and many European societies believe that organizations should compensate
people according to their merit. However, other norms, such as equality and
responsiveness to needs, also exist (Leventhal, 1976). If bonuses are distributed equally
to employees in an organization (a common occurrence), then the distribution of rewards
follows an equality norm. If bonuses were distributed according to need, then the
employee with six children would receive the largest bonus. Although use of a needs
norm seems rather unusual, one of the authors observed an organization that used need to
justify larger bonuses to male employees with dependents. It is obvious that the blanket
assumption that only the equity norm guides workers and organizations is naive.
Research is needed to explore when norms for equity, equality, and need are activated
and influence perceptions of fairness.

3. Cognitive distortions of inputs and outcomes. It is ludicrous to assume that people are
so simplistic that they resolve perceived inequity only through altering their inputs. In
fact, as already discussed, Adams’ (1965) original conceptualization included six
different methods of equity resolution. Unfortunately, researchers have failed to capture
the complexity of equity resolution because they have concentrated solely on behavioral
measures, such as quality and quantity of input. In a laboratory experiment testing
Adam’s equity theory, an overpayment situation was created and the effects were
observed over several sessions (Lawler, Koplin, Young & Fadem, 1968). Participants
initially resolved overpayment inequity by increasing the quality of their work, but in
later sessions, subjects enhanced perceptions of their qualifications to resolve the
inequity. Cognitively distorting inputs to match outputs is undoubtedly commonplace in
organizations. In a field demonstration of cognitive distortion, clerical workers
cognitively reevaluated their outcomes in response to underpayment inequity (Greenberg,
1989). After a pay cut, these workers increased the perceived importance of their work
environment in contributing to their payment equity. Clearly, until researchers have

explored other methods of equity resolution equity theory will remain a partially
developed theory.

4. What determines choice and impact of the referent? In most of the laboratory studies
that have tested equity theory predictions, subjects were provided with a comparison
other. The focus of the participant is limited to a referent that is assigned by means of
experimental instructions. However, in the outside world the referents are not so obvious
and are subject to a greater extent to the discretion of the employee. In his original theory,
Adams hypothesized that people will choose similar others in one’s work situation as the
basis for comparison of inputs and outcomes. The implication is that the more similar the
comparison other, the more attractive the referent and the more powerful the impact of
judged inequity. In an experimental test of this notion, researchers manipulated the
attractiveness of the referent by telling the subject that the referent other’s attitudes were
either similar or dissimilar to the subject’s attitudes (Griffeth, Vecchio, and Logan, 1989;
see figure 4.13). Consistent with predictions, participants in the overcompensation
condition were more likely to perceive their own performance as higher if they were
overpaid than if they were underpaid or equitably paid. Supposedly the self-perception of
higher performance restored equity. This only occurred if the referent was similar to
them. If the referent was dissimilar, the participants did not differ in their self-perceived
performance as a function of the level of payment they received relative to the referent.

People do not always choose other individuals in comparing their inputs and outcomes.
Goodman (1974) classified comparison others into three categories: similar others, self-
standards, and system referents. In Adams’ theory, workers were described as choosing
similar others as the individuals they would use for comparing their situation (e.g., How
does my situation compare to that of other similar workers?). These are persons who
sufficiently similar to the person to provide a comparison. Self-standards refer to
standards people set for themselves, such as comparisons of past and present work inputs.
One common example is the individual who focuses on how the current situation
compares to past situations (e.g., I’m making more now for what I do than I was last year
and how does it compare to what I personally want to obtain?). System referents refer to
implicit or explicit expectations between the worker and his or her organization, which
typically begin when the worker is hired (e.g., How much am I making in salary
compared to how my employer is doing in terms of profits and top management
salaries?). One can also distinguish between referents that are outside the work situation
(e.g., How are workers in other companies doing compared to me?) and those internal to
the work situation (e.g., How are workers in my company doing compared to me?). There
is some evidence that in evaluating the fairness of one’s pay workers are most likely to
rely on comparisons with other workers outside their organization in similar or related
occupations (Dornstein, 1988; Bygren, 2004). In a study involving Swedish workers,
Bygren (2004) found that comparisons with coworkers and one’s own past pay were of
minor importance in determining satisfaction with pay.

Figure 4.13: Self-rated Performance as a Function
of Equity of Payment and Similarity of the Referent to the Participant

The use of external referents would seem to depend on how much access the employee
has to salary information. To the extent that the organization prohibits disclosure of
salaries, employees judge the equity of their compensation by using external referents
such as the salaries in other organizations. Also, employees differ in their skill in
obtaining comparison information and those who are more educated may have more
success in finding good external referents. Goodman (1974) found that those managers
with more education tended to use external referents, or comparison others outside their
organization. The work on choice of referent suggests that testing equity theory only with
similar “other” referents in the same work situation does not capture the essence of the
equity resolution process and that researchers need to give more attention to comparisons
with external referents. Researchers also need to give more attention to the extent of
fluidity in choice of referents. Do workers stick with the same referents in judging the
fairness of their situation or do they shift what they use as a comparison over time and
situation perhaps as a way of resolving the inequity?

5. Individual differences. Given the complex judgments in which people must engage to
resolve inequity, equity theory is indeed a theory of the individual. It is therefore
surprising that little research has examined individual differences in reactions to inequity.
One individual difference variable is a person’s locus of control (Rotter, 1966).
Individuals with an internal locus of control tend to perceive events in the world as being

under their control and this belief possibly inclines them to take action in resolving
perceived inequities rather than resort to cognitive distortions of inputs and outcomes
(Mowday, 1991). Moral maturity is another possible individual difference variable
influencing how people resolve piece-rate overpayment. In one of the more interesting
laboratory experiments on this topic, people who scored high on a measure of moral
maturity were found to be more sensitive to overpayment inequity (Vecchio, 1981; see
figure 4.14). Those individuals low on moral maturity attempted to maximize their
outcomes and were not as concerned with equity resolution (see also Kohlberg, 1968).
Specifically, participants high in moral maturity behaved according to equity theory
predictions; they produced less with higher quality than equitably paid subjects. Subjects
low in moral maturity did not conform to overpayment predictions. This rather
provocative experiment underscores the potential importance of individual differences in
understanding the process of equity resolution.

Figure 4.14: Interaction of moral maturity X equity
conditions on quantity and quality of performance.

Correlational findings. So far the discussion has focused on the laboratory experiments
conducted to test this theory. There is also a large amount of field correlational research
examining the perceived fairness of the distribution of outcomes and various outcomes
(table 4.6). The results show that despite the factors shown in laboratory experiments to
moderate reactions to inequity, a variety of negative outcomes are associated with
perceptions of inequitable outcome distributions. The experimental research findings
show that it is possible for people to experience overpayment inequity and then work
hard to restore equity. These correlational findings indicate, however, that employers
should focus on making sure that the distribution of outcomes is perceived as fair rather
than attempting to motivate people by creating feelings of inequity. Perceptions of
distributive inequity are associated with lower levels of satisfaction, trust, organizational
commitment, organizational citizenship behavior, and performance. Perceptions of
distributive inequity also are associated with negative reactions such as withdrawal and
other negative reactions to their work and organizations.

K = number of correlations; N = number of participants; ruc = mean
uncorrected correlation weighted by sample size; rc = corrected
correlation; * = confidence interval for the correlation excluded zero.

Table 4.6: Correlations Between Distributive
Justice and Outcomes Found in Meta-Analysis

Procedural fairness

Distributive justice is concerned with whether the division of rewards among people is
fair. Procedural justice is concerned with whether the division was fairly decided. One
could have distributive justice without procedural justice, procedural justice without
distributive justice, both procedural and distributive justice, and neither procedural nor
distributive justice.

Attributes of fair procedures. There are norms that define whether the processes in
making decisions are fair or not.

1. Consistency. Decisions that follow consistent guidelines are judged fairer than those
that are inconsistent. Assume, as an example, two instructors set down as a rule that there
are no make-ups for missed assignments but that you can drop the lowest of the
assignments instead. One instructor consistently follows this rule whereas the other
makes exceptions. Students are more likely to perceive the instructor who follows the
rule consistently as procedurally just. A student may perceive this same instructor as
unfair if the instructor told the student that a makeup exam is not permitted. In this case,
the student feels unfairly treated, but the inequity results from a violation of distributive
justice rather than procedural justice. Consequently, workers could perceive a manager as

providing a fair distribution of outcomes but using unfair procedures in making these
distributions, or vice versa. The two-component theory of fairness, which we will discuss
later, hypothesizes that distributive justice moderates the extent to which workers react to
procedural unfairness. There is some evidence that when there is distributive fairness,
procedural justice does not influence attitudes and behavior as much as when there is
distributive unfairness. According to this research, assessment of whether outcomes are
allocated fairly is the first step and if workers perceive unfairness they proceed to
assessing procedural fairness of the situation before acting on the unfairness.

2. Justification. If the instructor provides a justification or explanation for the refusal to
allow a make-up exam this is also likely to help create a perception of procedural
fairness. In general, decision makers who provide justifications for their decisions are
seen as fairer than those who do not provide justifications. This aspect of procedural
justice is considered the core component of informational fairness.

3. Voice. This is another important factor defining procedurally fair decision processes.
By voice we mean that the employee has a chance to state his or her opinion and provide
input. It does not necessarily mean that the employee actually makes the decision. A
supervisor who shows bias against a particular group of employees is likely to undermine
procedural fairness for those who benefit as well as those who do not benefit.

4. Objectivity. Showing impartiality and objectivity are very important components of
procedural fairness. If the supervisor bases the decision on an ample amount of objective
information, the decision is seen as more procedurally fair than if the decision is highly
subjective and based on little data. One important benefit of having rules and procedures
in a bureaucracy is that employees will see the decisions as impartial rather than based on
personal likes and dislikes.

5. Availability of repeal. Still another characteristic of procedural fairness is the
availability of appeal. If an instructor does not allow students to question the accuracy
with which an exam was graded, the students are likely to perceive less procedural
fairness than if the instructor responds to their questions.

The results of a meta-analysis summarized in table 4.7 shows that perceptions of
procedural justice are beneficial across a variety of outcomes. To the extent that
employees report that there is procedural justice they are more satisfied, committed to the
organization, and trusting. They also have higher performance and show less negative
reactions.

The research on procedural justice has supported the importance of perceived fairness of
the decision processes used in distributing outcomes. As indicated in table 4.7, employees
who report that procedures are fair are more satisfied with their jobs and supervision, and
more trusting in and committed to their organizations. They are also more likely to
engage in citizenship behavior and perform their jobs effectively and less likely to
withdraw from the organization.

K = number of correlations; N = number of participants; ruc = mean uncorrected
correlation weighted by sample size; rc = corrected correlation; * = confidence
interval for the correlation excluded zero

Table 4.7: Correlations between Procedural Fairness
and Outcomes Found in Meta-Analysis

Explanations for procedural fairness effects. There are two primary explanations for why
procedural justice has all these beneficial effects and procedural injustice has such
negative consequences. One possibility is that procedural justice is the prelude to
achieving positive extrinsic outcomes. People generally feel that fair procedures will
allow them to achieve more benefits than unfair procedures. This is the so-called self-
interest theory of procedural fairness. The other explanation, the group theory, states that
procedural fairness is a sign that the employer or other decision maker positively regards
the group to which the individual belongs. For instance, allowing voice is tantamount to
recognizing that the opinions of the employee are worth listening to and that the
employee is competent enough to form a reasonable opinion.

Interactional fairness

Most of the research on justice has examined procedural and distributive justice, but two
additional types of justice appear distinguishable and influence the motivation of
employees. One of these is interactional fairness. This is very close to procedural justice
but some researchers consider it a distinct and separate basis for employee perceptions of
fairness (Bies, 2005). The reader can no doubt remember times when situations were fair
in terms of distribution of outcomes and the procedures used in making the distributions.
Yet, the treatment received was lacking in dignity, social sensitivity, and politeness. For
instance, a supervisor could hand out the yearly bonus check to employees based on an
objective, unbiased performance appraisal procedure. Distributive fairness could be
assured by making sure that the bonuses are proportional to the effort and outcomes of
the employees receiving them. At the same time the supervisor could interact with those
receiving them in a rude manner, perhaps conveying an attitude that none of them really
deserved what they were receiving. Likewise, the procedures could be lacking in
objectivity and the distributions inequitable but a supervisor could treat employees with
sensitivity and respect. Basically interpersonal justice is the more informal, human side of
justice that reflects the social skills of the supervisor, whereas distributive and procedural
justice are the formal components that are originate from organizational structure.

In addition to directly influencing employees, interactional justice also appears to
moderate the impact of procedural and distributive justice. In one study of strategic
business alliances in China, procedural justice more positively affected the performance
of the alliance when the parties in the alliance were treated with dignity and respect (Luo,
2007). Interpersonal fairness even appears to have physiological determinants. One study
found that interactional unfairness triggered the release of cortisol, a hormone that plays a
central role in the body’s response to stress (Yang, Bauer, Johnson, Groer, & Salomon,
2014). The results of a meta-analysis show that perceptions of interactional justice are
associated with several positive outcomes including higher satisfaction, commitment, and
citizenship (see table 4.8).

K = number of correlations; N = number of participants; ruc = mean
uncorrected correlation weighted by sample size; rc = corrected correlation;
* = confidence interval for correlation excluded zero.

Table 4.8: Correlations Between Interactional
Justice and Outcomes Found in Meta-Analysis

Informational fairness

Greenberg (1990) conducted a quasi-experiment demonstrating the effects of procedural
inequity on another outcome, employee theft (see figure 4.15). The employees in three
plants in a company were compared on the amount of theft that occurred over the same
time interval. In one plant, the plant manager announced a pay cut without providing
adequate justification. The manager followed a script in which he told them the decision,
gave a short reason, and then allowed very little time for questions and answers. In the
adequate justification plant, the plant manager announced the same pay cut but provided
a thorough explanation. In a third, control condition, there was no cut in pay to explain.

Figure 4.15: Greenberg’s (1990) Study of the Effects of
Underpayment Inequity and Justification on Employee Theft

Figure 4.16: Rate of Theft as a Function of the
Adequacy of the Explanation for a Pay Cut

As seen in figure 4.16, there was little difference on the amount of theft between the plant
that received a pay cut with an adequate explanation and the plant that did not receive a
pay cut. However, the plant that received a pay cut with an inadequate explanation
showed a large increase in theft during the period in which pay was cut.

The research on informational justice consists of fewer studies than the research on
distributive, procedural, and interactional justice. The findings of a meta-analysis of
studies conducted so far found that similar to the other types of justice, perceived
informational justice is positively associated with a variety of positive outcomes and
negatively related to negative outcomes (see table 4.9).

K = number of correlations; N = number of participants; ruc = mean
uncorrected correlation weighted by sample size; rc = corrected
correlation; * = confidence interval for correlation excluded zero.

Table 4.9: Correlations Between
Informational Justice and Outcomes Found in Meta-Analysis

Inter-relationships among types of justice

How do the four types of justice interact to affect employee attitudes and behavior? The
two-component model attempts to answer this question and the research so far has
provided support. The two-component model of justice suggests that distributive fairness
interacts with the other forms of justice in determining how employees respond to
perceived fairness and unfairness. According to this theory, people first assess the
fairness of procedures, the information that is provided, and their interpersonal treatment.
If these are perceived as fair, they are not as bothered by an unfair distribution of
outcomes. Where employees are most motivated to remedy injustices is where an unfair
distribution of outcomes is coupled with unfair procedures, poor treatment
interpersonally, and a lack of information (Skarlicki & Folger, 1997; Leung, Tong & Ho,
2004). When employees do not know the outcomes of others they may fall back on
cognitive heuristics in judging how their outcomes compare to others. One type of
heuristic is rooted in procedural, justice. We may assume in the lack of information on
outcomes that if the procedures or unfair and there is a lack of information provided by
management that our outcomes are also unfair (van den Bos, Lind, Vermunt & Wilke,
1997). A similar heuristic also may exist for interpersonal and information justice.

Points to ponder:

1. Describe the four types of fairness? Do you think we need all three or could we
combine some of them?

2. Adams equity theory proposes that both overcompensating and undercompensating
employees can be used to motivate employees to work hard and perform more
effectively. Under what conditions does the theory propose that each tactic could be used
to accomplish these objectives? What are the problems with using inequity in outcome
distributions to motivate employees to work harder and perform more effectively?
3. More than one type of norm exists for what constitutes a fair distribution of outcome.
Describe the different norms for outcome distribution and the conditions in which they
might be most salient.
4. If you were a manager of a group of employees what steps would you take to ensure
procedural fairness? Distributive fairness? Interactional fairness? Information fairness?
5. Describe the two-component model of justice and its implications for how the various
types of fairness might be combined in understanding and increasing worker motivation
to perform effectively.

Does Work Engage the Self?

Work can affect the way an employee thinks and feels about the self. This linkage of the
self-concept with work is a crucial motivational component in finding ways to increase
work motivation. So far the chapter has dealt with motivation as if it were a matter of
cool calculations and perceptions. In examining goals, expectancies, fairness, and
outcomes the employee was depicted as conducting a completely objective and rational
analysis of the situation and acting accordingly. An employee calculates the extent to
which working hard will fulfill important needs and the contingencies of working hard
for obtaining these outcomes. They assess the expectancies and valences for various
alternatives and chooses that option that has the best expected value. Goals are set for
future performance and the employee assesses the acceptability and commitment to the
goals. Goals that are acceptable and engage commitment lead to more effort to achieve
the goal to the extent that it is difficult and specific. The fairness of outcome
distributions, treatment, procedures, and information is similarly assessed and to the
extent that they are perceived as fair, employees invest more effort, vigor, and persistence
on their tasks. All of these processes can occur in an impartial manner in which
employees coolly size up the situation and acts accordingly. What happens when the
employee perceives that these factors have important implications for the way they define
and evaluate the self? And what are the consequences of employees’ self-concepts and
self-evaluations for their behavior and attitudes? In other words, what happens when
things become personal?

Finding meaning through self-conception and self-evaluation

People make sense of the world around them by constructing mental models of the
objects, events, and people in their environments. They are often not conscious of their
mental models, and these models are seldom sophisticated. Nonetheless, the search for
meaning is an integral component of what it means to be human. Making sense of work
and the organizations in which they perform this work are at the core of this search for
meaning. Additionally, making sense of the self and the connections of the self to work
and organizations is at the core of how they make sense of the world of work. People

construct self-conceptions in which they attribute to themselves personal attributes such
as intelligent, kind, hardworking, and aggressive. They derive these self-conceptions
from their experiences at work and outside work. The core of how they think about
themselves is probably set for most people at a relatively young age. People do not only
think about but also evaluate the self. They have ideas about what they consider to be the
ideal attributes and then judge the extent to which their self-conceptions are congruent
with these ideals. To the extent that they fall short, the self-evaluation is negative. To the
extent that they meet or surpass the ideal, the self-evaluation is positive. Self-evaluations
and self-conceptions form around specific, situationally based attributes. For instance,
people may form a self-conception and self-evaluation with regard to how well they
perform a specific task in their work. Salespeople might see themselves as very good in
face-to-face attempts to persuade clients but as not very good at keeping the office
organized. People also form more general self-conceptions and self-evaluations. At the
most general level are the entire collection of self-conceptions used by people to define
themselves and their global self-esteem. Global self-esteem is how individuals feel about
themselves in general. Low global self-esteem is a feeling of unworthiness,
incompetence, and failure, whereas higher global self-esteem is a feeling of worthiness,
competence, and success. The importance of the specific self-conception affects the
extent to experiences relevant to these situationally based self-conceptions affect global
self-esteem.

