Excerpts From:
MOTIVATING CRITICAL COMPUTER SYSTEMS OPERATORS:
JOB CHARACTERISTICS, CONTROLS, AND RELATIONSHIPS
A THESIS
SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL
OF THE UNIVERSITY OF MINNESOTA
BY
DALE HARRISON MCKNIGHT
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
Chair: Gordon B. Davis Co-Advisors: Norman L. Chervany and Fred D. Davis
Committee Members: Frank Miller, Akbar Zaheer
December, 1997
Note -- These excerpts include:
--The first in-depth examination of the Critical Information Systems Operator job
--Incrementing the Job Characteristics Model with Relationships/Trust
--Incrementing Management Contols theory with Relationships/Trust
--New Grounded Theory validity methods
--Demonstrations of thorough survey Construct Validity methods
--Empirical results that explain what motivates critical systems operators
--An explanation of the paradoxical results found for managerial controls
Copyright Dale Harrison McKnight 1997
All Rights Reserved
MOTIVATING CRITICAL COMPUTER SYSTEMS OPERATORS:
JOB CHARACTERISTICS, CONTROLS, AND RELATIONSHIPS
ABSTRACT
This study expands the explanatory power of two theories of motivation: the Hackman and Oldham Job Characteristics Model (JCM) and the economics-based Management Controls model (MCM). The JCM predicts worker motivation as a function of the worker’s job characteristics (e.g., skill variety), while the MCM predicts motivation as a function of managerial controls (e.g., incentives). These motivation theories each omit an explicit account of the roles of: a) supervisor/subordinate relationships, and b) workplace fairness perceptions, relying instead on how the job or its incentives are structured. This study adds explanatory power to these theories through two constructs: ‘Relationships’ (worker trust and liking towards the supervisor) and ‘System Trust’ (worker beliefs about the fairness structures of the workplace). The target application of this research is the critical computer systems operator. ‘Critical’ means the extent to which business transactions are interrupted when these systems are not available to their users.
This research was conducted in two phases at one site. Phase I explored factors important to keeping critical computer systems available to users almost 100% of the time. “Grounded theory” methods were used to analyze the semi-structured interviews. In Phase II, a questionnaire was administered to eighty-six operators to test the extent to which adding Relationships and System Trust to the JCM and MCM helped these models predict operator motivation.
The study contributes to research in four ways. First, Relationships and System Trust added predictive power to the JCM. Second, Relationships and System Trust added predictive power to the MCM. Relationships and System Trust supplement traditional views that job characteristics or management controls alone produce motivated workers. Third, the study validates measures for two newly conceptualized constructs: Relationships and System Trust. Fourth, it describes the highly motivating nature of the critical computer systems operator job.
This study also contributes to practice. Two paradigms have dominated recent corporate motivation practices: worker empowerment (based on the JCM) and incentive pay (based on the MCM). This research suggests that these paradigms will yield inadequate results unless worker/manager relationships and workplace fairness are also considered.
MOTIVATING CRITICAL COMPUTER SYSTEMS OPERATORS:
JOB CHARACTERISTICS, CONTROLS, AND RELATIONSHIPS
TABLE OF CONTENTS Page
Abstract ii
Chapter One: Introduction and Overview 1
Overview and Research Question 1
The Nature of the Critical Systems Operator Job 4
Advancing the Job Characteristics Model 6
Advancing the Management Controls Model 9
Summary and Contributions 12
Roadmap for the Study 13
Chapter Two: Methodology and Construct Validation 15
Approach of the Study 15
Phase I Conceptual Model Building Methodology 20
· Grounded Theory 20
· Phase I Research Framework 22
Phase II Model Building and Testing Methodology 24
· Measurable Constructs Used 25
· Instrument Pretest 27
· Instrument Pilot 31
Construct Validation Results 33
Hypothesis Testing Methodology 45
Research Site for the Study 49
Chapter Three: Nature of the Critical Systems Operator Job 55
Theory Building 56
· Nature and Importance of the Critical Computer System 56
· Management Information Systems Literature 59
· Management of Technology Literature 62
· Conceptual Model Building 64
· Hypotheses 66
Methodology Detail 76
Results of Hypothesis Testing 79
Discussion of Results 84
TABLE OF CONTENTS (continued)
Chapter Four: Job Characteristic Model--Adding Relationships 88
Theory Building 89
· JCM Related Research 89
· JCM Hypotheses 90
· Relationships and System Trust Related Hypotheses 91
Methodology Detail 96
Results of Hypothesis Testing 96
Discussion of Results 100
Chapter Five: Incentive Controls--Adding Relationships 103
Theory Building 104
· Definitions 104
· Controls Theory Overview 105
· Conceptual Model Building-Incentives 107
· Scientific Model Building-Incentives 110
· Hypotheses-Incentives 119
Methodology Detail 123
Results of Hypothesis Testing 125
Discussion of Results 127
