MSU LIBRARIES “ RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. THE ANTECEDENTS AND CONSEQUENCES OF PERSONAL CONTROL BY Joseph Thomas McCune A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1984 ABSTRACT THE ANTECEDENTS AND CONSEQUENCES OF PERSONAL CONTROL BY Joseph Thomas McCune The purpose of this study was to increase our under- standing of the construct of personal control in organiza- tions. Personal control was defined as one's perception of freedom in and control over work activities, events and outcomes. The important antecedents and consequences of the three dimensions of control (i.e. outcome control, ac- tivity control and perceived influence) were identified in a review of the research literature and used to develop and test a mediational model of personal control in a field setting. Questionnaire responses of 423 faculty (24 percent return rate) and 655 clerical workers (40 percent return rate) were analyzed using correlation and regression analyses. An examination of the personal control scale in- tercorrelations and their pattern of correlations with the antecedent and outcome variables supported the multidimen- sional conceptualization of personal control. The importance of personal control was demonstrated by the high correlations between the personal control scales and the outcome variables (e.g. satisfaction, psychological strain and turnover intention). Further, the multidimen- sional conceptualization of personal control explained more Joseph Thomas McCune variance in the outcome variables than any of the personal control scales alone. However, only partial support was found for the mediational model of personal control. Finally, the limitations, as well as the theoretical and practical implications of the study, were discussed and recommendations were made regarding future research involv- ing personal control. To Laura and in memory of my grandmother, Florence McCune ACKNOWLEDGMENTS I would like to take this opportunity to acknowledge and thank those individuals most responsible for my achieving a Ph.D. First and foremost, I would like to thank my dissertation committee, who, besides being my teachers and advisors, have also been my close friends. The guidance and support of my committee members through- out the dissertation process has made this project more bearable. I would like to thank Kevin Ford, who kindly agreed to serve on my committee at the last minute and greatly helped to improve the quality of the final product. The friendly advice and helpful criticisms of Steve Kozlowski throughout the dissertation procedure helped keep my dis- sertation focused and technically correct. I regret that John Wanous was not able to serve on my committee for its entire duration. His contributions, however, provided me with a different perSpective that enriched my present study and suggested future studies as well. Mary Zalesny provided an insight into this project that enabled me to sharpen my focus and helped broaden my understanding of the topic. Mary's comments and attention to detail were very helpful in making revisions, which greatly improved iii the readability and accuracy of the final paper. Above all, I would like to thank my chair, Neal Schmitt. Neal's advice, guidance, and support greatly exceeded the responsibilities of a dissertation chair. He was always available and willing to critique my work, help with problems, and provide encouragement. Neal also served as my advisor during my four years at Michigan State, and I owe much of my development as an I/O Psych- ologist to him. I would like to take this opportunity to express my sincere appreciation to my parents, Ruth and Edward, for their love and support, for encouraging my pursuit of knowledge, and instilling in me the value of an education. I would also like to thank my in—laws, Ethel and Lawrence Schembri, for their encouragement and support throughout my graduate work. And finally, I would like to express my thanks and love to my wife, Laura. Her love and friendship, not to mention her typing and editing, made my work so much easier. Further, her tolerance of my working nights and weekends, as well as her support and encouragement on a day-to-day basis, has provided me with the strength to continue and made my dream of a Ph.D. a reality. For these things I will always be grateful to her. iv TABLE OF CONTENTS Page LIST OF TABLES . . . . . . . . . . . . . . . . . . . vii LIST OF FIGURES. . . . . . . . . . . . . . . . . . . X CHAPTER I INTRODUCTION. . . . . . . . . . . . . . . l The Importance of Control . . . . . . . . . . . 1 Organizational Research . . . . . . . . . . . . 4 Personal Control in Organizations . . . . . . . 8 Purpose of Study . . . . . . . . . . . . . . . 12 CHAPTER II LITERATURE REVIEW. . . . . . . . . . . . 15 Outcome Control . . . . . . . . . . . . . . . . 15 Antecedents of Outcome Control. . . . . . . . . 1? Consequences of Outcome Control . . . . . . . . 22 Activity Control. . . . . . . . . . . . . . . . 26 Antecedents of Activity Control . . . . . . . . 29 Consequences of Activity Control. . . . . . . . 