Theories of self-enhancement and self-maintenance

What are the consequences of the self-concept and self-evaluation for work motivation?
Previous research and theory have emphasized the upside. People are described as
exhibiting higher levels of motivation to achieve organizational goals if they work in
organizations where they can fulfill the need for a positive self-evaluation through their
work activities. Numerous theorists have brought attention to universal needs to enhance
and maintain self-esteem. When organizations create conditions in which these needs are
fulfilled and employees can feel good about themselves, high levels of motivation and
performance follow. Also, work that allows the expression and maintenance of valued
self-concepts benefits health, life satisfaction, and well-being. The emphasis in much of
organizational theory has been on the benefits of high self-esteem in the form of increase
persistence and vigor with which people engage in work activities and the more accurate
perceptions they form of themselves and their situations. In an example of this self-
enhancement position, Shamir (1991, p. 416) proposed that:

“general job motivation will be enhanced to the extent that:
1. Job-related identities are salient in the person’s self-concept.
2. The job offers opportunities for self-esteem enhancement.
3. The job offers opportunities for increased self-worth.
4. Actions required on the job are congruent with the person’s self-concept
or can be performed in a way which is consistent with the person’s
self-concept.
5. Career opportunities on the job are congruent with the person’s
possible selves.

General job motivation will be lower to the extent that:
1. The job or its content contain elements that are detrimental to the
person’s self-esteem.
2. The job or its context contain elements that are detrimental to the
person’s self-worth.”

In another example, Sheldon and Elliott (1998) propose in their self-concordance theory
that employees exhibit higher intrinsic motivation and more satisfaction to the extent that
they are assigned goals that reflect their valued self-conceptions and “express enduring
interests and values” (p. 482). They propose that when people pursue goals for self-
concordant reasons they show more vigor and persistence than those working for
controlled reasons. Moreover, those successful achieving self-concordant goals
experience more well-being and positive psychological outcomes than those achieving
goals for controlled motivation. Those with initially high global self-esteem are more
likely to pursue self-concordant goals than those with lower self-esteem. By contrast,
goals are pursued for controlled reasons when external pressures and extrinsic rewards
pressure require the person to pursue the goals or when goals are pursued to avoid shame,
guilt or anxiety. Those with initially low global self-esteem are more likely to pursue
goals in the attempt to prevent negative outcomes. Correlational research using a self-
report measure of self-concordance provides some support for these hypotheses (Judge,
Bono, Erez & Locke, 2005; Sheldon & Elliott, 1998).

At the same time that providing positive work environments allows workers to enhance
their self-esteem and benefits the performance of organizational tasks, work
environments that threaten self-esteem can have negative consequences. Based on the
assumption that people have a basic need for self-esteem, Argyris (1962) argued that low
self-esteem persons and people having to cope with work environments that threaten their
self-esteem have little psychological energy to invest in the performance of their work
roles because most of their energy is spent defending against further loss of self-esteem.

Self-related constructs

Support for the self-enhancement position is provided in research showing that more
positive self-evaluations are associated with higher motivation, satisfaction, and
performance. These self-evaluations include self-efficacy expectations, global self-
esteem, organizationally based self-esteem, and core self-evaluations.

Self-efficacy expectations. A self-efficacy expectation is the personal belief of persons
that they can execute the behaviors necessary to perform tasks or goals. The concept of
self-efficacy expectation was introduced by Bandura (1982) in his re-evaluation of the
radical behavioral approach and BMod. He proposed that people did not mechanically
modify their behavior in response to outcome contingencies. Humans think about the
outcomes of their actions and part of this thinking is evaluating their capability of doing
what they need to do to obtain outcomes. The measures of self-efficacy have varied
across studies but Bandura recommended the use of a unipolar measure of confidence

that ranged from no confidence to complete confidence. There is some evidence that this
is the preferred approach (Sitzmann & Yeo, 2013).

In an example of a study examining the relation of self-efficacy to performance, graduate
students in business performed a managerial decision making task (Wood & Bandura,
1989). The participants rated on a ten-point scale (1 = no confidence at all to 10 = total
confidence). The strength of self-efficacy for performance of the task was the sum of the
nine confidence ratings. At the beginning of the simulation, half the participants were
randomly assigned to a condition in which they were told that performance on the task
reflected a basic cognitive ability that people do or do not possess. The remainder of the
students were assigned to a condition in which performance on the task was described as
an acquirable skill. Over task trials participants who believed the task was an ability
demonstrated a progressive decline in self-efficacy whereas those told that it was an
acquirable skill demonstrated an increase in self-efficacy. The correlation between self-
efficacy and performance was .30 during the second block of trials and increased to .49 in
the third and last block of trials.

The meta-analyses of the relation of self-efficacy to performance show that more positive
self-efficacy expectations are indeed positively related to performance (Stajkovic &
Luthans, 1998; Sitzmann & Yeo, 2013; see table 4.10). Positive self-efficacy
expectancies apparently provide an advantage by encouraging more focus of attention
and effort on effective task strategies whereas low self-efficacy expectations encourage a
self-focus that diverts attention away from the task and depletes cognitive resources that
could be devoted to effective strategies (Stajkovic & Luthans, 1998).

Despite finding an overall positive correlation between self-efficacy and task
performance, there are some important caveats. For one, when the correlation is assessed
using a between-person approach, the correlation is much higher than when a within-
person approach is taken. In a between-person approach, the correlations are computed
by measuring self-efficacy and performance for each individual participant and then
computing the overall correlation. In a within-person approach the correlations are
computed by measuring self-efficacy and performance across task trials for an individual
and then computing the correlation for that individual. The within-person approach is
more sensitive to the dynamics that occur over time and provides a more accurate picture
of the relation of self-efficacy and performance. When examined as a within person
process, the relation of self-efficacy to performance is found to vary over task trials and is
not always positive. People set goals, form self-efficacy expectations for achieving these
goals, show varying levels of effort and persistence in striving to accomplish their goals,
monitor the discrepancy between their performance and their goals, and subsequently
modify their behavior. Although the overall mean of the correlation between self-
efficacy and performance is positive, there is a mix of negative, null, and positive
correlations.

In their meta-analysis, Sitzmann and Yeo (2013) also found that when past performance
of individuals in within-person approach were controlled, the effect of self-efficacy on
future performance was minimal. They conclude that the correlation between self-

efficacy and performance using the within-person approach is spurious and that “self-
efficacy is primarily a product of past performance rather than a driver of subsequent
performance” (p. 558). In other words, the reason that people with self-efficacies perform
better than those with low self-efficacies is that high self-efficacy persons have been
more successful in the past. Moreover, they argue that “hiring people with high self-
efficacy or boosting self-efficacy may not generate any return on investment because
self-efficacy accounted for practically no variability in performance after controlling for
potential confounds” (p. 562).

K = number of correlations; N = number of participants; ruc = mean uncorrected
correlation weighted by sample size; rc = corrected correlation; * = confidence interval
for correlation excluded zero.

Table 4.10: Correlations between Self-Efficacy Expectations and Performance

Global self-esteem and organizationally based self-esteem. Global self-esteem is usually
measured with a scale such as the Rosenberg self-esteem scale. In measuring the
person’s evaluation of his or her total worth as a human, items are posed such as “All in
all, I am inclined to feel that I am a failure” and “I feel that I have a number of good
qualities.” The reader can take this scale and compare the results to test norms by going
to:

http://personality-testing.info/tests/RSE.php

In an attempt to measure self-esteem that is specific to the feelings of worth associated
with experiences in the organization for whom an individual is employed, Pierce,
Gardner, Cummings and Dunham (1989) developed the Organization-Based Self-Esteem
scale. This is a ten-item measure asking respondents to evaluate their value, worth, and
effectiveness in the organizations in which they work. Respondents are to keep in mind
the messages they receive from managers and supervisors and state the extent to which
they agree or disagree with each of the following questions:

I count around here There is faith in me.
I am valuable I am taken seriously
I am important I am cooperative.
I am helpful I can make a difference
I am efficient I am trusted

http://personality-testing.info/tests/RSE.php

The basic theoretical proposition is that organizational practices convey messages to
employees about their worth, and these messages shape how employees think about
themselves as employees. To the extent that they have positive organization-based self-
esteem, their motivation to work hard and their satisfaction with their jobs increase. A
meta-analysis of the research using Global (or general) Self-Esteem and Organization
Based Self-esteem (OBSE) found that both are positively related to a variety of
organizational factors and outcomes, but that organization based self-esteem is more
strongly related than global self-esteem (Bowling, Eschelman, Wang, & Kirkendall,
2010; see table 4.11).

K = number of correlations; N = number of participants; ruc = mean uncorrected
correlation weighted by sample size; rc = corrected correlation; * = confidence interval
for correlation excluded zero.

Table 4.11: Results of Meta-Analysis of Correlations
Between Self-esteem and Other Variables.

Employees who perceive more job complexity, autonomy, support, psychological
ownership, and supportive, structuring, and considerate leadership and less work-related
stressors (e.g., role ambiguity) and job insecurity, also reported higher organization based
self-esteem. In turn, higher organization based self-esteem was positively related to job
satisfaction, performance, organizational citizenship behavior, organizational
commitment, and physical and mental health, and negatively related to intentions to
turnover and actual turnover. Of course the reader should heed the usual warnings about

correlation is not causation. Whether self-esteem is the cause, the effect, or the covariate
of the other variables cannot be concluded from these correlations.

Core self-evaluations. The self-efficacy, global self-esteem, and organization-based self-
esteem research focuses on self-evaluation of worth, competence, and other valued
attributes. A somewhat different approach is taken in the core self-evaluation research.
Judge and his colleagues propose that self-esteem is only one of several personality
dimensions and that a higher-order construct subsumes both self-esteem and several other
dimensions. Core self-evaluations are defined as the “bottom-line evaluations that
individuals hold about themselves” and include self-esteem, beliefs about internal locus
of control, emotional stability, and generalized self-efficacy (Judge & Bono, 2001, p. 80).
Each of these four component traits is considered a separate psychological construct, but
the authors hold that they all contribute to the higher order or cardinal trait of core self-
evaluation. A scale to measure core self-evaluations was constructed that taps each of
these component four traits contained these items (Judge, Erez, Bono & Thoresen, 2003).

1. I am confident I get the success I deserve.
2. Sometimes I feel depressed (reverse scored).
3. When I try, I generally succeed.
4. Sometimes when I fail I feel worthless. (reverse scored)
5. I complete tasks successfully.
6. Sometimes, I do not feel in control (revere scored)
7. Overall, I am satisfied with myself.
8. I am filled with doubts about my competence (reverse scored)
9. I determine what will happen in my life.
10. I do not feel in control of my success in my career (reverse scored)
11. I am capable of coping with most of my problems.
12. There are times when things look pretty bleak and hopeless to me. (reverse scored)

The research using core self-evaluation measures tends to support the general hypothesis
that more favorable perceptions of the work environment and more favorable work
outcomes are associated with employee reports of positive core self-evaluations. Table
4.12 provides a summary of the results of a meta-analysis (Chang, Ferris, Johnson, Rosen
& Tan, 2012). Employees with more positive core self-evaluations are more likely to
perceive the organization as fairer and more supportive, as providing more autonomy and
skill utilization, and as imposing fewer stressors. Core self-evaluations are also positively
related to potential outcomes including satisfaction, commitment, intrinsic motivation,
organizational citizenship, and task performance. Finally, core self-evaluations are
negatively related to counterproductive work behavior, turnover intentions, and strain.

Not everyone is as sanguine about the construct of core self-evaluations (Chang, Ferris,
Johnson, Rosen, & Tan, 2012; Chen, 2012). The critics argue that the theoretical basis
for the notion of core self-evaluations is weak and confusing. Core self-evaluation is an
aggregate construct consisting of separate and independent constructs. The reader will
encounter other aggregate constructs in this text (e.g., emotional intelligence, employee
engagement). The combining of independent variables and the treatment of them as if

they represent the same thing is always problematic. Theory is needed to identify the
causal paths between each component construct and core self-evaluations and the
potential mediating role of core self-evaluations. There is also confusion over whether
other traits should be included in addition to beliefs about locus of control, emotional
stability, self-esteem, and generalized self-efficacy. Extraversion, negative and positive
affectivity, optimism, and other variables would seem to meet the criteria set forth by
Judge and colleagues for inclusion. Still another criticism is that the higher order factor of
core self-evaluations reflects response biases such as common method variance (see the
chapter on research methods) rather than a substantive psychological construct.

K = number of correlations; N = number of participants; ruc = mean
uncorrected correlation weighted by sample size; rc = corrected correlation; *
= confidence interval for the correlation excluded zero.

Table 4.12: Results of Meta-analysis of Correlations
Between Core Self-evaluations and Outcomes

Limitations on self-enhancement

Most of the research and theory using self-concepts and self-evaluation as constructs tend
to hypothesize that more is better. In other words, more positive self-concepts and self-
evaluations are hypothesized to yield almost unlimited benefits to both the organization
and the employee. This is a naïve position. For one, people desire to do more than
enhance self-esteem. They also seek to verify and maintain their self-conceptions. Some

have even argued that there is a basic need for self-consistency that may lead low self-
evaluation persons to reject positive outcomes and intentionally fail to confirm their
negative self-conceptions (Dipboye, 1977; Korsgaard, 1996; Swann, Griffin, Predmore &
Gaines, 1999). At the very least, the effective performance of a work role requires some
realism in self-assessments. To believe that one possesses more of positive attributes than
warranted by the evidence is likely to lead to task strategies that doom the employee to
failure. Take, for example, employees who comes to believe that they have very high
levels of task ability and are confident that future success is ensured. Such employees
may not be particularly bothered by failure and may even rest on past laurels. They may
not attempt learn from past failures or take efforts to improve. The negative consequences
of having self-evaluations are probably just as damaging as the negative consequences of
having unrealistic low self-evaluations.

The research on self-concept and self-evaluation provides support for designing work
environments that engage the self. Designing a work environment in which employees
can express and develop those personal attributes that they hold as ideals seems likely to
energize their efforts to achieve success in their work roles. It is important to note,
however, that the engagement of the self-concept can have downsides in the form of
distortions of the task and personal attributes. At the same time that organizations should
promote positive self-images, they need to (1) provide diagnostic and objective feedback
on performance to prevent distortions in perceptions of the work role and one’s own
knowledge, skills, and abilities, (2) set task goals to allow a realistic assessment of
progress, (3) provide clear and accurate information on job requirements, and (4)
encourage nonthreatening and constructive dialogues among employees so that they can
learn from previous performances.

Are Employees Intrinsically Motivated?

So far the discussion has focused on how to motivate through setting goals, providing
positive consequences, shaping expectancies for what will occur if the employee works
hard, making sure outcomes of hard work are positively valued by relating them to needs,
and fairly treating employees. Suppose that all these things have been implemented and
have succeeded in boosting the effort of employees. There is one more, very important
question to ask: Are the employees intrinsically motivated? There is an important
distinction between motivation that is internally driven and continues even in the absence
of external inducements such as the carrot and the stick. One can get a donkey to move
by offering a carrot and then whacking the animal with a stick if it fails to move. One
could also get a human to do something or not do something through positive and
negative reinforcement and punishments. But do these inducements work in the long run?

There are two perspectives on intrinsic vs. extrinsic motivation. Some practitioners and
theorists say that intrinsic rewards set in motion an entirely different motivational state
than extrinsic rewards. Among the first theorists to propose this distinction was Abraham
Maslow. He believed that intrinsic motivation is essentially what occurs when people
reach the top rung of the needs hierarchy and are motivated mostly by self-actualization
needs. In contrast to the lower level needs which operate according to a homeostatic

process and driven by the attempt to reduce deficiencies, the more people satisfy self-
actualization needs, the more they want. The other perspective is radical or strict
behaviorism. This approach proposes that there is one underlying motivational dynamic
and that apparent differences between intrinsic and extrinsic motivation are simply a
result of differences in reinforcement schedules. What we call intrinsic motivation, in
which people are driven to persist and exert effort in the absence of obvious rewards, is
simply the result of an intermittent or variable reinforcement schedule. Witness the
pigeon pecking for bugs under the leaves on the ground in the forest. The pigeon will
continue with this activity for long periods of time without finding a bug. Is this intrinsic
motivation? Is the pigeon pecking for the pure joy of pecking? A behaviorist
interpretation is that they are under a variable ratio reinforcement schedule. Similarly,
people at a slot machine in Las Vegas keep pulling the lever in the absence of reward. It
is the intermittent reinforcement schedule that is responsible for the high levels of vigor
and persistence, not some intrinsic satisfaction of pulling the lever.

This chapter takes the position that in attempting to understand human motivation in the
workplace, intrinsic and extrinsic motivation constitute two separate processes and are
more than the reinforcement schedule at work in the situation. Maslow’s theory has
already been discussed, so focus of attention is on three other theories in which the
distinction between intrinsic and extrinsic motivation is a central focus: Herzberg’s two-
factor theory, Hackman and Oldham’s job characteristics theory, and Deci’s cognitive
evaluation theory (also called self-determination theory).

Herzberg’s two-factor theory

Herzberg, Mausner, and Snyderman (1959) conducted an empirical study of what people
reported as sources of motivation in their jobs, and the results led them to the two-factor
theory. The researchers interviewed over 200 engineers and accountants about their jobs;
these workers were asked to recall job-related incidents that were associated with
especially good and bad feelings. These data indicated that good feelings were associated
with such job-related factors as achievement, responsibility, advancement, recognition,
and work (task) activities. Bad feelings were frequently associated with working
(environmental) conditions, supervision, salary, job security, organizational rules and
practices, and interpersonal relationships at work.

On the basis of these data, Herzberg proposed the motivator-hygiene theory, or as it is
more commonly known, two-factor theory. Two-factor theory assumes that everyone has
two types of needs, hygiene needs and motivator needs. Hygiene needs include factors
extrinsic to the work itself. These include providing status, security, good relationships
with supervisors, subordinates, and supervisors, salary, work conditions, and the
company policy and administration. Motivator needs include intrinsic factors, such as
opportunities for advancement, increased responsibility, personal growth, achievement,
recognition of successes, the challenging nature of the work itself, and opportunities and
achievement. The key points in Herzberg’s two-factor or motivator-hygiene theory were:

1. Satisfaction and motivation are synonymous. Satisfaction leads to motivation and
motivation leads to satisfaction.
2. Dissatisfaction differs from satisfaction and is not simply the lower end of the scale.
Dissatisfaction is an entirely different dimension than satisfaction and the factors
contributing to each differ.
3. Managers need to reduce dissatisfaction by providing hygiene factors (i.e., good pay,
security, good relationships, good work conditions, good policies and administration).
4. Managers need to motivate workers by providing motivator factors (i.e., opportunities
for growth, increased responsibility, recognition of successes, challenging work,
opportunities for advancement).