Chapter Six: Other Controls--Adding Relationships 133
Theory Building 134
· Conceptual Model Building-Accountability 134
· Scientific Model Building-Accountability 135
· Conceptual Model Building-Feedback 137
· Scientific Model Building-Feedback 137
· Conceptual Model Building-Micromanagement 140
· Scientific Model Building- Micromanagement 141
· Conceptual Model Building-Autonomy 142
· Scientific Model Building- Autonomy 142
· Scientific Model Building-Work Outcomes 143
· System Trust’s Impact on Motivation 145
Methodology Detail 145
Results of Hypothesis Testing 146
Discussion of Results 153
TABLE OF CONTENTS (continued)
Chapter Seven: Contributions, Limitations, and Future Research 157
Contributions 158
· To Theory 158
· To Practice 160
Study Limitations 163
· External Validity 164
Future Research 167
References 172
Appendix A Examples of open and axial coding 191
Appendix B Questionnaire Items by Construct 193
Appendix C Operator Questionnaire 201
Appendix D Supervisor Questionnaire 224
Appendix E Pretest Instrument A--Matching 227
Appendix F Pretest Instrument B--Categorization 228
Appendix G Pretest Instrument C--Sorting 229
Appendix H Pairwise Intercorrelation Matrices 230
Appendix I Descriptive Statistics 244
Appendix J Pairwise Intercorrelation Matrices--High Level Concepts 245
LIST OF TABLES
Page
Table 1 Instruments for Testing Management Controls/Relationships 26 Model
Table 2 Pilot Reliability Analysis 32
Table 3 Construct Level Cronbach’s Alpha Reliabilities 34
Table 4 Intercorrelations of Trust Constructs and Liking 38
Table 5 Mono-Trait, Mono-Method Analysis for Autonomy 41
Table 6 Correlations among CTE, Performance, and Two Autonomy Types 43
Table 7 Correlations among JCM Variables and Two Autonomy Types 43
Table 8 Reliabilities for High Level (Second Order) Concepts 46
Table 9 Intrinsic Motivation Orientation (IMO) Scale 78
Table 10 Job Characteristics Comparisons 79
Table 11 Intrinsic Versus Extrinsic Factors Reported 81
Table 12 Correlations between Less Secure Group and Other Attributes 82
Table 13 Job Characteristics Model Test Results 97
Table 14 Relationships and System Trust Test Results 99
Table 15 Effects of Extrinsic Motivation on Intrinsically Motivating Tasks 115
Table 16 Management Controls / Relationships Model—Correlation Tables 147
Table 17 Management Controls / Relationships Model—Regression Results 148
Table 18 Sensitivity Analysis for Relationships Moderation of Accountability 150
Table 19 Sensitivity Analysis for Relationships Moderation of Feedback 151
LIST OF TABLES (continued)
Table 20 Sensitivity Analysis for Relationships Moderation of
Micromanagement 151
Table 21 Sensitivity Analysis for Relationships Moderation of Autonomy 152
COMMONLY USED ABBREVIATIONS
JCM Job Characteristics Model
MCM Management Controls model
CSO Critical (computer) System Operator
MIS Management Information Systems
XYZCo Organization for the research site
GNS Growth Need Strength
CPS Critical Psychological States
LIST OF FIGURES
Page
Figure 1 Job Characteristics Model (JCM) 2
Figure 2 Motivating Nature of the Critical Systems Operator Job 6
Figure 3 Expanding the Job Characteristics Model 9
Figure 4 Management Controls Model 10
Figure 5 Advancing the Management Controls Model 12
Figure 6 Roadmap for the Study 14
Figure 7 The Operations Research Model 19
Figure 8 Phase I Research Framework 23
Figure 9 Management Controls / Relationships Model—Detailed Level 25
Figure 10 Job Characteristics Model (JCM)—Detailed Level 27
Figure 11 Nomological Network for Trust Constructs 37
Figure 12 Model of Construct Creation 192
Figure 13 Model of Construct Linkages 192
39
CHAPTER ONE: INTRODUCTION AND OVERVIEW
This chapter previews the topic, propositions, general methodology, and contributions of the study. It begins with a research overview that introduces the research question. Then it creates the broad propositions that later chapters will test in detail. Finally, it summarizes the contributions of the study and presents a guide that organizes the contents of later chapters.
Overview and Research question
This study introduces two constructs, Relationships and System Trust, that improve the predictive power of the Job Characteristics Model (JCM) and the Management Controls model (MCM) of worker motivation. System Trust means the belief that proper impersonal structures are in place to enable one to anticipate a successful endeavor (Lewis & Weigert, 1985; Shapiro, 1987; Zucker, 1986). In this study, the Systems Trust construct was operationalized as the worker’s belief that structures (i.e., processes, procedures) support or encourage fairness in one’s work environment. Relationships means the extent to which one holds positive feelings, beliefs and intentions towards another person. The Relationships construct was operationalized as trust in, and liking of, one’s supervisor. The Relationships definition carries a quality-of-relation focus that differs from the traditional definitions of relationships in: a) sociology, which focus more on behavioral and role interdependence (e.g., Blau, 1964), and b) social psychology, which focuses more on the ability of parties to influence each other (e.g., Berscheid, 1983).