36 Perceived Influence . . . . . . . . . . . . . . 4O Antecedents of Perceived Influence. . . . . . . 45 Consequences of Perceived Influence . . . . . . 52 Summary and Research Plan . . . . . . . . . . . 60 CHAPTER III METHOD. . . . . . . . . . . . . . . . . 83 Subjects. . . . . . . . . . . . . . . . . . . . 83 Procedure . . . . . . . . . . . . . . . . . . . 83 Instruments . . . . . . . . . . . . . . . . . . 83 Data Analysis . . . . . . . . . . . . . . . . . 90 CHAPTER IV .RESULTS AND DISCUSSION 93 ReSponse Rate . . . . . . . . . . . . . . . . 93 Representativeness of the Sample. . . . . . . . 93 Psychometric Properties of the New Personal Control Scales. . . . . . . . . . . 97 Internal Consistency. . . . . . . . . . . . 100 Comparability/Distinctiveness of the New Personal Control Scales. . . . . . . . . . . 100 Personal Control Scales/Dimensions Intercorrelations. . . . . . . . . . . . . . 103 External Consistency. . . . . . . . . . . . . . 105 Item-Scale Correlations . . . . . . . . . . . 107 Variance in the Dependent Variables Accounted for by the New Scales. . . . . . . 111 V Page CHAPTER IV RESULTS AND DISCUSSION (cont.) Test of Hypothesized Relationships Between Personal Control Dimensions and Ante- cedent and Outcome Variables . . . . . . . . 122 Test of the Personal Control Model. . . . . . . 128 CHAPTER V SUMMARY AND CONCLUSIONS . . . . . . . . . 145 Summary and Conclusions . . . . . . . . . . . . 145 The Personal Control Model. . . . . . . . . . . 150 Limitations of Study. . . . . . . . . . . . . . 153 Future Research . . . . . . . . . . . . . . . . 154 Practical Implications. . . . . . . . . . . . . 155 APPENDIX A DEMOGRAPHIC ITEMS. . . . . . . . . . . . 158 APPENDIX B ANTECEDENT VARIABLES . . . . . . . . . . 159 APPENDIX C PERSONAL CONTROL MEASURES. . . . . . . . 161 APPENDIX D OUTCOME VARIABLES MEASURES . . . . . . . 167 BIBLIOGRAPHY. O O O O O I O O O O O O O O O I O O O 175 vi Table 10 11 LISTS OF TABLES Summary of the Literature Reviewing the Relationship Between Antece- dent and Outcome Variables and the Personal Control Dimensions . Summary of the Literature Reviewing the Relationship Between Antece- dent and Outcome Variables and the Personal Control Dimensions . Hypothesized Relationships Between Antecedent, Personal Control and Outcome Variables . . . . . . . . Comparison of Sample and Population Demographic Characteristics for Faculty . . . . . . . . . . . . . Comparison of Sample and Population Demographic Characteristics for Clerical Workers. . . . . . . . . Chi Square Test of the Difference in Demographic Characteristics Between the Faculty Sample and Its Population. . . . . . . . . . Chi Square Test of the Differences in Demographic Characteristics Between the Clerical Worker Sample and Its Population . . . . Coefficient Alphas for the Per- sonal Control Scales. . . . . . . Intercorrelation Matrix of Demo- graphic, Antecedent, Personal Control and Outcome Variables . . Item-Scale Correlations for Per- sonal Control Measures. . . . . . Results of Hierarchical Multiple Regression Analyses of Parti- cipation/Influence Scales-- Vroom Psychological Partici- pation Scale (Vroom) and the Perceived Influence Scale (PI) on the External Variables . . . . vii Page(s) 61-63 70 73 95 96 98 99 101 104 108-109 112-114. Table 12 13 14 15 16 17 18 19 20 Results of Hierarchical Multiple Regression Analyses of Autonomy/ Activity Control Scales--JCI Autonomy Scale (JCI),JDS Autono- my Scale (JDS) and the Activity Control Scale (AC) on the Exter- nal Variables . . . . . . . . . . . Hypothesized Relationships and Em- pirical Correlations Between Personal Control Dimensions and Antecedent and Outcome Variables. . Results of the Hierarchical Multiple Regression Analyses of the Demo- graphic, Personal Control, Ante- cedent and Outcome Variables. . Results of Hierarchical Regression Analysis of Demographic, Antece- dent and Personal Control Vari- ables on Intrinsic Satisfaction . . Results of Hierarchical Regression Analysis of Demographic, Antece- dent and Personal Control Vari- ables on Extrinsic Satisfaction Results of Hierarchical Regression Analysis of Demographic, Antece- dent and Personal Control Vari- ables on Job Involvement . . . . . Results of Hierarchical Regression Analysis of Demographic, Antece- dent and Personal Control Vari- ables on Organizational Commitment Results of Hierarchical Regression Analysis of Demographic, Antece- dent and Personal Control Vari- ables on Physical Strain . . . . . Results of Hierarchical Regression Analysis of Demographic, Antece- dent and Personal Control Vari- ables on Psychological Strain . . . viii Page(s) 116-118 123 131 134 135 136 137 138 139 Table- 21 22 23 Results of Hierarchical Regression Analysis of Demographic, Antece- dent and Personal Control Vari- ables on Effort . . . . . . Results of Hierarchical Regression Analysis of Demographic, Antece- dent and Personal Control Vari- ables on Union Attitudes . Results of Hierarchical Regression Analysis of Demographic, Antece- dent and Personal Control Vari- ables on Turnover Intention ix Page(s) 140 141 142 Figure LIST OF FIGURES VIE Model . . . . . . . . . . . . . Determinants