To hear Herzberg himself describe his theory, check out this old, but interesting video at
the link provided below. Note that Herzberg essentially equates motivation to intrinsic
motivation. He labeled extrinsic rewards or incentives as “hygienes” and proposed that
they lead to “movement” (or KITA, a “kick in the ass”) not motivation. He labeled
intrinsic rewards or incentives as “motivators” and proposed they lead to intrinsic
motivation, which in Herzberg’s view was the only genuine type of motivation. We will
not use his terminology, but keep in mind his insights into intrinsic motivation.

Herzberg’s two-factor theory is really just an extension of the need theories that were
discussed earlier in this module. According to Herzberg, hygiene needs operate according
to a type of homeostatic model. When hygiene needs are not fulfilled, the worker is
dissatisfied. When hygiene needs are fulfilled, the worker is not satisfied but only “not
dissatisfied.” The fulfillment of the hygiene needs produces a state of neutrality, not
satisfaction. For example, if your garbage is not collected, you will undoubtedly become
dissatisfied. However, the fact that your garbage is collected is unlikely to make you
happy. Your attitude toward garbage collection is probably best described as a neutral
state that is disrupted only by the absence of garbage collection.

When motivator needs are fulfilled, the worker is satisfied; when they are not fulfilled,
the worker is not satisfied. According to Herzberg, the presence of hygiene factors will
reduce dissatisfaction but will only bring the employees to a state of “not being
dissatisfied.” They are not satisfied or intrinsically motivated if their hygiene needs or
met but will instead start expecting even more of the same. The point at which
dissatisfaction is eliminated keeps escalating so that providing hygienes never leads to
true satisfaction or intrinsic motivation. Similar to Maslow, Herzberg believed that it is
only through providing “motivators” in the form of such things as responsibility and
opportunities for growth that workers achieve a true state of satisfaction and intrinsic
motivation.

Although Herzberg’s two-factor theory has generated much enthusiasm over the years,
empirical research has not supported the major tenets of the theory. For one, the research
has shown that the motivator and hygiene factors are related to both satisfying and
dissatisfying situations (Dunnette, Campbell, & Hakel, 1967; House & Wigdor, 1967).

Also, critics have questioned the manner in which Herzberg, et al collected their data.
Herzberg used a data collection procedure called critical incidents; that is, he asked
workers to recall especially good and bad job-related incidents and then explain why they
felt the way they did. Because people prefer to present themselves in a favorable light,
they might attribute good incidents to themselves or their efforts (recognition,
achievement) and bad incidents to other factors in the work environment (supervision,
working conditions). Some of the critics of Herzberg’s theory have attributed the original
findings of Herzberg, Mausner, and Snyderman (1959) to this type of defensive bias
(Ondrack, 1974).

Given the lack of empirical support, one may question whether two-factor theory has had
any real impact on organizational research and practice. Two-factor theory’s lasting value
lies in the fact that it focused attention on the influence of motivators, such as
achievement and responsibility, on worker attitudes. Prior to Herzberg’s work,
organizational interventions had focused entirely on hygiene factors, for example, pay
and the work environment. Therefore, like the Hawthorne studies that preceded it, two-
factor theory made a lasting contribution despite its apparent flaws.

Self-determination theory

A second theoretical perspective on intrinsic motivation that has received more empirical
support than two-factor theory is Deci’s self-determination theory (also known as
cognitive evaluation theory). Deci proposed that not only do extrinsic rewards fail to
induce intrinsic motivation but they can actually hurt intrinsic motivation.

Does money destroy intrinsic motivation? Much of the attention in the original research
on Deci’s theory focused on the effects of one particular extrinsic reward…. money. The
attempts to link pay to performance are too numerous to enumerate and the failures of
these attempts are almost as numerous. There are many reasons that pay-for-performance
plans fail, not the least of which are the failures to properly measure performance and to
reward cooperation. Although linking pay to performance seems to make a lot of sense
from the managerial perspective, it is a risky approach to motivation. One potential
reason for the failure of pay-for-performance schemes is that it may destroy the intrinsic
motivation to perform the task. Suppose a supervisor lavished an employee with praise,
saying that his work was excellent and that he was a model employee. He would probably
start to envision the new stereo and clothes he would purchase with your hard-earned
raise. Imagine his surprise when the boss tells him that because he was such an excellent
worker, he was not being recommended for a raise. In response to his bewildered “Why?”
he is told that money destroys drive and incentive, so his only reward is praise.

Does this little scenario seem ridiculous? Probably, but some motivation theorists would
agree with the supervisor. According to Deci (1975) and his colleagues, the conditions
promoting intrinsic motivation are at odds with the current practice of rewarding
motivated behavior with money or similar extrinsic rewards. Drawing on the work of
deCharms (1968), Deci hypothesized that people have a need to feel that they are
competent and control their own lives. He assumed that they are intrinsically or internally
motivated to perform a task if it fulfills these two needs; under such conditions, people
will perform a task solely for the personal pleasure that the task itself provides. If,
however, people are offered external inducements, such as money, to perform an
intrinsically motivating task, they will perceive that external forces guide their behavior
and consequently lose their sense of competence and control.

Research testing self-determination theory. In a test of self-determination theory, Deci
(1971; 1972) conducted a series of laboratory experiments in which he manipulated
intrinsic motivation by giving subjects puzzles to solve; the tasks he gave were
challenging and inherently interesting. Participants were assigned to one of three
conditions. Some were paid hourly, others contingently (according to the number of
puzzles they completed), and still others were not paid at all. During the experimental
session the participants were allowed to take a break in which they were given free time
to use as they wished. Those who were originally paid per puzzle completed spent less
time working on the puzzles in their free time than the unpaid participants. Deci
concluded that, when people perceive that they are extrinsically motivated to perform a
task through external sources such as money, their intrinsic motivation to perform the
task decreases.

The attempts to replicate Deci’s counterintuitive results have led to mixed results.
Although some studies (Calder & Staw, 1975; Hamner & Foster, 1975) failed to clearly
replicate Deci’s results, others (Pritchard, Campbell, & Campbell, 1977) supported his
results. Although subsequent research has not technically refuted Deci’s original results,

the generalizability of his results to work settings is suspect. First, the differences
between reward conditions in the time spent working on the puzzles were statistically
significant, the practical significance of whether a person works a few minutes more on
an intrinsically motivating task is possibly trivial. A related issue is the short time interval
(the experimental session) subjects spent working on the puzzles. Although such tasks are
intrinsically interesting for short periods of time, they are not so appealing after hours or
days. Workers in organizations often must perform the same tasks daily for years. Third,
Deci’s findings seem to apply to those situations in which people do not expect payment,
such as a laboratory experiment, not to situations in which they expect payment for what
they do, such as the typical job (Staw, 1977). A fourth issue is that in most work
environments, management tries to motivate people who are unmotivated or extrinsically
motivated. Deci’s theory primarily concerns the effects of extrinsic factors on the
performance of participants who were already intrinsically motivated.

For all these reasons, the notion that we should not pay people contingent on their
performance because it will destroy their intrinsic motivation seems to have limited
generalizability to realistic job situations. Nevertheless, it remains alive and well in
discussions of using money to motivate students. Check out the following video that
describes a controversial proposal to pay students for grades. There is a considerable
amount of research in school settings to suggest that such schemes may actually have a
detrimental effect on student creativity.

Modifications in the theory. Deci’s self-determination theory has evolved over the years
so that the focus is no longer solely on the effects of money but any reward that lowers an
employee’s sense of competence and control. On the basis of the theory, one could argue
that if properly administered, a monetary reward system could even boost intrinsic
motivation if the rewards raise the employee’s sense of competence and control. The
theory would also argue that providing positive reinforcement contingent on successful
accomplishment of work are detrimental if it lowers the sense of control and competence.
If the supervisor reduces them to the role of a trained dog jumping for a biscuit, the
reward may negatively impact intrinsic motivation (i.e., their willingness to work without
the promise of rewards). Rewards should increase rather than lower the sense of
competence and control. This is very similar to what Herzberg said in the video that
appeared earlier about “jumping for jelly beans.” Although he framed his arguments
within his two-factor theory, Herzberg observed that this type of reward system could
diminish employee sense of control and dignity.

Job enrichment as a strategy for increasing intrinsic motivation

An important issue in discussions of how to increase intrinsic motivation is how to enrich
jobs so that they are inherently interesting and challenging. Herzberg and his colleagues
were among the first to bring attention to job enrichment as a means of increasing job
enrichment. Check out his lecture in a smoked filled room in the 60s.

Herzberg’s discussions of job enrichment were framed within his two-factor theory,
which has not received much empirical support in subsequent research. An alternative
theory, which has received much more empirical support, is the Job Characteristics
Theory of Hackman and Oldham (1976). This approach has already been discussed as an
example of a theory using individual differences in needs as a moderator of the effects of
job characteristics on motivation.

Job characteristics theory. This perspective has a straightforward assumption: Well-
designed work tasks fulfill the growth needs of workers (see Alderfer earlier in the
module), and the fulfillment of those needs motivates them. What Hackman and Oldham
called motivational potential is essentially intrinsic motivation. Motivational potential is a
high degree of vigor and persistence on the part of the employee as that employee
pursues the fulfillment of higher needs such as growth and autonomy. Hackman and
Oldham’s Job Characteristics Theory (1976) proposes that five fundamental task or job
dimensions trigger motivational potential. The core job characteristics that can lead to
high levels of intrinsic motivation are skill variety, task significance, task identity,
autonomy, and task feedback.

1. Skill variety refers to the degree to which one’s job requires the use of several skills.
For example, a manager’s job is high in skill variety because it requires many skills, such
as supervision, planning, delegating, negotiation, and decision-making.

2. Task significance indicates the importance of the work task to other people. A nurse’s
work tasks, for example, are high on task significance because of their relevance to other
people’s welfare.

3. The third dimension, task identity, refers to those characteristics of work tasks that
enable a worker to identify with a complete product or service. For example, a factory
worker who performs only one component of the manufacturing process may fail to see
the relevance of his or her individual job to the manufacture of the completed product,
such as a car.

4. A job is high in autonomy if workers have freedom or discretion in deciding how to
perform the job. The typical manager’s job is often fairly high and the factory
worker’s job fairly low on autonomy.

5. Task feedback refers to whether the performance of job tasks gives the worker
feedback about his or her work effectiveness. A teacher, for example, may not get much
task feedback in the process of teaching her students. She only becomes aware that the
students did not learn the material after grading their exams.

Hackman and Oldham hypothesized that each of the core job dimensions would
contribute to the development of three critical psychological states. Specifically, they
hypothesized that task variety, task identity, and task significance would lead to the

experienced meaningfulness of work; task autonomy would contribute to the
development of the experienced responsibility for outcomes of work; and task feedback
would lead to the knowledge of the results of work. The hypothesis is that these
psychological states are associated with positive work and personal outcomes, such as
high intrinsic work motivation, performance, satisfaction with work, and low
absenteeism and turnover. In addition, they proposed that growth need strength
moderates the effects of the core job dimensions on the outcomes. Specifically, Hackman
and Oldham proposed that the positive effects of the five job characteristics on the three
critical psychological states and the effects of the three psychological states on outcomes
are most pronounced for workers with strong growth needs. The relationships in the job
characteristics model are presented in figure 4.17.

Hackman and Oldham further proposed that the motivating potential of any job is not
simply an additive function of the five job dimensions, but rather a multiplicative
function:

MPS (motivating potential score) = ((skill variety + task identity + task significance)/3) X
autonomy X feedback

A job with a high MPS score should have more motivating potential than a job with a
lower score. From inspecting the formula, you can see that autonomy and feedback are
the most important factors in the MPS. If either is zero, the MPS is zero. However, as
long as the job has skill variety, task identity, or task significance the MPS is still
reasonably high as long as autonomy or feedback is high. The Job Diagnostic Survey
(JDS) was developed to measure the task dimensions, psychological states, and outcomes
set forth in the Job Characteristics Theory (Hackman & Oldham, 1975). The JDS is a
rather unusual instrument because the items for each component are typically scattered
across different sections of the survey, and more than one item format is used to measure
each component. The original purpose of this variability in item location and format was
to reduce response bias by forcing the respondents to carefully read each item. Some
example items from the JDS are presented in table 4.13.

Figure 4.17: Hackman and Oldham’s Job Characteristics Model of Work Motivation

Table 4.13: Examples of items from the Job Diagnostic Survey (JDS)

Research. A meta-analysis incorporating 28 studies that tested the job characteristics
model supported the predictions of the model (Loher, Noe, Moeller, & Fitzgerald, 1985).
The researchers found that the average correlation between each core job characteristic
and job satisfaction was positive and ranged from .32 (task identity) to autonomy (.46). In
a test of the moderating effects of growth need strength, the correlation between a
combination of the core job characteristics and satisfaction was .68 for workers who are
high in growth need strength and .38 for workers who are low in growth need strength.
These correlations are consistent with the model because the relation between job
characteristics and satisfaction is higher for workers high in growth need strength. Two
other meta-analyses, both employing larger samples of studies (Fried & Ferris, 1987;
Spector, 1985) also provided support for the predictions of the job characteristic model
for the relationships of the core job characteristics with satisfaction, motivation,
performance, and the mediating psychological states as well as support for the
moderating effects of growth need strength.

For a couple of short lectures on the Job Characteristics Theory check out:

DSyOpi8apF

rDSyOpi8apF

Points to ponder:

1. What are the arguments for and against distinguishing intrinsic from extrinsic
motivation?
2. Is it always important that workers are intrinsically motivated as opposed to
extrinsically motivated? What are the situations in which ensuring intrinsically motivated
workers is an important concern? What are the situations in which it may suffice to have
only extrinsic motivation?
3. Based on self-determination theory, how could one use monetary rewards to increase
intrinsic motivation?

Integrative Approaches to Work Motivation

At this point, you are probably asking which theory of work motivation is the best.
Which theory should managers use as a guide to improving the motivation of their
workers? The answer to this question is that all of them deserve consideration as useful
guides to practical action. Each theory addresses a crucial question in the attempt to
increase the vigor and persistence with which employees pursue organizational goals. In
addition, theoretical frameworks exist that can provide an integration some of the theories
discussed in this chapter. This chapter closes by considering several attempts to integrate
the various motivational perspectives that appear promising.

The motivational theories are reconcilable

Integration of the various motivation models is not possible if they represent competing
frameworks and generate conflicting hypotheses. Fortunately, many of the predictions are
not really inconsistent. Take, for example, VIE theory and goal-setting theory. At first
glance they seem to contradict one another but on closer examination the differences
between the two are reconcilable. VIE theory would seem to assume that difficult goals
produce little effort because expectancy (the effort-performance association) is low. On
the other hand, goal-setting theory would seem to predict that difficult goals lead to
higher performance.

Matsui, Okada, and Mizuguchi (1981) investigated these competing assumptions (see
figure 4.18). They assumed that difficult goals are associated with a lower expectancy of
successful performance. However, they also assumed that difficult goals would have a
higher positive valence because people value success on difficult tasks.

Figure 4.18: Means for Performance, Valence, and
Expectancy in Easy- and Hard-Goal Conditions.

Before goal- setting theory can predict motivated behavior, the valence of the goal must
overcome the low expectancy of
performance. The researchers had student subjects perform a clerical accuracy task
(number comparison) in either an easy- or difficult-goal condition. They found a
difficult-goal effect: subjects in the difficult-goal condition outperformed subjects in the
easy-goal condition. Subjects in the difficult-goal condition also had a lower expectancy
of success than those in the easy-goal condition. In addition, participants evaluated the
valence of the outcomes from goal attainment (such as a sense of achievement, clerical
ability, and persistence) more positively when they were given difficult goals than when
they were given easy goals. These results confirmed the researchers’ predictions that the
higher valence of difficult goals, if they are accepted, overcomes the low expectancy of
success. In this experiment participants were presented with one and only one goal that
we can assume they accepted.

Mento, Cartledge, and Locke (1980) suggest a two-stage process. Expectancies and
valences influence whether a goal is accepted. When participants are presented with
multiple goals to choose from, VIE theory will more accurately predict the choice of the
goal, i.e., they will choose the goal with the highest expectancy X valence. However,
after a goal is chosen and accepted, goal-setting propositions govern performance in
pursuit of that goal. Once accepted, increasing difficulty of the goal activates more effort
despite the lower expectancy of success. More recently, Kernan and Lord (1990) found
that valences and expectancies functioned differently in single-versus multiple-goal
situations. Using students who performed either a single clerical task or multiple clerical
tasks, they found that VIE variables were more influential in the multiple task (goal)
condition. The researchers reasoned that the cognitive VIE variables were more useful
when subjects had to choose which task to focus on (multiple goal), not how to perform a
single task goal.

Another apparent conflict that is reconcilable is between nAch theory and goal setting
theory. If you recall, the high need for achievement person works harder on moderately
difficult tasks than easy or impossible tasks. Again, this is compatible with VIE theory
and goal theory if one assumes that the effects occur after the high need achievement
person accepts and commits to a goal. Once acceptance and commitment are in place, the
high need for achievement person works harder the more difficult the goal and the more
specific the feedback from the goal up to the point that the goal is seen as impossible. At
that point the goal is rejected.

Integrating with VIE theory

Probably the best known of these refinements is Porter and Lawler’s (1968) proposed
integration that is depicted in figure 4.19. In this integration, equity, intrinsic vs. extrinsic
motivation, and needs are seen as influencing motivation via their impact on components
of VIE theory. The model assumes that, for effort to translate into a desired level of
performance, the person must have the ability to perform well (abilities and traits), and he

must understand demands of his job (role clarity). The model acknowledges that people
work for both extrinsic rewards, such as money and promotions, and intrinsic rewards,
such as pride in one’s work and a sense of accomplishment. The model also assumes that
the level of performance a person attains will affect the level of rewards he perceives as
equitable. Specifically, if a person expends a great amount of effort that culminates in
high performance levels, he will perceive that he deserves a substantial reward.
.

Figure 4.19: The Porter-Lawler (1968) Refinement of the VIE Model.

Work motivation as an unfolding process

One way to integrate the various approaches is to conceive of motivation as a process that
unfolds over time. The beginning point of the process is defined by the viscerogenic and
the psychogenic needs of the individual. Needs are essentially categories of goals. They
reflect stable traits of people and also vary in their importance to some extent with the
satisfaction and dissatisfaction of the needs. The relative importance of needs determines
what goals are considered. Goals that relate to unimportant needs are not given much
consideration while the focus of the individual is on choosing among the goals related to
more important needs. Mento, Cartledge and Locke (1980) suggest a two-stage process
that received support in the research of Kernan and Lord (1990). Expectancies and

valences influence which of several goals is accepted. However, after a goal is accepted,
goal-setting propositions govern performance. Thus, the specificity and difficulty of the
goal and the commitment to the goal shape one’s focus of attention, task strategies, vigor,
and persistence. Satisfaction with achieving the goals is influenced by several factors.
Satisfaction with performance increases to the extent that the outcomes achieved are
contingent on one’s goal accomplishments and lead to pleasure and the avoidance of pain.
Satisfaction also increases with the fairness of outcomes and the procedures by which the
outcomes are administered.