As depicted in Figure 1, the Hackman and Oldham (1975) Job Characteristics Model (JCM) posits that worker perceptions of their Job Characteristics (e.g., Skill Variety) lead to Critical Psychological States (e.g., Felt Responsibility) that, in turn, lead to motivational Work Outcomes (e.g., Job Satisfaction). These model linkages are moderated by the worker’s Growth Need Strength, an individual characteristic variable. The JCM focuses on the nature of the job itself, ignoring social or structural aspects of the worker’s environment.
Figure 1 Job Characteristics Model (JCM)
Growth Need Strength
Critical
Job Characteristics Psychological Work
States (CPS) Outcomes
In contrast, Management Controls models (e.g., Ouchi, 1979) posit that incentives or other controls improve worker motivation. The term “controls” means methods of attempting to ensure desired outcomes by trying to influence other people (Anthony, 1965; Lawler & Rhode, 1976). Management control occurs when managers use methods to try to influence employees to behave in certain ways. Control models generally ignore social and structural factors, focusing instead on extrinsic rewards or behavior control. For example, managers try to entice employees to work faster by offering them contingent financial incentives.
This study’s subjects were critical systems operators. Critical systems are computer systems that must be kept available to users, or else numerous business or operational transactions are interrupted. Transaction processing systems, used to conduct a firm’s daily business, often fall in the category of critical systems (Laudon & Laudon, 1995). Managers of critical systems try to keep their systems continuously available to system users. Hence, critical systems operators (CSOs) must be constantly alert to problems that might threaten the system. When a critical systems crashes, the operators are charged with restoring it within seconds or minutes, not hours. The researcher studied critical computer systems operators (CSOs) in two stages: exploratory (Phase I) and confirmatory (Phase II). The systems these operators ran were considered critical because thousands of users required that the systems be continuously available so they could perform their daily job function.
During the study’s Phase I interviews, it became evident that CSOs were clearly motivated by the nature of their job, but that controls and incentives did not have consistent, positive motivational effects on CSOs. In analyzing Phase I data, it became clear that worker relationships with superiors and their beliefs about the work environment also influenced their motivation. Some evidence for this effect also comes from the management literature (e.g., Cook & Wall, 1980; Locke, Latham & Erez, 1988). Therefore, the study’s research question is:
Do operator/supervisor Relationships and System Trust improve the ability of the Job Characteristics Model and the Management Controls model to predict critical systems operator motivation and motivational outcomes?
In other words, this study tested the extent to which operator/supervisor Relationships and System Trust added predictive value to the JCM and the MCM in the critical systems operator context.
The Nature of the Critical Systems Operator Job
The critical computer systems operator (CSO) is a subset of the class of information systems workers called “computer operators.” A literature search revealed that very little research has been done on computer operators. The management information system (MIS) literature focuses on system development, implementation, maintenance, and use issues, while covering few system operation issues (Berkeley, 1984; Ives, Hamilton and Davis, 1980; Swanson & Ramiller, 1993). Lyytinen & Hirschheim's (1987) exhaustive review of the MIS failure literature reported almost no research on system operation issues.
In fact, in the 1970s and 1980s, the traditional computer operator job was viewed as a quasi-clerical function that did not merit intensive study (Couger & Zawacki, 1980). In their survey of over 1200 computer operations employees, Couger and Zawacki (1980: 33) reported that “employees in DP operations perceive their jobs to be deficient in the key characteristics that produce motivation and lead to increased productivity. The motivating potential score (MPS) of these jobs is lower than that of any of the other 500 jobs in the Hackman/Oldham data base.” MPS, derived from the scores of the five JCM job characteristics, represents how motivating a job is. Describing computer operations as a data processing ‘stepchild,’ Couger and Zawacki suggested that only “the ‘sledgehammer’ of a catastrophic event such as a flood or bombing” could “draw attention to computer operations.” (1980: 34)
This study draws attention to the job of the critical systems operator—a job that does not fit the Couger and Zawacki computer operator profile. In the critical system context, the threat of catastrophic system downtime is so large that it produced motivating potential scores for the eighty-six critical systems operators in this study that were more than double that of the traditional computer operator Couger and Zawacki measured. This study’s informants operated three critical transaction processing systems at a large U. S. corporation fictionally name XYZCo. During Phase I interviews, critical systems operators (CSOs) at XYZCo were found to be highly skilled and motivated individuals who performed an extremely interesting and challenging job. For example, the task of diagnosing and fixing system outages was reported to be exhilarating, satisfying, and yet full of pressure. These CSOs were found to be primarily intrinsically motivated, in that they more often mentioned that they enjoyed their job and its challenge than they mentioned extrinsic job rewards. From Phase I data (discussed in more detail in Chapter Three), it was proposed that (see Figure 2):