of E —9 P Expectancies Determinants of P —9 O Expectancies VIE Model . . . . . . . . . . . . . The Job Characteristics Model . . . The Antecedents of Activity Control Leader Behavior Continuum . . . . . Hypothesized relationships of indiv- idual and Organizational Variables and Role and Expectancy Percep- tions and Satisfaction and the Appropriateness of Participation in Decison Making. . . . . . . . Model of the Antecedents and Outcomes of Personal Control. . . . . . . Page(s) l6 l8 18 23 28 35 46 56 72 CHAPTER I INTRODUCTION The Importance of Control Theorists and researchers have long attested to the im- portance of the feeling of control. Angyal (1941) proposed that human beings have a tendency toward self-determination in that they strive to resist external influences and attempt to control the physical and social forces in their environment. Woodworth (1958) observed that individuals seem to exert a great deal of effort toward producing ef- fects on their environment even when these effects do not satisfy basic needs. Drawing on the work of Anygal and Woodworth, White (1959) theorized that individuals have an innate need to strive for "competence" through effective interactions with their environment. According to White, when individuals are able to produce changes in their en- vironment, they experience feelings of satisfaction and efficacy. May (1972) asserted that one needs a sense of mastery over one's fate to maintain feelings of self-esteem and well-being. Thus, the perception of control has been theorized to be an important human need or value that is necessary for one's sense of well-being. Laboratory researchers have demonstrated the importance of control with a variety of different research paradigms. Overmier and Seligman (1967) found that laboratory animals eXposed to inescapable electric shocks exhibited a severe aversive affective and behavioral reaction (i.e. learned 1 2 helplessness). Several different types of manipulations have been used to elicit the learned helplessness reaction, including long periods of restraint and monotony, punish- ment of an appetitive response, and punishment of mistakes on insoluble problems (Mineka and Kihlstrom, 1978). The factor common to all of these manipulations involves the organism's loss of the ability to control an important en- vironmental outcome. The explanation most often offered for these results is that the organism learns it cannot control (i.e. avoid or escape) the aversive events, and even when environmental contingencies change making control possible, the organism does not recognize the change in con- tingency and passively accepts its fate. Thus, the animal learns it is "helpless." Researchers have also investigated the learned help- lessness phenomena with human subjects. Fosco and Geer (1971) manipulated the amount of control that their sub- jects had by varying the number of insoluble problems each one received. Subjects were administered an electric shock for each problem they were unable to solve. Thus, the high- control groups were able to avoid more shocks than the low- control groups. The results indicated that the low-control groups performed significantly poorer on soluble problems given after the experimental manipulation than the high-con- trol groups. In similar studies, low-control subjects re- ported feeling frustrated and helpless (Roth and Bootzin, 1974); and helpless, passive and hostile (Krantz, Glass and L Cu :5 r1 \ ~< h.- 3 Snyder, 1974). Thus, laboratory research with both animals and humans indicates that exposure to low-control environ- ments can result in negative affective and behavioral reac- tions. The learned helplessness research, while demonstrating the negative effects of low-control environments, provides little direct evidence of the importance of individuals' feelings of control (i.e. personal.control). Geer, Davidson and Gatchel (1970) investigated the effects of per- sonal control in a study ostensibly designed to measure subjects' reaction times. In the first phase of the experi- ment, subjects were told to press a button at the onset of receiving a painful 6-second electric shock. In the second phase, the perceived control group was told that the dura- tion of the electric shock would be decreased if they were able to decrease their reaction time. The control group was told that the duration of the electric shock would be decreased for the remainder of the study. The duration of the electric shock was reduced to 3 seconds for both groups in the second phase of the experiment. The results indi- cated that the perceived control group had significantly lower levels of autonomic arousal in the second phase than did the control group. Thus, the belief that one is able to modify or reduce the occurrence of an aversive event, whether or not that belief is true, appears to ameliorate the effects of aversive stimulation. Other researchers have investigated the positive ef- "'1 If, .77? "1, 7‘" 4 fects of perceived control over aversive events. Glass and Singer (1972) manipulated perceived control by informing the perceived-control