Satisfaction with goal accomplishment in turn affects one’s commitment to future
attempts to pursue the chosen goals. Over time as the result of continuing cycles of goal-
oriented activity and the administration of outcomes for this outcome, the person
becomes intrinsically motivated and values the activity for the sake of the activity rather
than any external consequences. This seems most likely when the person develops a
sense of competence and self-determination and feels that he or she is growing as a
person.

Consider an example from everyday life. Assume that has had her morning coffee, she
reviews several alternative goals that she can pursue during the day. One is mowing the
lawn. This task has a high probability of success (let’s say an expectancy of .9) and a low
valence (let’s say a 3). Another goal is buying groceries and this has an expectancy of .9
and a high valence of 8. Consistent with VIE theory she chooses to accept the goal of
buying groceries (.9 X 3 = 2.7 vs. .9 X 8 = 7.2). Once she has chosen the goal of buying
groceries, the characteristics of the goal kick in and become the primary motivating force.
To the extent that she set a specific, difficult grocery-buying goal (e.g., I’ll get these
specific items to stock my pantry through the next two weeks), accepts the goal, and is
committed to accomplishing the goal, she demonstrates vigor and persistence in the
pursuit of the goal. Needs are essentially categories of goals and enter into the process at
the very beginning by dictating the valence of goal accomplishment. Her low valence for
mowing the lawn might reflect the low importance she places on needs for esteem (what
does she care that neighbors think poorly of her because her lawn?). Her high valence for
buying groceries might reflect the high importance she places on social needs (She needs
to buy groceries so she can entertain her friends). The relative importance of needs can
fluctuate with circumstances, e.g., her social needs are relatively unimportant one day but
highly important another. Fairness plays a role, especially if she has a partner who is
expected to share in the chores. Some theorists suggest that the fairness of the outcomes
and procedures primarily influences how satisfied she is with the outcomes of her actions.
Satisfaction, in turn, influences the commitment to the goal and thereby affects the vigor
and persistence with which she pursues the goals in the future. For example, even though
she exerts vigor and persistence in the pursuit of her grocery buying goal, if she feels that
her partner’s job is to shop and that he or she has not been fair in sharing these duties, she
is not as committed to the task in the future. Finally, there is the question of how
intrinsically motivated she is to pursue the grocery buying goal. To the extent that she
feels that buying groceries is a task over which she has control, that she is competent in
performing the task, that the task is meaningful, and that the task allows her to grow as a
person, the task may become valued for the actions involved rather than any extrinsic

outcome. On the other, to the extent that she feels that buying the groceries is something
she must do to avoid pain (e.g., hunger, disapproval of friends, arguments with my
partner) or to obtain extrinsic rewards (e. g., satisfaction of hunger, approval of friends,
peace with my partner), she may not perform the task with the same vigor or persistence.
Once the threat of pain or the promise of reward is absent she might just order a pizza and
skip buying groceries altogether.

Control theory as an integrative framework

The concept of the negative feedback loop is at the core of control theories of human
behavior. Cybernetic theorists originally used the negative feedback look to explain how
closed mechanical systems self-regulate (Wiener, 1948). The basic components of control
theory are the input from the environment, a sensor which detects the input, a standard
against which the input is compared, a comparator which performs the comparison, a
negative feedback loop consisting of information on the extent of deviation of the input
from the standard, and the effector or entity that acts on the negative feedback to reduce
the discrepancy. The classic example is a thermostat used to regulate temperature in a
room. The thermostat contains a standard in the form of a temperature setting. The
temperature is monitored and the current temperature of the room is an input to the
system. The input is compared to the set temperature that is the standard. When the
temperature drops below the standard, a signal is sent to the air conditioning unit (the
effector) and the heat is turned on. When the temperature rises above the set temperature,
the signal to the air conditioning unit triggers cooling. Thus, the discrepancy between
room temperature and the set temperature serving as the standard is reduced. Basically a
negative feedback loop acts to maintain homeostasis and stability in the system. Control
theory also posits that in most complex systems there is s hierarchy of standards.

In applying control theory to work motivation, theorists view the worker as a closed
system who acts to reduce discrepancies between standards and inputs (e.g., Klein, 1989).
Needs, goals, values, norms for fairness (e.g., need, equity, equality), and ideal self-
conceptions are standards. There are numerous potential standards that could govern their
behavior, and the ones that become standards depend on their salience and importance to
the employee. Take, for example, a person who sets a goal to produce 15 units each week
and actually produces 13. The discrepancy between the goal and the number produced
constitutes a negative feedback loop and sets into motion actions to reduce the
discrepancy. Consequently, the person exhibits vigor, persistence, and choices that have
the effect of increasing the number of widgets and reducing the discrepancy from the
standard. What about an employee who produces more than 15 units? In a theory of work
motivation that adheres strictly to control theory this would also constitute a discrepancy
and would trigger a negative feedback loop. The employee would slack off to bring
performance back in line with the goal.

Control theory applications to human behavior usually posit a hierarchy of standards. At
the highest and most abstract level are the general beliefs, ideals, and values such as
incorporated in the ideal self-concept. Each of the lower, subordinate standards in the
hierarchy is concerned with an aspect of the higher standard. Take, for example, a person

who holds an ideal self-concept as a leader. At the next level there are standards that are
necessary to becoming a leader such as persuasiveness, vision, and organization. Each of
these standards have subordinate standards governing more specific enactments of the
higher level standard. Being persuasive means giving an effective presentation at next
week’s executive committee meeting that changes the opinions of others on the
committee, and successfully getting the employees who reports to you to buy into your
decisions. Each of these standards govern even more specific actions such as putting the
powerpoint together for the presentation, writing a talk, and setting the agenda for the
meeting. Each higher level standard specifies more concrete goals at the next lower level.
Not all standards are salient at any one time. The level in the hierarchy that is salient
determines which standard is most influential.

There are criticisms of using control theory as an integrative framework to explain work
motivation. Locke (1994) complains that “The present author has never understood the
appeal of control theory…Perhaps the machine metaphor gives it a “scientific” aura, but
the aura, upon closer examination, seems largely illusory” (p. 18). A major criticism is
that it is basically a homeostatic model that assumes that humans seek to reduce
discrepancies. Humans often seek to increase discrepancies despite the tensions
associated with these discrepancies. Examples are the high need achieving person who
seeks challenge and risk and the self-actualizing person who is not content with the status
quo but seeks growth experiences. Another criticism is that control theorists inevitably
must incorporate traditional motivational constructs to adequately explain work
motivation. People do not simply sense discrepancies and then act to reduce the
discrepancies, they perceive events, attribute causes to these events, experience a wide
range of emotions, and assess expectancies and valences of outcomes. Klein (1989)
incorporates all of these into his control model of work motivation. But one is left
wondering what is new. Control theories of work motivation fail in providing integrative
frameworks and more typically present a patchwork of ideas borrowed from other
motivational theories. Advocates of a control theory approach have attempted to counter
many of these criticisms. They argue that behavior that appears to be aimed at increasing
discrepancies can be explained by incorporating a hierarchy of standards. Acts to move
beyond a standard may reflect the salience of higher level, more abstract standards. As
such, they may represent attempts to reduce discrepancies from a higher level, more
abstract standard in the hierarchy of standards. Other proponents have posed strong
arguments for the value of control theory as a useful meta-theory that can pull together a
variety of motivational theories. The author has always struggled with the use of control
theory. I am skeptical that it advances our understanding of work motivation, but future
research may prove otherwise.

Points to ponder:

1. It is clear that the motivational theories do not necessarily conflict and that there is
room for reconciliation among them. What are these points of reconciliation?
2. Describe how VIE theory could be used as a framework for integrating the various
approaches.

3. Describe the unfolding process approach to integrating the various work motivation
theories.
4. What is the control theory approach to integrating work motivation theories? What are
the criticisms and potential advantages of this approach?

Conclusions

The chapter began with examples of how motivation is frequently used to explain the
achievements (or lack thereof) of individuals, organizations, and entire countries.
Motivation is often seen as a trait that inherent in personality and determines whether an
individual is productive in work activities. As seen in the subsequent discussion,
motivation is more complex than these commonsense notions. Contrary to everyday use
of the term, no one lacks motivation. Everyone is motivated to do something, even
though it is not always what the organization would prefer. Motivation is the choice that
employees make among tasks and goals, and they may choose to pursue organizational
goals or other goals that conflict with organizational interests. Motivation is also the
vigor and persistence with which they work to accomplish a goal. Indeed, there are traits
involved. For instance, a high need achievement person is more motivated by challenging
situations than a low need achievement person, whereas a high need affiliation person is
more likely to pursue goals with vigor and persistence when they are instrumental for
achieving close relations with others. In large part motivation is a state consisting of the
cognitive, behavioral, and emotional processes that mediate work situations and
responses to these situations.

A common distinction when talking about motivation theories is between content and
process theories. Need theories are considered content theories of motivation in that they
focus on what it is that motivates people. On the other hand, the other theories we have
discussed are process theories that have attempted to explain how, for example, needs
and rewards are translated into motivated behavior. One pervading theme across all
process theories is an emphasis on purposeful, conscious thought, typically in the form of
specific types of cognitive processes. The need theories discussed earlier tend to focus
only on what motivates workers and delve to a lesser extent into the processes by which
motivation occurs. Each of the motivation theories possesses certain weaknesses, but
each theory does an admirable job of describing motivation under certain conditions for
some people.

When attempting to understand why a person is motivated to work with vigor and
persistence on a task, the key questions to pose are

(1) What are their goals?
(2) What are the consequences of the individual’s actions?
(3) What do they expect to occur from their actions?
(4) What do they need and value?
(5) Do they perceive the work situation as fair and just?
(6) Does the work situation engage worker self-concepts and self-evaluations?
(7) Are they intrinsically motivated to pursue task activities and goals?

The psychological research on work motivation tends to cluster around theories pertinent
to each of these questions and at first glance appears rather disconnected. However, there
are frameworks that have proven useful in integrating the various approaches including
expectancy/valence theory, the unfolding model, and control theory. Although the
statements of some of the theoretical positions may appear at first glance to contradict
other theoretical positions, a closer examination reveals that are reconcilable. Rather than
testing the theoretical predictions of any one theory or pitting competing theories to
determine which one is “right,” integrative approaches view the various theories as pieces
of the motivation puzzle and recognize that all are needed to gain a complete picture.
Moreover, all of these theories are valuable in suggesting practical interventions for
creating a work situation in which employees are fully engaged in the pursuit of
organizational goals. Here are a few:

1. Carefully examine the jobs that people will perform and their motivational
characteristics, e.g., using the JDS. Select employees whose needs fit these jobs. If the
jobs are challenging and varied and the employees are allowed autonomy, a high need for
achievement person is more productive than a person lower on this need.

2. Once hired, assess the needs that are most important to employees and provide rewards
for goal accomplishment that fulfill these needs. Keep in mind that past frustration may
have led employees to deemphasize growth needs and focus more on security and more
basic needs.

3. Consider what can be done to get employees to place more importance on needs
important to performing the work. If high need for achievement is important to
successful performance, pep talks, motivational seminars, and motivational training may
make sense. Changing the needs of employees is not easy, however, and may require a
change in the culture of the organization.

4. Take action to make sure that employees have high self-efficacy expectations, i.e., the
expectancy that if they try they can achieve task success. This may require a variety of
interventions, including redesign of the work itself, instilling self-confidence and
boosting self-esteem, training to improve competence, and removing barriers to
successful performance.

5. Create a work environment in which employees see the pursuit of organizational goals
as instrumental to fulfilling valued self-conceptions and increasing their self-esteem. The
danger is that employees who see their performance of the job as crucial to their self-
esteem may engage in self-serving responses such as defensiveness and distorted
perceptions. At the same time that managers link the work environment to the self, they
should ensure that employees possess crucial knowledge, skills, and abilities, receive
constructive feedback on performance, and understand what is expected.

6. Make sure that employees perceive that successful performance is positively
instrumental to achieving valued outcomes, i.e., outcomes that fulfill their important

needs. This is accomplished through creating actual instrumentalities and then
communicating these instrumentalities. To the extent that these rewards are few and far
between, the instrumentality is low. We need to make sure that employees perceive a
high instrumentality of effort and persistence for achieving rewards, so we actually
reward those who work hard. This is in contrast to reward programs where a very few
people are selected a recognized for good performance. In addition to setting up a reward
program where there are actual contingencies, we need to communicate the
instrumentalities directly and make a show of people getting rewarded for their efforts.

7. Make sure that employees set specific, difficult goals for accomplishing the tasks they
are performing. If the tasks are complex or not well learned, the specificity and difficulty
of the goals need to start at a lower level and then escalate as the tasks are mastered. To
the extent that the task is novel and is likely to remain complex and full of surprises, try
your best goals are often superior to specific goals. It is possible in some cases to break a
complex task into simpler subtasks. For each subtask we can then set difficult, specific
goals.

8. If specific, difficult goals are assigned, we need to explain and justify why we made
these assignments.

9. At the same time that goals are set that are specific and challenging, provide feedback
on how well the employees are doing in working to achieve the goals. Specific, frequent
feedback is recommended, but also consider the nature of the task. On complex,
ambiguous tasks and when employees are still learning, more general and less frequent
feedback may be more beneficial until the employees achieve some level of mastery.

10. In the administration of rewards, assigning of tasks and goals, and the general
treatment of employees, employers should provide procedural fairness by making
decisions on these matters consistent and free from bias, allowing employees to
participate in the decisions or at least have voice, and providing a means of appealing
decisions.

11. We should give careful attention to how employees are assessing what they are
receiving in the way of benefits for their investments in their work relative to the benefits
and investments received by others. One way is through maintaining internal equity in
compensation through job analysis and evaluation (see the module on job
analysis). Another approach is through making transparent how people are being
rewarded. Transparency requires, however, that employers put in place a reward system
that is actually fair.

12. We should make sure that people feel a sense of self-competence and self-
determination and that extrinsic rewards such as money are not overemphasized to the
neglect of other, more internal rewards, such as pride in a job well done or the joy of
helping others. This is accomplished through a variety of interventions, including job
enrichment, human relations oriented supervision, reward systems, performance appraisal
and feedback, the training and development of employees, and employee participation.

CHAPTER 5: ATTITUDES TOWARD WORK AND
ORGANIZATIONS

Introduction

Think about a current job or a job held in the past. Now describe the pros and cons of this
job. In response to this request, most people can easily generate the things they like or
dislike about their jobs. Consider these comments of employees of organizations chosen
as the best places to work in 2016 by Glassdoor (https://www.glassdoor.com/Best-Places-
to-Work-LST_KQ0,19.htm):

“The people that work here are some of the best I’ve met in my professional career. The
culture is good. The founders are great people and I believe they have the best intentions
for the company, the employees, and our community. There is a lot of opportunity to
learn from different teams and possibly switch roles as departments grow and new team
form.”
“Working at … has been a great experience so far. When they say they take an active
interest in you and what you want to do with your career, they mean it. From ongoing
meetings with my manager to development opportunities offered free of charge, I feel
like I have a lot at my disposal to be successful here. It’s also a really fun place to work–
and not just because there are dogs around you. Everyone you pass in the hallways seems
to be smiling and enjoying their jobs, which makes me think I could enjoy it here for
years to come.”
“Leadership places a heavy emphasis on employee growth across all divisions, from
tuition reimbursement to offering opportunities to take on challenges outside your core
responsibilities. -Transparent culture from top to bottom. -Management actually listens to
employees, and makes quick changes to structure if/when needed. -Flexible working
hours. -People here are incredibly smart and passionate, genuinely fun to work..”
On the negative side consider some of the reviews posted on Glassdoor
(https://www.glassdoor.com/index.htm)
“Overwhelming, too many hours, no days off. It’s like an insane asylum or the twilight
zone. So many old career carriers never wanting to retire and looking down on the
CCA’s. They don’t care about children or families at all. I’m ready to quit, it isn’t worth it!
Life is too short to miserable and exhausted.”
“You cannot live on this job alone. The work is physical and demanding and can do
some harm to your body. The manager plays favorites and some people get away with
doing as little as possible. It is dirty and you are exposed to the elements in the winter.
Very little chance for advancement and not enough hours to support existence. Peak
Season can easily burn you out.”

“Most of the employees who are in a supervisory or managerial role are family and/or
friends of other managers. There is a serious nepotism problem. The list goes on. I need
to stop here. My best advice is DO NOT work here unless you want your work-life to be
a living hell.”

https://www.glassdoor.com/Best-Places-to-Work-LST_KQ0,19.htm

https://www.glassdoor.com/index.htm

These quotes reveal the attitudes of the employees toward their jobs and the organizations
for whom they work. Psychologists define an attitude as a state of readiness to respond in
a particular way in one’s thoughts, feelings, and actions to an object in the environment.
It is a predisposition or leaning formed on the basis of the information that one acquires
through past experience with the object that is the target of the attitude. The object that is
the focus of the discussion in this chapter is the work situation of employees, including
the jobs they hold and the organization that employs them. The three components of a
work-related attitude are illustrated in figure 5.1. A work-related attitude is reflected most
obviously in the feelings expressed about the work (e.g., “…a living hell”). They are also
expressed in the beliefs about the job (e.g., “The work is physical and demanding”) and
the behavioral inclinations of the respondent (e.g., “I’m ready to quit”). An employee
holding a positive attitude is more likely to have positive feelings, beliefs, and behavioral
inclinations than an employee with a negative attitude.

Figure 5.1: The Evaluative, Cognitive, and
Behavioral Components of Job Satisfaction

Employees form work-related attitudes that are relatively stable and have important
consequences. They influence not only the effectiveness with which employees perform
their jobs, but also their health and well-being. This chapter reviews the research and
theory for three work-related attitudes: job satisfaction, job involvement, and
organizational commitment. In the discussion that follows, the review of the research is
organized around two primary questions. How is the attitude measured, and what are the
facets or components of the attitude? What are the correlates? What is the process by
which the attitude forms? After considering whether satisfaction, involvement, and

commitment reflect the same of different constructs, the chapter concludes with a
discussion of the practical implications.

Job Satisfaction

Of the three attitudes examined in this chapter, psychologists have devoted the most
attention to job satisfaction.

How is job satisfaction measured?

Although work-related attitudes consist of cognitive, behavioral, and evaluative
components, the measurement of job satisfaction has focused more on the evaluative
component than either the cognitive or behavioral components. Psychologists often say
that whatever exists is measurable. As nebulous as attitudes may seem, they too are
measurable using quantitative scales. Investigators have used two methods of measuring
job satisfaction and other work-related attitudes, interviews and standardized self-report
questionnaires.

Interviews. One approach is to interview the person and assess their satisfaction on the
basis of their answers to questions. Take, for example, Grace Clements’ description of
her job, the making the molded inner lining of suitcases in a plant that manufactures
luggage:

“All day long is the same thing over and over. That’s about ten steps every forty seconds
about 800 times a day…. I daydream while I’m working. Your mind gets so it
automatically picks out the flaws [in the luggage]…. You get to be automatic in what
you’re doing and your mind is doing something else…. I hope I don’t work many more
years. I’m tired. I’d like to stay home and keep house.”
Contrast Grace’s description of her job to that of Kay Stepkin, the director of a small
nonprofit bakery that produces and sells bread:
“We try to have a compromise between doing things efficiently and doing things in a
human way. Our bread has to taste the same way every day, but you don’t have to be
machines. On a good day it’s beautiful to be here. We have a good time and work hard
and we’re laughing…. I think a person can work as hard as he’s capable, not only for
others but also for his own satisfaction…. I am doing exactly what I want to do…. Work is
an essential part of being alive. Your work is your identity. It tells you who you are….
There’s such a joy in doing work well.”
Grace’s feelings about her job are quite different from Kay’s: Grace views each work day
as a wearisome burden, whereas Kay anticipates each work day with eagerness and
challenge. A reasonable inference from their comments is that Kay is much more
satisfied with her job than Grace. (By the way, these quotes are taken from Working:
People talk about what they do all day and how they feel about what they do, by Studs
Terkel, 1974; check out the following link for another example Studs Terkel provides
from his interviews about how a meter reader finds job satisfaction).

Also, check out this enactment of some of the interviews from Studs Terkel’s book
Working by a theatre group.

Interviews such as those can produce a rich body of information and provide valuable
insight into job satisfaction. They suffer, however, in that they fail to produce a
standardized, objective, quantitative measure of work-related attitudes that researchers
can track and examine over time. Moreover, interviews are costly, time consuming, and
subject to numerous subjective biases. For that reason, psychologists typically have used
self-report questionnaires in which employees are asked to indicate their level of
satisfaction on numerical scales.

Single item questionnaire measures. Asking a general question about satisfaction and
having participants respond on a numerical scale is more convenient and less expensive
than interviews. It also allows the gathering of a large amount of data from large samples
of workers and provides a standardized metric. Yet, having respondents answer one
general question about satisfaction has its downsides. For one, single item measures do
not provide much information about the sources of satisfaction or dissatisfaction. Also,
responses to these measures appear biased in a positive direction.

In national polls respondents are asked simple questions such as “Are you satisfied with
your job?” Summaries of the results over time show a remarkably high percent of
respondents stating that they are satisfied with their jobs (see figure 5.2 for the results of
General Social Survey from 1972 – 2014).

Source: General Social Survey (https://gssdataexplorer.norc.org/variables/2838/vshow)
Respondents were asked “All in all, how satisfied would you say you are with your job?”
Responses alternatives were Not at all (1), Not too satisfied (2), Somewhat satisfied (3),
Very satisfied (4).

Figure 5.2: Job Satisfaction among U. S. Employees from 1972 to 2014

The findings stand in stark contrast to frequent accounts in the media that the workforce
is deeply dissatisfied and that dissatisfaction is increasing (e.g., Job Satisfaction Is at an All Time
Low. We’re Shocked. Shocked, 2010; retrieved from http://www.onlineinvestingai.com/blog/2010/01/12/job-

https://gssdataexplorer.norc.org/variables/2838/vshow

http://www.onlineinvestingai.com/blog/2010/01/12/job-satisfaction-is-at-an-all-time-low-were-shocked-shocked/

satisfaction-is-at-an-all-time-low-were-shocked-shocked/). Although there are ups and downs in
reported satisfaction, it is not uncommon to find 80 – 90% of respondents stating in
response to one-item questions that they are satisfied. In the General Social Survey,
87.9% indicated that they were very satisfied or somewhat satisfied with their jobs in
2002. In the same survey, 89.6% indicated that they were very satisfied or somewhat
satisfied in 2014. Based on analyses of several surveys, researchers of one article
concluded that mean job satisfaction scores were high and the small changes that did
occur from year-to-year were random fluctuations rather than meaningful trends
(http://www.siop.org/tip/Apr13/08_Bowling.aspx).

A similar satisfaction scale was used in surveys conducted across European countries.
The results are summarized in Figure 5.3. There is a range in the mean job satisfaction
found for countries with Denmark, Norway, Great Britain, and Switzerland at the top on
satisfaction and Turkey and East European countries at the bottom. However, the results
suggest that the majority of employees across all countries tend to report they are
satisfied with their jobs.

Further evidence of the satisfaction levels of employees comes from a survey of the G8
and BRICS nations:

http://en.rocketnews24.com/2013/06/19/japan-second-worst-in-g8-for-employee-
satisfaction/
The percentage saying that they are satisfied was 60% or higher for all of these countries
with the exception of respondents from India.
Canada: 91%
Great Britain: 91%
Germany: 86%
Italy: 86%
U. S.: 86%
Brazil: 86%
South Africa: 81%
Japan: 81%
France: 81%
China: 73%
Russia: 60%
India: 39%

An unresolved issue is whether the results of surveys using one-item questions reflect
true levels of satisfaction or whether they are the consequence of biases. People may
respond more positively to single item satisfaction measures because of social desirability
bias. In other words, they state that they are more satisfied than they really so that others
will view them as normal and happy people rather than as malcontents. Researchers can
reduce social desirability bias with surveys that are anonymous and confidential.
However, bias would still remain because of self-deception. This term refers to the fact

http://www.onlineinvestingai.com/blog/2010/01/12/job-satisfaction-is-at-an-all-time-low-were-shocked-shocked/

http://www.siop.org/tip/Apr13/08_Bowling.aspx

http://en.rocketnews24.com/2013/06/19/japan-second-worst-in-g8-for-employee-satisfaction/

that people are not always totally aware of their own level of satisfaction and do not want
to think of themselves as unhappy.

Respondents were asked “All in all, how satisfied would you say you are with your job?”.
Responses alternatives were Not at all (1), Not too satisfied (2), Somewhat satisfied (3),
Very satisfied (4).

Figure 5.3: Job Satisfaction in European Countries

Multi-facet questionnaire measures. As a consequence of both social desirability and self-
deception, responses to one-item job satisfaction questionnaires are biased in a positive
direction with the majority of people expressing positive attitudes. Although the use of
one-item attitude measures continues to dominate national surveys, psychologists have
attempted to deal with the potential problems with self-report questionnaires that measure
attitudes toward specific facets of the job. In job satisfaction questionnaires, it is common
to ask about facets such as the pay, one’s supervisor, one’s coworkers, the working
conditions, the job itself, benefits, and other characteristics of the work and the work

context. It is also common to include an overall job attitude measure that consists of the
average of responses to the ratings of the separate facets.

To illustrate, consider four multi-facet job satisfaction questionnaires: the Hoppock job
satisfaction scale, the Faces scale, the Job Description Index (JDI), and the Minnesota
Satisfaction Questionnaire (MSQ). While not exhausting the possibilities, they provide
examples of some of the alternatives to measuring job satisfaction.

1. The Hoppock scale of job satisfaction. The true beginning of scientific, quantitative
research on job satisfaction actually began back in the 1930s with Robert Hoppock who
was a graduate student interested in applying newly developed scaling (measurement)
techniques to the study of job satisfaction. He devised questions that people answered by
referring to a scale ranging from 100 (extreme dissatisfaction) to 700 (extreme
satisfaction). Although this seems like a common and simple task, collecting such data
was a novel concept in the 1930s. With the aid of his father-in-law, he collected data
from most working adults in New Hope, Pennsylvania. Hoppock discovered that
approximately 88% of those sampled were satisfied with their jobs. He also found that
the most satisfied workers were in the professional, managerial, and executive
occupations (Hoppock, 1935). It is interesting to note that the percentage of respondents
saying they are satisfied with their jobs is still high and has not changed a great amount in
the years since Hoppock’s research.

Hoppock’s research is significant for two reasons. First, he developed one of the first
contemporary job attitude surveys, and, therefore, provided a template for the data
collection method that researchers would use in subsequent job attitude research. Second,
researchers have replicated Hoppock’s results repeatedly for over 55 years. In figure 5.4,
the job satisfaction of the employees from the five categories of occupations that he
sampled is provided, together with the rank order of responses in each category and the
average or means response in each category. (Remember that the scale ranges from 100
to 700, with higher scores indicating greater satisfaction.)
Kahn measured satisfaction of workers in seven occupational categories with a job
satisfaction survey in which the scale values ranged from 1 (very dissatisfied) to 5 (very
satisfied). The pattern of results is very consistent across these two samples collected
approximately 40 years apart.

If one takes into account the scaling differences in the two job satisfaction studies (that is,
100-700 vs. 1-5), another point becomes obvious. Considerable overlap exists among
categories, with many laborers having reported as much job satisfaction as many
professionals. Also, across all occupations, the average satisfaction score is skewed or
distorted toward the high (satisfied) end of the scale. The high job satisfaction reported
by most workers across all occupations is also an enduring finding of job satisfaction
research. The existing data on relative levels of job satisfaction totally dispute the
anecdotal perception that most people are unhappy with their jobs. Of course, these
conclusions are based on mean or average responses; any one worker may report extreme
dissatisfaction with his or her job.

Figure 5.4: Differences Among
Occupations on Job Satisfaction (Hoppock, 1935)
Compare Hoppock’s results with data reported by Kahn (1981; excerpted from Quinn &
Shepard, 1974; see figure 5.5).

Figure 5.5: Differences Among Occupational
Groups on Job Satisfaction (Quinn & Shepard, 1974)

2. Faces scale. The second oldest of the scales discussed here is the Faces Scale (Kunin,
1955). As depicted in figure 5.6, this scale requires respondents to indicate their degree of
job satisfaction with a facet of the job by checking the drawing of the human face that
most closely approximates their feelings. Both a male and a female (Dunham & Herman,
1975) version of the Faces Scale exist, and either form is appropriate for both male and
female workers. Although the scale looks simple, Kunin (1955) actually used a complex

scale construction procedure to develop a set of expressions that spans extremes in
attitude and also captures equal intervals between the extremes.

The Faces Scale is very quick and easy to answer and is generally acceptable as an
approach to measuring satisfaction with the separate facets of a job such as salary,
coworkers, supervision, and work conditions. The Faces Scale is the preferred method if
workers have little time to fill out attitude surveys and or minimal reading or language
skills.

Figure 5.6: Faces Scale of Job Satisfaction

3. Minnesota Satisfaction Questionnaire (MSQ). The MSQ is the creation of Weiss,
Dawis, England, and Lofquist (1967) and consists of 100-item job satisfaction scale with
five items, each composing 20 facet scales (see table 5.1). The facets of the MSQ include
satisfaction with advancement, compensation, coworkers, responsibility, and working
conditions. The scale is scored as a global measure by summing across the 100 items as
well as a facet measure by summing across the five items in any facet scale. Because of
its length, most researchers use a 20-item short form of the MSQ. Typically, this 20-item
version is used to assess global satisfaction. Factor analyses of the short form reveal two
dimensions or facets within the 20 items, labeled intrinsic and extrinsic satisfaction. The
intrinsic items refer to aspects of the work or job itself (“The chance to do different things
from time to time”), and the extrinsic items refer to conditions or situations external to
the job or work (“The way company policies are put into practice”).

The authors of the MSQ systematically drew upon previously published scales and
organizational theory to develop their scale. Their goal was to produce a job satisfaction
instrument with superior measurement properties (i.e., good reliability and validity). As a
global measure of job satisfaction, the MSQ achieves this goal admirably. However,
researchers express some reservations about the adequacy of the intrinsic and extrinsic
dimensions as separate facets of global job satisfaction (Cook, Hepworth, Wall, & Warr,
1981). Consequently, most users of the MSQ only score for global satisfaction.

Table 5.1: Sample Items from the Minnesota Satisfaction Questionnaire

4. The Job Description Index (JDI). The JDI is unlike the previously discussed job
satisfaction measures in that it was developed specifically to measure satisfaction with
different job components or facets. Smith, Kendall, and Hulin (1969) created the JDI by
systematically developing and testing over 100 potential items. The authors created a
scale that is easy to read and answer. Respondents are required only to check off the
adjectives or phrases describing their jobs (see table 5.2). The five facet satisfaction
scales are the work itself, pay, promotion opportunities, supervision, and coworkers.

Some researchers sum across the five facet scales to obtain an acceptable measure of
global job satisfaction (Hulin, Drasgow, & Komocar, 1982). However, because the facets
are only moderately intercorrelated, the authors do not recommend summing the facets to
obtain an overall score. Instead, they advise using the job in general (JIG) scale, a
measure contained in the latest version of the JDI and measures global job satisfaction
(Ironson, Smith, Brannick, Gibson, & Paul, 1989).

The JDI is the most widely used measure of job satisfaction. Research on the
measurement properties of the JDI has shown that it provides reliable and valid job facet
satisfaction scales. One possible weakness of the JDI is that the five job facets are not
comprehensive by today’s standards. Four facets measure extrinsic satisfaction, or
satisfaction with work conditions, and only one measures intrinsic satisfaction (i.e., the
work itself). However, if interest does not extend beyond the five facet scales, the JDI is
definitely the preferred measure of job facet satisfaction.

Table 5.2: Examples of Items from the Revised Job Descriptive Index

Points to Ponder:

1. Consider your own attitudes toward a course you are taking and/or a job you have held.
Where would you place your overall attitude on the Faces Scale?
2. What separate facets can you identify in relating your attitudes toward the course or
job? How do your attitudes vary across the facets? Which are most important in
determining your overall attitude toward the course or job and why?
3. Identify cognitive, behavioral, and affective components of your attitude toward the
course or job.
4. To what extent do you believe your attitudes influence how well you perform in the
course or job? Why or why not?
5. How has job satisfaction usually been measured? Which scale is
more appropriate for workers with limited education? Which scale would tell a manager
if his or her employees are dissatisfied with their pay?

Work environment characteristics associated with job satisfaction

Using measures of satisfaction with the job in general and with the individual facets of
the work, researchers have collected a huge amount of data on how various
characteristics of the work environment are related to satisfaction. In considering these
findings, remember that these findings are based on correlational research. The
correlations provide evidence of “potential antecedents” of job satisfaction, but do not
constitute proof that the correlates actually caused satisfaction. Given that much of the
research was conducted by measuring both situational factors and job satisfaction with
self-reports, one could just as easily conclude that the satisfaction of the worker led to
perceptions of the situation as vice versa. Still, the findings suggest potential leverage
points for employers seeking to increase worker satisfaction.

Job characteristics. Based on a meta-analysis of the correlations found between the core
job characteristics and general job satisfaction, it is clear that people are more satisfied
with their jobs to the extent that their job is high on the factors measured in the Job
Characteristics Scale that was discussed in the last chapter (Fried & Ferris, 1987). In
other words, they are more satisfied the more varied the tasks they perform (r = .45), the
more identity the job possesses, i.e., the tasks have a meaningful beginning and end (r =
.26), the more significant the job’s impact in the organization and on other people (r =
.35), the more autonomy and input the employees have (r = .48), and the more feedback
they receive on their performance (r = .43).

Fairness. People are more satisfied with their jobs if they perceive that they are treated
fairly. A meta-analysis of the research on job satisfaction and justice found that workers
were more satisfied with their jobs to the extent that they perceived that there was
procedural justice (r = .62) and distributive justice (r = .56) (Colquitt, Conlon, Wesson,
Porter and Ng (2001). In other words, job satisfaction is associated with situations in
which people feel that they are allowed to have some input to decisions and the
procedures used in the decisions are consistent, objective, justified, and open to appeal.
Also, job satisfaction increases to the extent that people perceive that outcomes are
equitably allocated.

Supervisory behavior. People are more satisfied with their jobs to the extent they perceive
their supervisors as task oriented (i.e., they are organized and emphasize high
productivity) and considerate (i.e., they are concerned about subordinate well-being and
needs). Supervisors who are characterized by a laissez faire orientation (i.e., exert little
leadership and are disorganized) are likely to have dissatisfied employees.

Cohesiveness of the group. People are more satisfied to the extent they work in cohesive
work groups and like their coworkers. The authors of one meta-analysis of the
relationship of group cohesion to satisfaction outcomes concluded that group cohesion
was the strongest predictor of job/military satisfaction (Oliver, Harman, Hoover, Hayes,
& Pandhi, 1999). Similarly, the authors of a meta-analysis examining the relation of
group cohesion to satisfaction in virtual teams concluded that the strongest effect was

between group cohesion and satisfaction (weighted correlation = .57) (Lin, Standing, &
Liu, 2008).

Participation in decision-making. People are more satisfied to the extent that they
perceive that they are allowed to participate and have input into important decisions. The
correlation of participation and job satisfaction is larger in studies in which both
satisfaction and participation are measured with employee self-reports (Wagner, Leana,
Locke & Schweiger, 1997). This suggests that some method bias is at work that might
inflate the correlation. However, the findings suggest that participation is related to
satisfaction, albeit at a modest level, and that the benefits are most apparent when the
participation involves the generation, processing, and use of information not already
available to the supervisor.

Pay. It is not surprising that people are more satisfied the more they are paid (Judge,
Piccolo, Podsakoff, Shaw, and Rich, 2010). What is surprising, however, is that the
relation is not strong. The authors of one meta-analysis found a correlation corrected for
statistical artifacts of only .15, indicating a relationship that is much smaller than usually
thought to exist (Judge et al, 2010). Although the corrected correlation between the
amount paid and pay satisfaction was larger, this relationship also was small (.23).

Working conditions. People are more satisfied to the extent that the work environment is
clean, low on stress, safe, and comfortable.

Flexibility of work schedules. People are more satisfied to the extent that they have
flexible working schedules that allow them balance work and family and other non-work
concerns.

Perceived discrimination, harassment, and bullying. Employees report more negative
work-related attitudes to the extent that they perceive they are the targets of
discrimination and harassment in the workplace. Based on a meta-analysis involving 79
correlations and a total sample of 14,452 employees, Triana, Jayasinghe and Pieper
(2015) reported a mean uncorrected correlation of -.32 and a mean corrected correlation
of -.38 (corrected for sampling error and unreliability) between perceived racial
discrimination and work-related attitudes. Willness, Steel, and Lee (2007) found in a
survey of employees that reports of sexual harassment were associated with lower job
satisfaction. Nielsen and Einarsen (2012) found in their meta-analysis that employees
were less job satisfied to the extent that they reported bullying in the workplace.

Points to ponder

1. To what extend do you believe there are individual differences among employees in
how the various variables influence job satisfaction?
2. The variables listed in this section are called “potential antecedents.” For each one,
describe the dynamics by which the variable might shape and cause the level of job
satisfaction of workers? In your answer describe the potential mediators and moderators
of the relationship.

3. Think of the jobs that you have held. How did these jobs compare on the
characteristics listed above? To what extent did these factors influence your attitudes
toward the job?

Personal characteristics associated with job satisfaction

In addition to situational characteristics, it is important to consider the personal
characteristics of employees that they bring to the work situation and may influence their
job satisfaction. These individual differences include traits such as personality
characteristics and demographic characteristics such as race, age, and sex.