group that they could terminate the occur- rences of highly aversive noise for the remainder of the session by pushing a button. The control group was provided no such button or instructions. The researcher asked the perceived-control group not to use the button (and few did), so the button represented potential control. The results indicated that the perceived-control group had higher per- formance levels on proofreading tasks and reduced physio- logical reactions to the noise. In a similar study, Penne- baker, Burham, Schaeffer and Harper (1977) found that the perceived-control group reported fewer physical symptoms than the control group reported. These results provide evidence of the importance of perceived control to indi- viduals' performance and well-being. In sum, the perception of control over one's environ- ment has been theorized to be an important human need. Further laboratory researchers have demonstrated the nega- tive effects of exposure to low-control environments, as well as the positive effects of individuals' perceived con- trol over aversive events. These research results suggest that personal control could be an important variable in un- derstanding organizational behavior. Organizational Research The concept of control has assumed a somewhat different meaning in organizational research than the one used by 5 psychological theorists and laboratory researchers. Tannen- baum (1962) defined control as any process in which an in- dividual, group or organization determines (i.e.intentionally affects) what another individual, group or organization will do. In contrast, Tannenbaum conceptualized freedom as the extent to which an individual determines his or her own be- havior. Using Tannenbaum's terms, the psychological theor- ists and laboratory researchers, for the most part, defined control as freedom. The organizational researchers, in con- trast, define control as control. Organizational researchers tend to View control from the organization's perspective in the sense that organiza- tions attempt to control the activities and outcomes of its organizational members (Ouchi and Maguire, 1975). The purpose of organizational control is to maintain order, coordinate activities, ensure conformance to rules and facilitate achievement of organizational goals (Tannenbaum, 1962). The traditional management approach has been to direct and control employees as much as possible to ensure a stable and certain work flow. Time clocks, regulations, close supervision and simplified jobs are among the methods utilized to achieve stability and control. Thus, the con- trol of organizational members is an accepted and necessary function of management. A trend toward allowing employees greater discretion, however, has developed over the past three decades. Lawler (1976) has warned that some organizational control systems can produce dysfunctional effects, including employee re- 6 sistance to the programs, rigid bureaucratic behavior, and strategic behavior. In his study of alienation in organi- zations, Blauner (1964) found that bureaucratization, cen- tralization and rigid rules led workers to perceive little control over the methods they use to do their work, result- ing in a sense of powerlessness and alienation. Drawing on Blauner's research results and generalizing from the learned helplessness research conducted with both animal and human subjects, Martinko and Gardner (1982) developed a model of the determinants of organizationally induced helplessness (OIH). The OIH model proposes that certain organizational conditions, including centralized decision making, noncon- tingent reward systems, unrealistic work goals and low-scope jobs, are primary determinants of passive and maladaptive behavior in organizations. Thus, considerable evidence suggests that too much organizational control and not enough freedom in organizations may result in negative affective and behavioral reactions among organizational members. The importance of control in organizations, however, is not limited to the proposal that excessive control pro- duces aversive effects among organizational members. Human- istic theorists have argued for a more democratic or par- ticipative style of management to provide workers with the opportunity to satisfy higher order needs on the job (Likert, 1961; McGregor, 1960). Further, organizational researchers have recommended increasing the amount of con- trol employees have over a variety of work functions, in- 7 cluding decision-making operations (Vroom, 1960), the sett- ing of performance goals (French, Kay and Meyer, 1966), the selection of work methods and work pace (Hackman and Oldham, 1 980) and the choice of work rewards (Lawler, 1971, 1981). Several popular organizational development techniques (e.g. quality circles and semi-autonomous work groups) provide workers with increased control over their worklives (French and Bell, 1984). Some countries have even adopted legisla— tion mandating organizations to establish work councils and other participative work structures (Jenkins, 1973). In sum, a trend away from excessive