Demographic characteristics. Does job satisfaction vary as a function of demographic
characteristics such as age, sex, and race? For example, are women, blacks, and older
people more or less satisfied than the typical young, white, male worker? The research on
the relation between personal characteristics and job satisfaction has yielded findings
showing correlations that are quite small and often close to zero. When large differences
are found among demographic subgroups, they are often surrogates for other variables,
such as occupational level, education, and tenure. For example, women, minorities, and
younger workers are often overrepresented in lower status jobs where job satisfaction is
lower. In the research on demographic characteristics only the age of employees appears
to show a consistent relationship to job satisfaction.

a. Sex. The research on sex differences in job satisfaction reveals very small differences
for both global job satisfaction measures (Brush, Moch, & Pooyan, 1987; Mukerjee,
2014; Smith, Olsen, & Falgout, 1991) and measures of facet satisfaction (Andrisani &
Shapiro, 1978; Hodson, 1989; Hulin & Smith, 1964; Mason, 1995; Saad, 2013; Tuch &
Martin, 1991; Weaver, 1978). Also, the differences found between men and women has
varied across studies. Some research provides support for a “happy female worker”
hypothesis and shows that women express more satisfaction on global measures than men
(Clark, 1997; Mukerjee, 2014; Sousa-Poza & Sousa-Poza, 2000). Other research findings
show that women do not differ substantially on job satisfaction from men on most facets,
with the possible exception of salary where they tend to express greater dissatisfaction
(Saad, 2013). Based on an analysis of international data collected by Gallup, Tay et al
(2014) conclude that women are equal to men in their job satisfaction but that women are
more likely to report that the work situation provides an opportunity to do their best.
These mixed findings are surprising given that women tend to have lower salaries and
work in lower skill, more routine jobs than men. As stated in the title of one article, “Why
aren’t women more dissatisfied?” (Hodson, 1989). The author in this case speculated that
women use different comparison groups that lessen their dissatisfaction and are
socialized to talk less about their dissatisfaction. Obviously more research is needed to
determine the extent to which men and women differ on job satisfaction and why.

b. Ethnicity. The results are complicated with regard to ethnicity. The research typically
shows that black employees express lower satisfaction than white workers (Forgionne &
Peeters, Slocum & Strawser, 1982; Mukerjee, 2014; Tay, et al, 2014; Toch & Martin,
1991; Weaver, 1977, 1998). Based on an international sample, Tay et al (2014) report

that while black employees tend to report lower job satisfaction, Hispanics tend to report
higher satisfaction, and Asian workers report the same satisfaction as white workers.
Typically, the ethnic differences are small, but a more recent study explored 1997 data
from two national surveys conducted in the U. S. and found that black respondents
expressed substantially more dissatisfaction than white respondents (Mukerjee, 2014).

There is a need to address the cultural and structural reasons in attempting to make sense
of the differences on job satisfaction among ethnic groups. Structural reasons refer to
organizational factors that may contribute to racial differences; for example, relative to
whites, blacks may not have ready access to mentoring relationships with (white) senior
managers. The ethnic gap in satisfaction reflects in part the fact that black employees
experience lower pay and poorer work conditions (Toch & Martin, 1991). However, there
is some evidence that the racial gap remains even after salary, occupational level, and
level of benefits are controlled and that perceived discrimination is the more important
determinant of the dissatisfaction expressed by black employees (Mukerjee, 2014).
Perceived discrimination is a powerful predictor of negative work-related attitudes as
shown in a recent meta-analysis (Triana, et al, 2015). Cultural reasons refer to cultural
differences in values or expectations. One researcher concludes that workers from
different ethnic groups differ on reported job satisfaction both because of differences in
how they are treated at work and in their own personal value systems (Moch, 1980). As
in the case of gender differences on job satisfaction, a lot of unanswered questions remain
in attempting to explain the lower levels of satisfaction reported by black employees.

c. Age. The research on age differences on job satisfaction shows that job satisfaction
increases with age for all demographic groups, including women and minorities (Bourne,
1982; Hunt & Saul, 1975; King & Jex, 2004; Ng & Feldman, 2010; Rhodes, 1983; Tay,
Ng, Kuykendall, & Diener, 2014)). Satisfaction is positively associated with age, even
after controlling for the employee’s tenure (Tay et al, 2014). This trend occurs for
satisfaction with all facets except for satisfaction with promotions, where satisfaction
appears to decline with age (Ng & Feldman, 2010). The increases in satisfaction with age
appear somewhat stronger for employees who are minorities, those with longer tenure on
the job, and those without a college education (Ng & Feldman, 2010). One explanation
for these findings is that workers are initially dissatisfied with their jobs but eventually
seek work that they find more satisfying. Another explanation is that people actually
become more satisfied with (or perhaps more realistic about) their lives, including their
work lives, over time. In support of this latter possibility, older workers often perceive
they have fewer job options or alternatives and resign themselves to their job situations or
(Pond & Geyer, 1987). Another explanation is socioemotional selectivity theory (Ng &
Feldman, 2010). According to this theory, increases in job satisfaction reflect changes
that occur with age in how employees invest their time and energy. Younger workers
invest more time and effort in acquiring information that furthers their careers. Older
workers place more importance on emotion regulation goals and invest more time and
effort in social activities. Older workers seek out those who have provided emotional
support and experience more gratification from their social interactions. More research is
needed to test the alternative explanations for the positive relationship of age and
satisfaction with the work itself. So far the findings are inconclusive and the reasons for

the age differences remains to be further explored (Barnes-Farrell & Matthews, 2007;
Janson & Martin, 1982; Ng & Feldman, 2010).

d. Generational differences in job satisfaction and other work-related attitudes. Related to
age are generational differences in job satisfaction and other work-related attitudes.
Traditionalists born between 1920 and 1945 are stereotyped as conservative, disciplined,
and hard working. Baby boomers born between 1945 and 1960 are stereotyped as
materialistic, driven by the work ethic, and stressed. Generation Xers (i.e., born between
1960 and 1980) are stereotyped as skeptical and individualistic. Millennials (i.e., born
between 1980 and 1999) are stereotyped as cynical, socially conscious, self-absorbed,
and less engaged in their job, organization, and career.

Thousands of articles have been written in popular media about the implications of these
purported generational differences. But what does the research show? The research
comparing the generations on job related attitudes shows that older generations are
slightly more satisfied with their jobs than younger generations, but that there are no
meaningful or systematic differences among the generations on work-related attitudes
(Costanza, Badger, Fraser, Severt, & Gade, 2012). In short, researchers have found that
the overall levels of job satisfaction have changed very little since the first surveys in the
1930s. At any one point in time younger workers tend to report somewhat less
satisfaction with their jobs than older workers, but pronouncements about large
generational differences are unsupported in the research.

Personality. Some people tend to be satisfied or dissatisfied regardless of the situation in
which they find themselves. This view places great importance on the personality traits of
the individual and would suggest that the level of job satisfaction expressed by an
individual is stable over time and situations. Four of the five Big Five personality factors
(conscientiousness, extraversion, agreeableness, and neuroticism) are related to job
satisfaction in a manner that appears consistent with theory and commonsense. Table 5.3
summarizes the results of two meta-analyses.

People who are conscientious, agreeable, and extraverted were found to be more satisfied
with their jobs than those who were less conscientious and extraverted. Also, people who
were higher on neuroticism were less satisfied with their jobs than those low on
neuroticism. In this particular meta-analysis, four of the five personality traits were
related to job satisfaction at a conventional level of statistical significance, but the size of
the correlations were large enough to be considered substantial only for
conscientiousness and emotional stability (i.e., neuroticism). Even stronger than the
correlations found for the Big Five personality traits were the correlations between
positive and negative affectivity and job satisfaction (Levin & Stokes, 1989; Ng &
Sorenson, 2009). Positive affectivity is the disposition to experience positive states such
as happiness, joy, elation, and pride. Negative affectivity is the disposition to experience
aversive emotional states, such as distress, agitation, pessimism, and dissatisfaction. Not
surprisingly, job satisfaction is higher for those with higher positive affectivity and lower
negative affectivity.

Is job satisfaction stable across time and jobs? The common assumption is that job
satisfaction is a response to the situation, but everyday observations of people at work
suggest that there is a trait component. The reader no doubt has observed some people
who are dissatisfied with their current situation but appear to be disgruntled no matter
what their situation. Also observed are people who are satisfied regardless of the
situation. Are these people simply prone to some level of satisfaction that is rooted in
their personality?

k = number of correlations; N = number of participants; ruc = mean correlation weighted
by sample size (uncorrected); rc = corrected correlation; * = confidence interval for
correlation excluded zero.

Table 5.3: Results from Meta-analyses of the
Correlations of Big Five and Affectivity

Several researchers have presented compelling evidence of a trait component in job
satisfaction. They have shown that people who are relatively satisfied tend to remain
satisfied regardless of the job whereas those who are relatively dissatisfied remain
dissatisfied regardless of the job. Staw and Ross (1985) reported this on the basis of an
investigation of the job satisfaction of a national sample of 5,000 male workers over a
five-year interval. Regardless of whether the workers changed jobs or occupations, the
researchers found that job satisfaction reported by individuals was stable over time. A
meta-analysis of test-retest of job satisfaction over approximately 3-year intervals reveals
a mean corrected correlation of .50 (Dormann & Zapf, 2000). The authors in this case
suggest that because the correlations were not corrected for restriction of range, this is an
underestimate. The findings of subsequent research demonstrate that work-related

attitudes reflect not only personal disposition to be satisfied or dissatisfied but also the
characteristics of the jobs that workers perform (Steel & Rentsch, 1997). Nevertheless,
there is a remarkable stability in job satisfaction and personal dispositions appear to play
an important role.

a. Adaptation level and opponent process. Both of these factors suggest that each
individual has a point of equilibrium in their job satisfaction and that stabilities in job
satisfaction over time reflect reactions to job events that maintain or restore this
equilibrium (Bowling, Beehr, Wagner & Libkuman, 2005). Adaptation level theory
proposes that past exposure to stimuli establishes a frame of reference that going forward
is used in judging and reacting to the same stimuli. A person who has experienced only
boring and menial jobs in his career evaluates his current boring job positively against the
adaptation level formed from past experiences. He might even consider the current job as
satisfying relative to this adaptation level. If a job is encountered that is much greater
than the adaptation level, the person might respond quite positively due to the contrast
effect but over time with continued exposure to interesting, challenging work, a new
adaptation level emerges that is used to evaluate the work. According to this explanation,
employees are always adapting to changes in work situations and as a consequence their
job satisfaction remains stable. This is somewhat similar to Herzberg’s theory who stated
that the positive boost coming from hygiene factors in the job are always temporary.
Eventually the worker asks “what have you done for me lately?” Adaptation level theory
would propose that this can occur for all features of the work, not just hygiene factors
such as compensation. The point of equilibrium (the set point) is often considered a
personal disposition reflecting the personality and biology of the individual.

Opponent process theory proposed that job satisfaction emanates from a person’s
physiological states (Landy, 1978). This theory assumes that when one experiences an
extreme emotional state, central nervous system mechanisms attempt to bring one back to
a state of emotional equilibrium or neutrality. Unlike adaptation level theory, the
experience of an emotion that creates disequilibrium does not lead to simple return to
neutrality but instead activates an oppositional process. This oppositional process
involves an emotional state that is opposite to the emotion that created the disequilibrium.
An example might clarify opponent process theory. Consider an employee who is
typically unhappy about his job and the small salary increases. One day, to his surprise,
he receives a large salary increase. After a moment of elation, he becomes sad and
depressed. According to the theory, the magnitude of the opponent process changes over
time, increasing each time it is activated. Consequently, upon receiving future salary
increases, the employee’s eventual opponent process reaction is considerably more
negative than prior reactions. The effect is to keep the employee at about the same level
of job satisfaction over time.

b. Genetics as a factor in the stability of job satisfaction. Based on the findings of stability
of job satisfaction over time some researchers have gone so far as to suggest that work-
related attitudes are genetically based and inherited. Arvey and his colleagues (Arvey,
Bouchard, Segal, & Abraham, 1989) were able to cast some light on this proposition in
an unusual study with 34 pairs of identical twins who were reared apart. To the extent

that no systematic differences in the work environments exist between each sibling pair,
differences between an individual and his or her twin would seem to reflect a genetic
influence on satisfaction. The researchers discovered that approximately 30% of the
variance in job satisfaction in the sample of identical twins was due to genetic factors.
Many of the twin pairs also had similar lifestyles and jobs. From these data, the
researchers speculated that people are genetically programmed to seek out certain
occupations and to respond to them in particular ways. More recent research assessing the
proportion of variance in job satisfaction attributable to heritability reveals that all the
genetic influence on job satisfaction is explained by personality (Hahn, Gottschling,
KÖnig, & Spinath, 2015). Identical twins tend to be similar on extraversion, emotional
stability, and conscientiousness, and it is these personality traits that account for that
proportion of variance in job satisfaction is the consequence of genetics. These more
recent findings confirm that about 30% of the variance is heritable but that both genetics
and the environment of the worker play important roles.

The implication is that people are predisposed to satisfaction or dissatisfaction and
organizations have less influence over employee attitudes than commonly believed. The
research findings for job satisfaction as a disposition are both provocative and disturbing.
If job enrichment and other changes in the work environment are not especially
worthwhile, then perhaps employers should focus their efforts on selecting employees
who are prone to satisfaction and rejecting those who are prone to dissatisfaction rather
than attempting to make work environments more satisfying. The ethical dilemmas
posed by these possibilities are obvious. On the brighter side, the findings show that
environmental factors are still important with genetic factors accounting for only 30 –
70% of the total variance. In addition, it is important to keep in mind that correlations
reflect the rank order of individuals on a measure and that changes can still occur on the
variable in question even as the rank order remains the same. As noted by Arvey et al.
(1989), “Job enrichment efforts may, however, have the intended effect of raising mean
levels of job satisfaction for the individuals involved, even though rank ordering of
individuals is preserved” (p. 191). In other words, changing the job to enhance intrinsic
satisfaction will make everyone more satisfied at the same time that some workers remain
relatively more satisfied (or dissatisfied) than others.

Person-environment fit as an antecedent to job satisfaction

So far the review has focused on the separate influence of characteristics of the work
environment and personal characteristics of the employee. In addition to a direct relation
connection of job satisfaction to the work situation and personal characteristics, these two
antecedents also appear to interact in determining job satisfaction. More specifically,
employees whose personal characteristics provide a good fit to the environment report
higher job satisfaction. The research has shown that people are more satisfied with their
jobs to the extent that they have attitudes, values, personality traits, and abilities that
provide a good fit to the various characteristics of the work environment. The correlation
between fit and satisfaction are substantial with one meta-analysis reporting that job
satisfaction increased with increased fit to the job (r = .56), supervisor (r = .44), group (r
= .31), and the organization (r = .44) (Kristof-Brown, Zimmerman & Johnson, 2005). The

authors of this meta-analysis also found that good fit to the work environment was
positively correlated with job performance. A good fit possibly is the result of the traits
and other personal characteristics that employees bring to their work situations. This
suggests that hiring people to provide a good fit to the work environment as a strategy for
enhancing employee job satisfaction and job performance. Given that these are
correlational findings, there are alternative explanations. For instance, it is also possible
that the more satisfied and higher performing employees are able to modify their work to
fit their personal characteristics or modify their personal characteristics to fit their work.

Points to ponder

1. Pose a hypothesis for why older employees might express higher levels of satisfaction
than younger workers? How would you conduct research to test this hypothesis?
2. Do you believe it is ethical to hire employees based on their likely satisfaction with the
job and the organization?
3. Describe the personality traits of the satisfied employee?
4. If the level of satisfaction of an employee tends to remain stable over time does this
mean that organizations do not need to do things to increase satisfaction? Why or why
not?

How does job satisfaction form?

Several potential factors associated with the work situation and the personal
characteristics of employees are related to job satisfaction. The next question concerns
the processes that mediate these antecedents and the level of employee satisfaction. For
instance, why does an employee with a low degree of task variety, autonomy,
significance, feedback, and identity form a relatively negative attitude toward the job
while an employee with high degree of these characteristics form a more positive
attitude? Three general processes appear to determine the formation of attitudes and serve
as mediators of the relation between the antecedents and employee job satisfaction.

1. Comparisons against personal standards: What do I have vs. what do I need, value, or
consider important?
2. Social comparison and social information: What do others have compared to what I
have and what are the attitudes of others?
3. Emotional and physiological states: What are my general mood and emotion and my
physiological state?

Comparison against internal standards. Comparison theories of satisfaction are a class of
cognitive, process-oriented theories of job satisfaction (see figure 5.7). Comparison
theories describe workers as going through a process in which they consider how much of
some characteristic they actually have in their present jobs and how much of this
characteristic they would like to have in their present jobs. The perceived discrepancy
between the two (actually have and would like to have) determines the level of
satisfaction that employees express with the job and the various job facets. The
characteristic in question is typically framed in terms of different types of comparisons,

such as needs, values, desires, expectations, or aspirations. For example, the
questionnaire might ask how much opportunity for social interaction there is in the
present job and how much social interaction is preferred. The larger the discrepancy
between the amount present and the amount desired, the lower the reported job
satisfaction (Porter, 1962). The smaller the discrepancy, the higher the reported job
satisfaction.

One type of “standard” is a human need. Among the needs discussed in the chapter on
motivation were needs for achievement, affiliation, esteem, power, and the like. Using
the comparison against internal standards approach, the need is an ideal against which
workers compare their current state or situation. Porter (1962) incorporated the concept
of attained versus desired needs in his Need Satisfaction Questionnaire (see table 5.4).
Porter assumes that in forming attitudes toward their jobs, employees mentally compute
the discrepancy between how much they are receiving in their jobs across the specific
facets of the job and how much there should be of each of these facets. If one takes the
sum of all the “should” ratings and subtracted from this total the “how much” ratings,
higher need satisfaction in the job is reflected in a larger, more positive total. Some
versions of the scale also have respondents rate the importance each need.

Figure 5.7: Cognitive Judgment Approach to Work-Related Attitudes

The concept of values is potentially useful in understanding worker attitudes. Although
workers may need a certain amount of autonomy or pay, Locke suggests that their
attitudes will depend on how much they value these factors. Because people work for
valued outcomes, values should predict job satisfaction beyond the influence of needs.
Regardless of whether values or needs constitute the basis of work-related attitudes, both
Porter and Locke propose in their theories that workers go through a cognitive process in
which they focus on their private standards in forming their work-related attitudes.

Table 5.4: The Need Satisfaction Questionnaire

Critics of the need comparison theories have identified both conceptual and statistical
problems with the idea of comparisons against internal standards. Need comparison
theory assumes that people make logical and rational comparisons between what they
have and what they need. Contrary to the notion that people make calculated judgments,
psychologists have accumulated evidence showing that people typically are not especially
rational decision makers and rarely engage in methodical comparison processes (Slovic,
Fischoff, & Lichtenstein, 1977). For example, they often do not go through each of their
needs determining how much they want and then check off how they have in forming an
opinion. Instead, form an attitude quickly based on very little information. The
calculation of job satisfaction by subtracting what the subject has (or what relevant others
have) from what he or she desires or needs has also been criticized. A problem with this
use of difference scores is that what participants believe they currently have frequently
does as well in predicting outcomes as the difference between what they have and what
they need (Wall & Payne, 1973). This is most likely to occur to the extent that
respondents are similar in their self-reported needs.

Locke (1976) proposed a somewhat different type of comparison theory. The value
comparison theory of work-related attitudes is based upon the premise that people have
positive attitudes toward their jobs if their jobs provide them with what they desire or
value. He defines needs as imperative for survival whereas he defines values as highly
subjective desires that determine satisfaction but are not necessary for survival.
According to Locke, values affect the range of responses more than the actual level of
satisfaction. For example, if employees value autonomy in their jobs, any increase or
decrease in the level of autonomy will cause wide variations in their job satisfaction.
However, if the employees do not value autonomy, then they are indifferent to changes in
the level of job autonomy.

Social comparisons. When satisfaction results from comparisons against personal
standards, workers focus on their own needs, aspirations, expectations, or values to assess
their work-related attitudes. Another source of attitudes is the information obtained from
our social environment (Goodman, 1974). In the social information processing approach,
work-related attitudes originate from looking outside of one’ self and focusing instead on
the various social cues in the environment (see figure 5.8). These social cues suggest how
one should feel and think about facets of the job.

Social information (which is in the upper left hand box) can come in different forms. One
version is in the form of direct comparisons that workers make with other people (i.e.,
comparison others). As discussed in the last chapter, Adam’s equity theory of motivation
proposes that workers ultimately determine their job satisfaction by comparing their
relevant job inputs and outputs to referent (comparison) others. In determining her level
of job satisfaction, a worker might consider that she brings an MBA and ten years of
business experience to a responsible managerial job. A coworker, her referent other, has
only a bachelor’s degree and eight years of experience. The coworker’s job is also a
managerial position, but with less responsibility and a slightly higher salary.
Consequently, the worker, after weighing the various inputs, outcomes, and amounts
received, feels dissatisfied with her pay.

Salancik and Pfeffer (1978) proposed another source of work-related attitudes in their
social information processing theory. They proposed that people form job attitudes by
observing the attitudes of relevant others or by attending to available contextual
information. If a worker takes a new job and everyone on the job seems unhappy and
depressed, this serves as a cue that the job is a dissatisfying one and the worker might just
form an attitude consistent with what others are experiencing. In a laboratory experiment
testing this hypothesis, participants were given either an interesting or a dull task to
complete and then heard coworkers who were accomplices of the researcher state that the
task was either interesting or dull (White & Mitchell, 1979). Consistent with Salancik and
Pfeffer’s predictions, participants who heard coworkers describe the task as interesting
were more satisfied with the task than research participants who heard coworkers
describe the task as dull.

Figure 5.8: Social Information Approach to Work-Related Attitudes

Although a few field tests have supported social information processing theory (e.g.,
Griffin, 1983), most of the tests of the theory have used college students who performed
artificial tasks in the laboratory for short periods of time. The laboratory research
demonstrates that expressions of job attitudes by others “can” influence workers’
perceptions of the objective characteristics of their jobs. The extent to which these effects
occur in actual organizational settings is open to debate. In organizational settings, so
many other factors are present that the simple comments of fellow workers probably
affect work-related attitudes less than the laboratory results suggest. The experiments
conducted so far have provided only weak support for the hypothesis that social cues
influence perceptions of the task and job satisfaction (Zalesny & Ford, 1990).
Nevertheless, social information deserves our attention as one of several potential sources
of work-related attitudes.

Affective and physiological states. The affective and physiological reactions of the
employee are particularly important mediators of the effects of the workplace events and
other antecedent variables on job satisfaction. A common visualization of affect is in
terms of the circumplex in figure 5.9. One can distinguish among affective responses on
two dimensions. First, there is the extent to which the affect is positive or negative.
Second, there is the intensity of the affect and the extent to which the affect reflects a
state of arousal and activation or reflects a relative quiet state.

Employees experience on a daily basis both events that evoke positive affect (e.g.,
feelings of pride from being recognized by one’s boss for a job well done, excitement
over helping a client) and negative affect (e.g., feelings of rejection resulting from
criticism from peers for not going along with the group, boredom associated with tedious,
routine tasks). As depicted in figure 5.10, one could conceive of job satisfaction as
shaped by the events in the workplace that evoke varying levels of positive and negative
affect. Job satisfaction is viewed by some theorists as an equilibrium point that emerges
over time from repeated experiences of positive and negative affect. If the preponderance
of one’s affective experiences are negative, then one could expect a low level of job
satisfaction to emerge as the equilibrium. If the preponderance of one’s affective
experiences are positive, then one could expect a high level of job satisfaction to emerge
as the equilibrium. Once formed, the average job satisfaction that emerges over time is an
equilibrium that serves as a “set point” influencing how the employee reacts affectively
to future work events. The theory borrows from evidence that the body is self-regulating
and will respond to conform to genetically determined set points. For instance, some
research suggests that people are born with a genetically determined set point in the form
of the body weight to fat ratio. When this ratio crosses a critical threshold and is either
higher or lower than the set point, the person’s physiology kicks in to bring the ratio back
in line with the set point. Similarly, affect event theory proposes that people have affect
set points and that when satisfaction, mood, happiness, and other affective responses rise
above or fall below critical thresholds, individuals respond in ways that reestablishes
equilibrium. This theory of job satisfaction also is relevant to control theory, a meta-
theory that was discussed in the last chapter as a potential framework for integrating the
various theories of work motivation.

Figure 5.9: The Affect Circumplex
Two types of affective experiences occur in both work and non-work settings. Mood
states represent general negative or positive affective states that are not related to any

specific event. Emotion represents a more specific affective reaction to specific events.
Moods are also generally thought of as having minimum impact on thought processes,
whereas emotions, especially when intense, disrupt thinking.

In one study examining the relation of mood and job satisfaction, university employees
reported their mood and job satisfaction three times a day, at 9 am, 12 pm and 3 pm, for 2
weeks (Judge & Ilies, 2004). Mood was measured by having the employees rate their
mood at each of these times on a series of adjectives. Job satisfaction was measured by
having the employees rate their satisfaction with the job at that moment. For instance, one
question asked “Right now each minute of work seems like it will never end” and another
asked “Right now, I find real enjoyment in my work.” As one might expect, ratings of
more positive affect were associated with ratings of higher job satisfaction and ratings of
more negative affect were associated with ratings of low job satisfaction. However, while
current mood measured on one day was related to job satisfaction on the same day mood
was not strongly related to job satisfaction measured on subsequent days. The effects of
current mood on job satisfaction appeared relatively short-lived.

Figure 5.10: Daily Ups and Downs in Job
Satisfaction as a Result of Affective Events on the Job.

Over time one could expect recurring moods and emotions at work would lead to more
permanent work-related attitudes. This is exactly what Weiss and Cropanzano (1996)
state in their Affective Event Theory (see figure 5.11). They propose that in addition to
comparisons against personal standards, social information, and personal dispositions,
that emotions and moods have a direct effect on work-related attitudes. Positive and
negative events at work affect affective reactions. The attitudes that employees hold

toward the facets of their work (e.g., coworkers, supervisor, work itself, etc.) form around
the strong and recurring affective reactions to the work events that a person experiences.
Attitudes are also directly influenced by the objective and perceived features of the work
(e.g., how much pay is received). The primary message of Affective Events theory is that
experiences at work evoke strong feelings that become associated with the job and play a
major role in shaping work-related attitudes. The influence of these affective reactions is
independent of rational comparisons of work facets against internal or external standards.

Figure 5.11: Affective Events Theory

Points to ponder

1. Use the affective event theory of job satisfaction to describe how you formed attitudes
toward a job that you have held.
2. In forming an attitude toward a job, describe how comparisons against personal
standards conflict with social comparisons.

3. Imagine that you are a supervisor of work group. How would you use what you know
about the processes by which attitudes form to make sure that the employees in your
work group are satisfied?

Outcomes of job satisfaction/dissatisfaction

Why should researchers in I/O spend so much time and effort on the topic of job
satisfaction? Possibly as a consequence of the human relations movement and early
research such as the Hawthorne studies, a lot of social scientists and practicing managers
came to believe that a satisfied worker is a more effective worker. The research over the
last 100 years or so generally supports this hypothesis to some extent. Although the
strength of the relationship is far less than the strongest human relations advocates would
propose and there are variables that appear to moderate the relationship, higher job
satisfaction is associated with a variety of positive outcomes. Table 5.5 summarizes the
correlations found between job satisfaction and outcomes in meta-analyses with
thousands of employees. As seen here, a satisfied employee

***is less absent from work,
***is less likely to leave the job (turnover),
***is less frequently late to work,
***tends to show more organizational citizenship (behaviors that help out coworkers and
others but are not required as part of the job),
***is less likely to show counterproductive behavior such as fighting, theft, and the like,
***receives more positive supervisory performance ratings
***performs more effectively on objective measures
***rates his or her own performance more positively
***is more committed to the organization.
***evaluates his or her life satisfaction and happiness more positively
***reports less role stress (conflict and ambiguity)
***is mentally healthier
***is physically healthier

This is an impressive list of potential benefits of higher job satisfaction, but the link
between job satisfaction and these other variables is not a simple one. To repeat an earlier
assertion, these are the results of correlational research and one cannot conclude on the
basis of the research that job satisfaction causes changes in these variables and these are
outcomes of job satisfaction. One can only say that they are potential outcomes and are
potentially caused by job satisfaction. Also, it is important to note that the correlations in
table 5.5 are statistically corrected and are higher in most cases than what would typically
find in an individual study. Also, as already mentioned, a variety of variables appear to
moderate many of these relationships. Now let us turn to some of the potential
moderators of the relationship of job satisfaction to performance, attendance, and life
satisfaction.

Are satisfied workers high performing workers? There is a small positive correlation
between job satisfaction and employee performance. One meta-analysis of research

reports an average uncorrected correlation of only .14 (Iaffaldano & Muchinsky,
1985). Note, that a higher correlation was reported on the previous page (a correlation of
.30), but the correlation reported in table 5.5 is corrected for statistical artifacts. When
uncorrected correlations are reported, it is common to find a correlation between
satisfaction and performance in the low teens. There is a small, positive correlation
between job satisfaction and job performance, but it is a SMALL relationship.

K = number of correlations; N = number of employees; ruc = mean correlation between
job satisfaction and variable, weighted by sample size; rc = mean corrected correlation
between job satisfaction and variable; * = confidence interval for the correlation
excludes zero.

Table 5.5: Correlations of Job Satisfaction with other Job-Related Variables.

Moderators of the relation between satisfaction and performance. Contemporary
researchers hypothesize that the correlation between satisfaction and performance is
moderated by other variables. Two of the moderator variables explored in the previous

research are the extent to which rewards are provided contingent on performance and the
specificity with which attitudes and performance related behavior are measured.

1. Contingency of rewards on performance. One variable that has shown promise as a
moderator of the relationship between satisfaction and performance is the contingency
between performance and rewards. The hypothesis is that among workers who receive
higher rewards, such as praise, promotions, or money, for good performance and lower
rewards for poorer performance, show a positive correlation between satisfaction and
performance. By contrast, among workers whose rewards are unrelated to performance,
the correlation between satisfaction and performance is negative.

Cherrington, Reitz, and Scott (1971) demonstrated the influence of rewards on the
satisfaction-performance link in a laboratory experiment. Two groups of research
participants were given a task to perform; one group received rewards based on
performance and the other group did not (see table 5.6). For the group that was
appropriately rewarded, the correlation between satisfaction and performance was
positive, for the group that was not appropriately rewarded, the correlation between
satisfaction and performance was negative. However, when the satisfaction-performance
correlation was computed for both groups combined, the correlation was near zero.

2. Specificity and correspondence. Fisher (1980) voiced an interesting perspective on this
issue. She maintained that the low relationship between satisfaction and performance is
partially a function of how researchers typically measure satisfaction and performance.
Two aspects of the measure that appear important are the level of specificity and the
correspondence between the attitude and the behavior.

Some years ago, Fishbein and Ajzen (1975) proposed that general attitudes are better
predictors of general behaviors and specific attitudes are better predictors of specific
behaviors. Following this reasoning, one could expect a relatively low correlation if one
variable is measured with global, nonspecific measures and the other variable is
measured at a more specific level. For example, a low correlation is likely if job
satisfaction is measured with a nonspecific measure (e.g., in general, how satisfied are
you with your job) and performance is measured on a specific facet of performance (e.g.,
how good is the performance in dealing with customers). One could also expect a low
correlation if job satisfaction is measured with a specific measure (e.g., how satisfied are
you with that part of the job involving customer relations) and performance is measured
with a nonspecific measure (e.g., in general how well does the employee perform). The
satisfaction and performance measures also should reflect logically related facets, i.e.,
that are high on correspondence. For example, satisfaction with coworkers is more likely
to show a strong relationship to performance in relating to coworkers, the facet of
performance that corresponds to the satisfaction measure, than to some other facet of
performance less correspondent with the satisfaction measure (e.g., performance in
dealing with customers).

Table 5.6: Correlations between Responses to Satisfaction
Measures and Performance with and without Performance Contingent Rewards

Is there a causal relation between job satisfaction and job performance? The small,
positive correlation between job satisfaction and job performance could reflect a relation
in which satisfaction causes performance. Also possible is a reversed causality (i.e.,
performance causes satisfaction) or a reciprocal relation (i.e., satisfaction causes
performance AND performance causes satisfaction). There are also situations in which
the correlation between satisfaction and performance are the result of a third variable
rather than a direct causal connection between satisfaction and performance. For instance,
ability could constitute a third variable that causes both satisfaction and performance and
accounts for the correlation between them. Another potential third variable is employee
tenure (i.e., number of years in the job). Longer tenure employees might cause higher
performance as the result of tenure leading to superior training and experience. Longer
tenure employees might also cause higher satisfaction as employees adjust to the job,
their coworkers, and supervisors. A positive correlation between performance and
satisfaction in this case might reflect these two causal relationships rather than a causal
relationship between satisfaction and performance.

Locke and his colleagues (Locke, 1970; Locke & Latham, 1990) proposed the High
Performance Cycle model to describe the possible linkages between satisfaction and
performance (see figure 5.12). This model uses the motivational framework of goal-
setting theory and predicts that high goals and high success expectations lead to high
performance. High performance, in turn, produces rewards, satisfaction, and commitment
to future goals. If the rewards for performance are appropriate and contingent on high
performance, it follows that employees are satisfied with their work and anticipate future
satisfaction. The most direct and immediate effect of job satisfaction is to increase
commitment to the organization and its goals. Organizational and goal commitment then
lead to the setting of specific, difficult personal goals, which for all the reasons discussed
in the motivation chapter, lead to high performance. As discussed in the previous chapter,

there are factors that moderate the effect of specific, difficult goals including feedback
ability, task complexity, situational constraints, and the degree of commitment of the
person to personal goals.

Figure 5.12: Locke and Latham’s High Performance Cycle Model
Relation of job satisfaction with life satisfaction and well-being. Since the early days of
research on job satisfaction, behavioral scientists have been fascinated with the relation
between job and life (nonwork) satisfaction (Weitz, 1952; Dubin, 1956). There are three
potential models of job and life satisfaction: a spillover model, in which satisfaction in
one area influences the other; a compensatory model, in which a lack of satisfaction in
one area is compensated by the other; and a segmentation model, in which satisfaction in
one area is independent of the other (Kabanoff, 1980). For example, if a person is
dissatisfied with her job, she may also express dissatisfaction with other facets of her life
(e.g., family, friends) (spillover model); she may turn to other areas of life for fulfillment
(compensatory model); or she may or may not express dissatisfaction with other facets of
her life (segmentation model). In a meta-analysis of the research on the relation between
job and life satisfaction and general happiness, the investigators reported on the basis of
53 studies and over 29,000 respondents that job satisfaction was positively related to life
satisfaction (r = .40) and happiness (r = .36) (Bowling, Eschlman & Wang, 2010). The
correlation between career satisfaction and life satisfaction is also positive and high, with
a meta-analysis of nine studies showing an r = .43 (Erdogan, Bauer, Truxillo &
Mansfield, 2012). Bowling, Eschlman and Wang (2010) concluded from their analyses
that there is more support for a spillover model in which satisfaction in one domain

affects satisfaction in the others than for either the compensatory or segmentation models.
Higher satisfaction with work spills over into one’s life outside of work and leads to
higher satisfaction with these other domains. Likewise, higher satisfaction with life
outside work spills over into higher satisfaction with work.

Sex of the employee appears to moderate the correlation between job satisfaction and life
satisfaction (Tait, Padgett & Baldwin, 1989). Interestingly, the relation between life and
work satisfaction is stronger for men than for women before 1974. After 1974, the job-
life correlation increases sharply for women, so that few male-female differences
currently exist. One possible explanation for why the differences in job satisfaction
between the sexes have diminished is that both job and life opportunities (and attitudes)
for women are now more similar to those of men. Another moderator is the status of the
job. There is evidence that in lower level, nonprofessional jobs, the correlation between
job satisfaction and life satisfaction is lower (r = .19; Susskind, Borchgrevinck, Kacmar
& Brymer, 2000) than in higher level, professional jobs such as college professors,
physicians, and self-employed. Indeed, the correlations are often in excess of .60 in
higher level jobs (Cunningham & De LaRosa, 2008; Near & Sorcinelli, 1986; Sorcinelli
& Near, 1989; Judge, et al, 1998; Thompson, Kopelman & Shriesheim, 1998). The higher
correlations between job and life satisfaction probably reflect the higher levels of
involvement of professionals and the centrality of work to self-esteem (Steiner &
Truxillo, 1989).

As the career opportunities for women have increased, one might ask how this has
affected the well-being of their spouses. The research so far is with heterosexual couples
and the findings show that wives who are employed show better mental health than wives
who are not employed. However, at least one study also found a negative association of
wives’ employment and the mental health of their husbands in which the working wife’s
mental health benefitted but the mental health of the husband suffered as a consequence
of having a working wife (Staines, Pottick & Fudge, 1986). The authors of this study
suggest that husbands of working wives perceive themselves as less adequate providers
than did husbands of wives who did not have outside employment, thus accounting for
the husbands’ lower levels of both job and life satisfaction and presumably their lower
mental health. These findings are troublesome but may be outdated. Having a working
wife today is the norm rather than the novelty that it was fifty years ago. Also, a question
that has not received attention in the research is how the work and life satisfaction of each
partner in a marriage affects the work and life satisfaction of the other. Perhaps it is the
quality of an individual’s work life rather than whether the person works or not that
affects the well-being of the spouse or partner. These and other questions deserve much
more attention than they have received so far in the research.

The relation of satisfaction to withdrawal behaviors. Withdrawal behavior is defined as
any type of behavior that removes the worker from the work setting regardless of the
circumstance provoking the behavior. Employees can withdraw by actually leaving their
jobs for another (i.e., they turnover). Short of actually quitting, they may instead be
absent or late for work.