organizational control of employees has developed along with a tendency for organizations to provide greater freedom for its mem— bers; It seemed likely that this trend toward increased free- dom and control for organizational members would be accom- panied by an increased awareness and concern for organiza- tional members' perception of control over their work envi- ronment (i.e. personal control). This does not appear to be the case. Despite the centrality of personal control in this trend, little research involving individuals' percep- tion of control has been conducted in organizations. The importance imputed to perceived control by psychological theorists and laboratory researchers further suggests the need for more research involving personal control. The present study will focus on the elucidation of the con- struct of personal control in organizations. 8 Personal Control in Organizations A major problem impeding the research involving con- trol is the variety of different meanings used to define personal control. Personal control will be defined in this study as an individual's perception of freedom in, and con- trol over his or her immediate environment. This defini- tion combines aspects of White's (1959) need for competence, Tannenbaum's (1962) definition of freedom and the laboratory researcher's conceptualization of personal control (e.g. Glass and Singer, 1972). In an organization, personal con- trol would likely include one's perception of freedom from the control of others, as well as his or her perception of control over work activities, materials and the rewards received for work. While this definition of personal control provides a useful starting point, it is not sufficiently detailed for it to be operationalized. Several theorists, however, have developed well-defined conceptualizations of related vari- ables--perceived freedom (Steiner, 1970) and personal con- trol (Bazerman, 1982). Perceived freedom is comprised of two independent dimensions--perceived outcome freedom and perceived decision freedom. Perceived outcome freedom in- volves one's judgment of the availability and desirabiltiy of the outcomes he wishes to obtain. Perceived decision freedom concerns one's perception of volition when deciding whether or not to seek a specific outcome and when choosing whether to seek one outcome rather than another. Steiner 9 (1970) viewed perceived freedom from an exchange theory perspective and proposed that high perceived freedom exist- ed when an individual perceived that his or her desired ac- tivities and outcomes were unimpeded by the necessity to expend energy or endure social sanctions. Bazerman (1982) conceptualized personal control as be- ing composed of two unique dimensions--activity control and outcome control. Activity control involves the discretion an organization provides the individual concerning the me- thods to use to perform his or her job. Outcome control is the degree to which outcomes are contingent on performance. Bazerman used Lawler's (1973) conceptualization of expec- tancy theory to describe the two components of outcome con- trol: 1) the effort to performance and 2) the performance to outcome expectancies. Bazerman's (1982) research focused on determining the optinual level of control that an organization should pro- vide to an employee. He proposed that the optimal control state is one in which an individual's ability to use con- trol is congruent with the amount of control provided him or her by the organization. In a laboratory study using college students, Bazerman found that performance was high- er in the congruent condition than either the under-control or over-control conditions. Bazerman's conceptualization of personal control pro- vides a useful framework for the development and operation- alization of the construct of personal control in an organ- 10 iZation. Further, Steiner's two dimensions of perceived freedom are included in Bazerman's conceptualization of personal control, making Bazerman's the more complete de- finition. Bazerman's definition of personal control has the added benefit that constructs similar to its two di- mensions of control-~outcome control and activity control-- have been operationalized and studied by other researchers. Outcome control has been studied as a component of the expectancy theory of motivation (Lawler, 1973; Vroom, 1964). The concept of activity control is essentially the same as that of Hackman and Oldham's (1975) job characteristic of autonomy. Bazerman's definition of personal control is also consistent with the definition of personal control de- rived from the work of the psychological theorists and lab- oratory researchers. Thus, Bazerman's definition of per- sonal control will be used to define and operationalize personal control in this study. While outcome control and activity control appear to be important components of personal control, they do not fully characterize personal control in an organization. The missing component in Bazerman's conceptualization of personal control is perceived influence. James, Gent,Hater and Cor ay (1979) defined perceived influence as the amount of influence an employee perceives he or she has over his or her supervisor's decisions. Being able to influence the decisions that impact on one's job would appear to be an integral component of personal control. 