1. Absenteeism and attendance. Absenteeism is costly both because of the direct cost of
having to pay the absent employee his or her wage but also because of indirect costs
associated with the absence of paying replacement workers, lost productivity, stress
among workers picking up the slack, reduced morale, and lowered quality. A study
conducted by Mercer Consulting estimated that unplanned absences, the type of absences
that employers seek to reduce, constitute a cost of about 8.7 percent of the payroll
(Kronos, How employee absenteeism hurts your bottom line,
http://www.kronos.com/ads/absence/38/Kronos_absence_out_sick ).

In a hypothetical company of 1000 employees paid an average annual wage of $43,000
with a total payroll of $4,300,000, the cost of unplanned absences is $2,494,000. A
published peer review study estimated the cost of employee absenteeism at 15% of
payroll (Navarro & Bass, 2006). Multiplied across the economy, the cost of absenteeism
easily runs into the many billions of dollars annually.

One potential source of absenteeism is employee dissatisfaction. The hypothesis that
dissatisfaction leads to absenteeism is intuitively obvious, but empirical research has not
supported a strong link between the two variables (Breaugh, 1981; Hackett, 1989;
Nicholson, Brown, & Chadwick-Jones, 1976). Hackett (1989) found in his meta-analysis
that the corrected correlation between overall job satisfaction and frequency of
absenteeism was a mere -.09. The findings of some research suggest that nonattitudinal
variables, such as a worker’s previous history of absenteeism are better predictors of
absenteeism than satisfaction (Ivancevich, 1985).

One problem in the research on satisfaction as a predictor of attendance is the inability to
distinguish between voluntary and involuntary absenteeism. One would expect that job
satisfaction is more strongly related to voluntary absenteeism than involuntary
absenteeism (e.g., failure to show for work because of illness, accidents, or other
unavoidable events). The low correlations found between job satisfaction and attendance
behavior is probably due in some part lumping together of both types of behavior.

Also, the decision to show up at work on time is more complex than simply a matter of
how satisfied the employee is with his or her job. Chadwick-Jones, Nicholson, and
Brown (1982) attempted to capture this complexity in a model of employee attendance in
which absenteeism was depicted as a consequence of social exchange processes. The
Chadwick-Jones et al. model assumes that an exchange relationship exists between the
organization (employer or supervisor) and the employee. This social exchange is
influenced by group norms and expectations. In any organization, the employee soon
discovers how much absenteeism is tolerated and the price that he or she will pay for
absenteeism. For example, an employee may learn that one or two unexcused absences
per month are tolerated if he is willing to work overtime and weekends when needed.

2. Turnover. As in the case of absenteeism and performance, the correlation of work-
related attitudes including job satisfaction with turnover is small. The corrected
correlation between job satisfaction and turnover revealed in meta-analyses of the
research is only -.14 (Tett & Meyer, 1993).

http://www.kronos.com/ads/absence/38/Kronos_absence_out_sick

Again, the low correlations found between job satisfaction and turnover probably are due
in part to the failure to distinguish between voluntary turnover and involuntary turnover.
A higher correlation seems likely if the employee can freely choose whether they stay or
go. This possibility was examined in a survey of 75,000 massage therapists that examined
the relation between job satisfaction and voluntary and involuntary intent to leave the
occupation (Blau, et al, 2009). A correlation of -.45 was found between job satisfaction
and voluntary intent to turnover, a relationship that is much higher than typically found.
Authors of another study examined the relation to turnover of HRM practices likely to
enhance and detract from job satisfaction (Shaw, Delery, Jenkins & Gupta, 1998). The
researchers surveyed the rate of turnover among truck drivers in 227 trucking firms and
distinguished between voluntary turnover in the form of quit rates among drivers and
involuntary turnover in the form of dismissal of drivers. To the extent that HRM directors
in these firms reported adherence to practices that were procedurally just and reported
higher pay benefits, there was a lower voluntary turnover. To the extent that the directors
reported electronic monitoring of drivers and higher required time on the road, both of
which are viewed as potential sources of dissatisfaction, there was higher voluntary
turnover. Involuntary turnover was related to a different set of predictors. In particular,
the use of valid selection procedures in the hiring of drivers, combined with a low
selection ratio (i.e., there are few positions to fill and a high number of applicants for
these positions), was associated with a low involuntary turnover rate in the form of
dismissal.

Another likely reason that job satisfaction and turnover is so low is that the process by
which satisfaction leads to turnover is complex and involves a variety of mediator and
moderator variables. Considerable research on job satisfaction and turnover has found
that dissatisfied workers are prone to quit their jobs, but this relationship is both indirect
and complex. One complication is that turnover is a process that unfolds over time and
the relation to satisfaction is mediated by a complex set of events. Mobley and his
colleagues (Mobley, 1977; Mobley, Horner, & Hollingsworth, 1978) developed a model
of the turnover process that attempts to explain the indirect and complex link between
satisfaction and turnover (see figure 5.13).

According to the model, dissatisfaction with the job triggers thoughts of quitting and a
search for another job. If the search process reveals favorable and available alternatives,
the employee develops an intention to quit, which directly predicts quitting (a type of
turnover). However, if the search process reveals unfavorable and/or no available
alternatives, the employee develops instead an intention to stay on the job, which directly
predicts staying. In a test of the models’ validity, researchers collected data on job
satisfaction, thinking of quitting, intention to search, intention to quit and turnover 47
weeks later with 200 hospital employees (Mobley, Horner, and Hollingsworth (1978).
The results were supportive of the model: Job satisfaction was more highly related to
thinking of quitting and intention to search than to actual turnover. However, intention to
quit or stay was significantly related to actual turnover.

Turnover is costly insofar as the employer must replace those who leave and train the
replacements, and the remaining employees must pick up the slack created by those who
leave. But is turnover always bad in the long run for the effectiveness of the
organization? Not necessarily. One study of retail sales employees found that over half of
the salespeople who left were rated by their supervisors as marginal or unsatisfactory in
their performance (Hollenbeck & Williams, 1986). The researchers concluded from these
findings that half of the turnover was functional. An organization can benefit from
functional turnover when the poorer performing employees leave and more capable
employees are found to replace them.

Another factor complicating the relationship between satisfaction and turnover is the
external labor market. Because of the personal risks involved in quitting one’s job (such
as not being able to meet financial obligations), turnover, more than absenteeism, is
affected by the availability of alternative jobs. Turnover is higher in good economic
times, when many jobs are available, than in poorer economic times, when fewer jobs are
available (Carsten & Spector, 1987). As a consequence, the correlation between job
satisfaction and turnover is likely to be stronger during economic prosperity than during
economic adversity. Because workers cannot freely leave the organization during periods
of economic adversity, the turnover rate is not a good indicator of workers’ true feelings
about whether they would prefer to leave or stay. The author of this text observed a
public sector organization that made this incorrect assumption, almost costing the
organization its most valued employees. Management assumed, because the agency’s
turnover rate had dropped considerably, that pending employee development programs
were unnecessary. Fortunately, data about current market conditions and current
employee attitude statistics in the agency convinced management that employees were
still disgruntled and would leave as soon as jobs outside the agency became available.
The employee development programs then proceeded as originally scheduled. In this
case, management incorrectly assumed that employees were satisfied when, in fact,
dissatisfaction was rampant but few employees were leaving because of the lack of
alternative jobs.

A third complicating factor is the extent to which individuals are embedded in
relationships within the organization. Individuals who are highly embedded in their jobs
and organizations are tightly connected to social situations that keep them in their jobs
and organizations (Mitchell, Holtom, Lee, Sablynski & Erez, 2001). Job embeddedness
increases with the number of close social relationships, the fit of individual values and
goals to the goals and values of the organization, and the degree of sacrifice that would
be required if the individual left the organization. Embeddedness can originate from
links, fit, and sacrifice associated with nonwork as well as work situations.

Figure 5.13: Mobley’s Intermediate Linkage Model of Turnover.
An employee who is unhappy with his job and has few relationships that bind him to the
organization, might still stick with the organization because of close ties to family,

friends, and the community. The research has generally supported the importance of
embeddedness. In a meta-analysis involving over 40 samples and over 30,000 employees,
on-the-job embeddedness and off-the-job embeddedness were predictive of turnover
(Jiang, Liu, McKay, Lee & Mitchell, 2012). On-the-job embeddedness had an
uncorrected correlation of -.15 (p < .05) and a corrected correlation of -.19 with turnover. Off-the-job embeddedness had an uncorrected correlation of -.10 (p < .05) and a corrected correlation of - .12 with turnover. There was also evidence that intentions to turnover mediate the relation of embeddedness to turnover. The less embedded an individual, the stronger their intention to leave. Intentions, in turn, lead to lower job performance and a search for alternative jobs. In the final link in the model, lower performance and search behavior lead to actual turnover. 3. Interrelationships among absenteeism, lateness, and turnover. This discussion has lumped absenteeism, lateness, and turnover together under the rubric of withdrawal behavior. This leads to the question of whether these behaviors are, in fact, interrelated. Apparently they are not highly interrelated. In one meta-analysis, the researchers report that the corrected correlations were only .26 between lateness and absenteeism and .25 between absenteeism and turnover (Berry, Lelchook & Clark, 2012). The corrected correlation between lateness and turnover was a mere .01. The relation of job satisfaction and organizational citizenship behavior (OCB). In any organization or group some people are cooperative, while others only do the minimum in their dealings with their fellow employees. Some employees not only cooperate but generally go out of their way to contribute to the group and the organization even when it is not required in their job descriptions. The latter type of behavior has gone by a variety of labels, including organizational citizenship (Organ, 1988), prosocial behavior (Brief & Motowidlo, 1986), and extrarole behavior (Katz & Kahn, 1978). There are small differences in the definitions of these concepts, but they are all concerned with individuals going beyond what is required in their jobs and attempting to contribute to the organization. This chapter uses the term organizational citizenship behavior (OCB), and defines it as behavior that employees perform with the intention of benefitting the organization or individuals. Additionally, employees perceive OCBs as behaviors that they are not required to perform and do not expect personal gains when they do. Among the behaviors that illustrate OCBs are assisting coworkers with job-related and personal matters, helping customers, and representing the organization favorably to outsiders (Brief & Motowidlo, 1986). Given that organizations are only effective to the extent that employees are willing to collaborate, how can an organization cultivate OCBs? The high correlation between job satisfaction and OCBs suggests that the best strategy is to create a satisfying work environment. OCBs appear less likely to emerge in a high-stress environment than in a low-stress environment. Leaders who are considerate and employee oriented are more likely to encourage such behavior among subordinates than inconsiderate taskmasters. Also, the warm and friendly atmosphere that exists in some organizations encourages OCBs, whereas the cold, punitive atmosphere of other organizations discourages such behavior. Researchers in one study found that salespersons were more likely to show TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 307 OCBs in the form of prosocial behavior when they felt secure in their work situation and trusted management (Puffer, 1987). Points to ponder: 1. Describe the potential outcomes of job satisfaction. Describe how these “outcomes’ could act as causes of satisfaction rather than vice versa. 2. What are the variables moderating the relationship between satisfaction and performance? 3. Although there is a small positive correlation between satisfaction and performance, dissatisfaction can lead to high performance and satisfaction to low performance in some circumstances. Describe and illustrate how this might occur. 4. In what ways can higher turnover benefit an organization? 5. One should not assume based on turnover and absenteeism that employees are satisfied or dissatisfied with their jobs. Why not? Job Involvement So far the discussion has focused on one work-related attitude, job satisfaction. Another work-related attitude that has received attention by I/O psychologists is job involvement, which is defined as the extent to which a person psychologically identifies with his or her job (Kanungo, 1982; Lodahl & Kejuer, 1965; Rabinowitz & Hall, 1977). Jobs occupy a central role in job-involved people's lives. There are benefits and costs of high job- involvement. On the positive side, high job-involvement leads to higher levels of effort. One can identify highly job-involved workers as the people who work overtime to finish a project and take pride in their job-related accomplishments or products. On the negative side, any change in employment conditions or status can seriously affect their psychological well-being or self-esteem. When these people are unable to perform their jobs, for example, because of illness, layoffs, or retirement, they may become discontented or depressed. Measurement of job involvement The most widely used measure of job involvement is a 20-item scale developed by Lodahl and Kejner (1965) (see table 5.7). They administered the 20 items to researchers (Cook et al., 1981) reported that the scale (both the full 20 items and shorter versions) possesses acceptable reliability and validity. Researchers also found that the involvement scale is multidimensional, although they have not systematically attempted to investigate the different dimensions. Probably because the overall measurement qualities of the total involvement scale are adequate, little interest has been generated to explore these different dimensions conceptually. As a result, more than 25 years later researchers are still not certain exactly what Lodahl and Kejner's 20-item scale measures. In their review of the job involvement literature, Rabinowitz and Hall (1977) concluded that some of the scale items measure a stable value orientation and some measure situational or specific job factors. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 308 The implied differences between the value orientation items and the situational items in the job involvement scale also present problems in providing feedback to organizations about the job involvement of their employees. If the value orientation items are more important in defining job involvement, then organizations could do little to change employee job involvement. The most practical approach, in this case, is to select workers who initially express high levels of job involvement. If, however, the situation is the more important determinant of job involvement, organizations should consider increasing employee involvement through various organizational change strategies. The empirical research tends to support both the value orientation and situational perspectives (Noe & Schmitt, 1986; Noe & Steffy, 1987). Table 5.7: Sample Items for Lodahl & Kejner Job Involvement Scale Correlates of job involvement The variables related to job involvement are generally similar in direction and magnitude to those found for job satisfaction (Cook et al., 1981; Brooke, Russell, & Price, 1988). Job involvement shows a small positive relationship to job performance. Older workers report higher levels of job involvement than younger workers across many types of workers and cultures (Morrow & McElroy, 1987; Rhodes, 1983). No sex differences are apparent in job involvement: Women report similar levels of job involvement to men (Chusmir, 1982; Rabinowitz & Hall, 1977). The behavioral correlates of absenteeism and turnover are also associated with job involvement; highly job-involved workers are less apt to be absent from work or quit their jobs (Brooke, 1986). TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 309 Points to ponder: 1. Is it healthy to base one’s self-esteem on one’s work? Why or why not? 2. Is your self-esteem based mostly on your work? If not, what are the sources of your self-esteem? 3. Consider the jobs you have held. In which of these was your job-involvement high? In which of them was your job involvement low? Why did you possess the level of job- involvement that you possessed? 4. Do you think employers should attempt to increase job-involvement so that employees base their self-esteem on how well they perform? Organizational Commitment Another work-related attitude that has received increasing attention is organizational commitment, which is defined as a workers' identification with and attachment to a particular organization. Like the highly job-involved individual, the highly committed worker takes his or her work seriously. However, the allegiances of the highly committed worker reside with the organization, not the job or work. Three components of organizational commitment Organizational commitment embodies three concepts: readiness to exert effort on behalf of the organization, acceptance of organizational goals and values, and desire to remain with the organization (Cook et al., 1981). Although, similar to job satisfaction, organizational commitment is a more global and enduring employee attitudes about the whole organization and therefore is probably less influenced by daily events (e.g., a disagreement with the supervisor). Researchers have proposed and investigated different dimensions of organizational commitment. One dimension is affective commitment, which is an emotional attachment to and involvement and identification with the organization. A second dimension is behavioral and is called continuance commitment. This dimension refers to the perceived costs associated with leaving the organization, such as giving up pension plans and profit sharing (Becker, 1960; Hrebiniak & Alutto, 1972). A third dimension is normative commitment, which is a felt obligation on the part of the employee to stay with the organization. The items associated with the three commitment dimensions are summarized in table 5.8. Recent research has empirically supported theoretical distinctions between the three types of commitment (Allen & Meyer, 1990; McGee & Ford, 1987; Meyer, Paunonen, Gellatly, Goffin, & Jackson, 1989). In the Meyer et al (1989) study, affective commitment correlated positively with job performance and continuance commitment correlated negatively with job performance in a sample of first-level managers in a food service company. Specifically, high emotional attachment to the organization was associated with high performance, whereas high behavioral attachment (attachment only for extrinsic benefits) was associated with low performance. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 310 Table 5.8: Sample Items from the Allen and Meyer Organizational Commitment Questionnaire Correlates of organizational commitment The correlates of organizational commitment depend on which component of organizational commitment that is measured (Brooke et al, 1988; Cook et al., 1981). Figure 5.14 summarizes the research findings on the antecedents, correlates, and consequences of affective, continuance, and normative commitment to the organization. The associations found for affective commitment are particularly noteworthy: Individuals with high affective commitment are less likely to be absent from or quit their jobs compared with their less committed coworkers (e.g., Williams & Hazer, 1986; Farkas & Tetrick, 1989). They are also likely to show higher levels of performance, well-being, health, and organizational citizenship behavior. Continuance commitment is negatively related to turnover but shows mixed findings for other outcomes. Normative commitment related to the outcomes in a direction similar to affective commitment but much less research has been conducted for this component of commitment than for affective commitment. Romzek (1989) found that those individuals who reported higher levels of organizational commitment also reported higher nonwork and career satisfaction two years later. TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 311 +, 0, and – indicate respectively a positive, zero, and negative relationship between the variables. Figure 5.14: Three-Component Model of Organizational Commitment. An issue that remains unexplored is the possibility that there is a downside to overcommitment. Over 50 years ago, Whyte (1956) warned about this danger in his book The Organization Man. Whyte claimed that the "organization man" identified so totally with his work group or organization that he substituted the organization's goals and beliefs for his own. In essence, according to Whyte, the overcommitted worker submerged his own identity for the good of the organization. The potential damaging effects on well-being, ethics, and creativity are easy to imagine. Subsequently, researchers have echoed Whyte's concerns (Mowday, Porter, & Steers, 1982; Randall, 1987). Despite the potential problems associated with too much commitment, researchers tend to view commitment as an unmitigated good, particularly for the TEXT COPYRIGHT ROBERT L. DIPBOYE 2016 ALL RIGHTS RESERVED 312 organization (e.g., Lawrence, 1958; Ouchi & Wilkins, 1985). Future research on commitment will perhaps examine the limits of employee commitment. Commitment as an exchange process An interesting theoretical view of organizational commitment, like the Chadwick-Jones approach to absenteeism, treats commitment as a social exchange process. Specifically, workers become more committed to their employing organization if they perceive that the organization values their contributions and cares about their well-being. Eisenberger, Huntington, Hutchison, and Sowa (1986) provide support for this exchange process and conclude that "employees' commitment to the organization is strongly influenced by their perception of the organization's commitment to them" (p 500). Over the last several decades some observers suggest that the management of organizations has changed the psychological contract that employees have with employers. No longer can employees count on employment for life in an organization but can look forward to a career of layoffs, reorganizations, acquisitions, and other disruptions. This violation of psychological contract poses a threat to productivity and innovation. As Kanter (1989) notes, "If people are encouraged to rely on themselves, then how can the corporation rely on them? A response is urgently require