11 Organizational researchers have explored the relation- ship between individuals' perceived influence resulting from participation in decision making and several indivi- dual and organizational outcomes (e.g. job satisfaction) (T031, 1970; Vroom, 1960). Tannenbaum (1962, 1966) also investigated organizational members' perceptions 0f influ- ence in decision-making operations. Tannenbaum used mana- gers' perceptions of influence to develop control graphs for organizations. These graphs illustrated the total amount of control in the organization, as well as the steep- ness of the organization's hierarchy of control. Tannen- baum's work, while interesting, is not relevant here since his conceptualization of control is used to describe and understand organizations rather than individuals. Personal control will be defined as consisting of the following dimensions: 1. Outcome Control - the degree to which an individual believes he or she is able to cause or control impor- tant work outcomes and consists of: a) Effort to Performance Expectancy - an individual's subjectively-determined judgment of the probability that his or her effort will result in a certain level of performance. b) Performance to Outcome Expectancy - an individual's subjectively-determined judgment of the probability that a certain level of performance will lead to a particular outcome (Lawler, 1973). 12 2. Activity Control - the degree to which an individual perceives that his or her job provides substantial freedom, independence, and discretion in scheduling the work and in determining the procedures to be used in carrying it out (Hackman and Oldham, 1980). 3. Perceived Influence - the degree to which an individ- ual perceives him or herself as having an influence on decisions made by their supervisors (James, et a1., 1979; Vroom, 1960). Purpose of Study Personal control has been defined as one's perception of freedom in, and control over work activities, events and outcomes. Psychological theorists have proposed that per- sonal control is an important human need. Laboratory re— searchers have demonstrated the negative effects of exposure to low-control environments, as well as the positive effects of perceived control over aversive environmental outcomes. A trend toward providing organizational members with more freedom and control over their worklives has developed over the past three years. Unfortunately, little research has investigated personal control in an organizational setting. Bazerman (1982) developed a multidimensional conceptu- alization of personal control, which was used to provide a framework for the operationalization of personal control in this study. It was necessary, however, to include the con- struct of perceived influence along with Bazerman's two di- mensions of control to conform to this study's definition 13 of personal control derived from the work of psychological theorists (i.e. Tannenbaum, 1962; White, 1959) and labora- tory researchers (e.g. Glass and Singer, 1972). Thus, per- sonal control was defined as consisting of the following dimensions: outcome control, activity control and perceived influence. Constructs similar to each of these dimensions of control have been studied by organizational behavior re- searchers. Researchers, however, have failed to recognize the similarity of these three constructs because of the dif- ferent perspectives from which they were developed, opera- tionalized, and investigated. Expectancy of control has been studied in the context of motivation theory, autonomy as a task characteristic, and perceived influence in re- gards to decision-making strategies. Further, few research- ers have studied more than one of these dimensions of con- trol at the same time. Thus, our knowledge of how the dif- ferent aspects of control are interrelated, affect each other or complement each other is limited. The purpose of the present study is to increase our understanding of personal control in organizations. To- ward this end, the antecedents and consequences of each of the three dimensions of control will be identified through a review of the research literature. This information will be used along with the eXpanded version of Bazerman's multi— dimensional definition of personal control to develop a media- tional nodel ofpersonal control inorganizations . Each of the 14 dimensions of control will be operationalized using exist- ing or newly deve10ped instruments. These instruments will be used to empirically test the model of personal control and to determine the relationships among the three dimen- sions of control. It is hypothesized that the three dimensions of control are much more similar than one would expect, given the dif- ferent theoretical orientations and practical applications associated with each. The extent to which these three con- trol variables are interrelated will be determined. The research literature involving each dimension of personal control will now be reviewed to determine: 1. how each dimension of control has been operationalized 2. the important antecedents of each dimension of per- sonal control 3. the important personal and organizational outcomes of each dimension of personal control. CHAPTER II LITERATURE REVIEW Outcome Control Valence-Instrumentality-Expectancy (VIE) theory is a process theory of motivation because it describes a framework for understanding individuals' desire to exert effort in a particular situation. VIE theory hypothesizes that indivi- duals decide whether to exert effort on the basis of.ex- pectancies of future outcomes. The basic premise of VIE theory is that an individual will tend to exert effort in a particular situation if the effort is expected to result in a level of performance that will be rewarded with a suf- ficiently attractive outcome. The general model (see Fig- ure 1) is composed of a number of components: 1) Effort-to— Performance Expectancy (E —9 P) is an individual's subjec- tively determined judgment of the probability that his or her effort will result in a certain level of performance, 2) Performance-to-Outcome Expectancy (P —9 O) is an indi- vidual's subjectively determined judgment of the probability that a certain level of performance will lead to a particu- lar outcome, and 3) Valence (V) isthe amount of positive or negative value an individual places on the outcome. Thus, a person's motivation is a function of effort-to-performance expectancies, performance-to-outcome expectancies, and the valence of the outcomes (Nadler and Lawler, 1977). These elements combine in a multiplicative fashion where motiva- tion (M) = (E ——>p) x ZEP ——>0) (v8. 15 16 Amocwam>v mmaoopdo.Ai Ahhma .umHqu can HoHUmzv Hmooz mH> buommm A monmEuomumm A, muaaabm .H musmflm coflum>fluoz .17 The ideal motivational state, according to VIE theory, is one in which an individual believes him or herself cap- able of performing at a level that will bring about a highly valent outcome. In other words, the individual be- lieves he or she can control the occurrence of an important outcome through his or her own behavior. Thus, the ex- pectancy that one's effort will bring about an important outcome is an essential dimension of one's perception of control. Antecedents of Outcome Control Lawler proposed a model of the determinants of E —9 P and P —+ O expectancies. According to this model, the ac- tual situation is the most important determinant of one's E —+ P expectancy. Communication from others, past exper- iences in similar situations, and one's self-esteem would also affect one's E —a P expectancy (see Figure 2). P —+ O expectancies are also influenced by the actual situation, communication from others, and past experiences in similar situations. Lawler (1973) proposed that a number of other variables were important, including the valence of the out- comes and the individual's locus of control (see Figure 3). Locus of control is conceptualized as a personality dimen— sion that involves the generalized expectancy of whether one believes his or her actions will be rewarded. Rotter (1966) defined locus of control as: When a reinforcement is perceived by the subject as—-not being entirely contingent-upon' 18 Self-esteem Past experieht;57ih777777“‘-\\1\‘11\\\--~7>' similar situations } E —->P Actual situation Communication from others Figure 2. Determinants of E-4'P GXDGCtanClGS (Lawler, l973) Past experiences in similar situations Attractiveness of outcomes..Ip‘\\‘\\\M\\\\W>‘> \_ P —$>0 *77 Locus of Control. I—:--7\P///> Actual situation Communication from others Figure 3. Determinants of P-—9»0 expectancies (Lawler, l973) 19 his action, then, in our culture, it is typi- cally perceived as the result of luck, chance, fate, as under the control of powerful others, or as unpredictable because of the great com- plexity of the forces surrounding him. When the event is interpreted in this way by an in- dividual, we have labeled this a belief in ex- ternal control. If the person perceived that the event is contingent upon his own behavior or his own relatively permanent characteris- tics, we have termed this a belief in internal control. (p. l) The locus of control measure is sometimes used to identify two contrasting personality types--low scores on this measure indicate that an individual has a high inter- nal locus of control, while a high score on this measure identifies those individuals who exhibit a high external locus of control. Internal locus of control individuals (i.e. internals) tend to believe that they can control events and that they are personally responsible for the outcomes that they receive. External locus of control in- dividuals (i.e. externals), in contrast, tend to believe that they cannot control events and attribute outcomes to luck, fate, or other forces external to them. An individual's locus of control might influence his or her perception of outcome control in two ways. First, internals may perceive greater outcome control than exter- nals exposed to the same situation since they see them- selves as having greater control in general. Second, in- ternals may tend to seek jobs or remain in jobs that pro- vide Opportunities for outcome control. Research results have consistently found a negative relationship between locus of control and expectancy (Spector, 1982). 20 Szilagyi and Sims (1975) investigated the relationship between locus of control and both E —9vP and P .9.0 expec- tancies. The results indicated a consistent negative rela- tionship between locus of control and both types of expec- tancies with correlations ranging from -.02 to -.25 for E —>vP expectancies and -.2 to -.39 for P —9rO expectancies for five levels of hospital employees. Lied and Pritchard (1976) reported similar findings with the exception that E'—> P expectancies had a higher correlation (i.e. -.40) with locus of control than did P <>'O expectancies (i.e. -.20). Kimmons and Greenhaus (1976) found that internals had higher mean P -910 expectancy scores than did externals. Thus, locus of control appears to be an important antece- dent of outcome control. An additional personality variable related to one's perception of P —9 O expectancy should also be considered. A considerable body of research has found that individuals tend to have an exaggerated perception of the amount of control that they have over environmental outcomes. Hens- lin (1967) has observed that dice players clearly act as if they can control the outcome of the dice. They throw the dice differently when trying for certain numbers and believe that effort and concentration will pay off. Ward and Jenkins (1967) have demonstrated that people perceive causal relationships in the absence of objective contin- gency. This phenomenon has come to be called the "illusion of control" and is defined as: 21 . . .an expectancy of personal success probability inappropriately higher than the objective prob- ability would warrant. (Langer, 1975, p. 311) Not everyone appears to exhibit an illusion of con- trol, however. Researchers have found that depressed in- dividuals do not appear to have an illusion of control (Alloy and Abramson, 1979; Colin, Terrell and Johnson, 1977). Lewinsohn, Mischel, Chaplin and Barton (1980) found that depressed individuals rated their performance in an unstructured group discussion in a manner similar to that of independent observers. Nondepressed individuals, in contrast, rated themselves significantly more positively than the observers judged them to be. Lewinsohn, et a1. (1980) concluded: Nondepressed people may thus be characterized with a halo or glow that involves an illusory self-enhancement in which one sees oneself more positively than others see one. (p. 210) Alloy and Abramson (1979) found that nondepressed students overestimated how much control they had over objectively uncontrollable outcomes associated with a rate of success (e.g. winning money) and underestimated the amount of con- trol they had over objectively controllable events that were associated with failure (e.g. losing money). De- pressed students, on the other hand, accurately estimated the amount of control they had in each of these conditions. This phenomena is not limited to chronically depressed individuals. Rather, it appears that one's current mood state has a major impact on one's perception of control. In a laboratory study, researchers studied the effects of 22 induced mood states (i.e. elated or depressed) on partici- pants' perceptions of control over uncontrollable events (Alloy, Abramson and Viscusi, 1981). The results indicated that naturally depressed students who were temporarily made elated in the laboratory did exhibit an illusion of control when judging the amount of control that they had over un- controllable events. In contrast, naturally nondepressed students who were temporarily made depressed showed no illu- sion of control and accurately judged their personal con- trol over the event. Thus, the extent to which an indivi- dual feels depressed will lower his or her perceptions of outcome control. In sum, one's mood state, as well as the variables identified by Lawler (1973), may influence an in- dividual's perception of outcome control. Consequences of Outcome Control Organizational behaviorists have conducted consider— able research concerning the consequences of individuals' expectations of control rather than the effects of control on the individuals. Nadler and Lawler (1977) have developed a model relating VIE components to effort, performance, and satisfaction (see Figure 4). The model proposes that effort is the prime consequence of the VIE components. Perfor- mance, however, is the result of the combined forces of ability, effort, and problem solving strategy employed. Satisfaction is hypothesized to result from the intrinsic and extrinsic rewards received for one's performance. Thus, according to this model, one's expectancies of control will 23 Ability Extrinsic Outcomes v f \ Motivation -—-9’Effort .Performance Satisfaction (E—iP)XEEP—90) (V) A; 9 " E k. f ‘I‘ Intrinsic Outcomes 1 Observed and Problem Solving Actual Experience ——> Approach in Similar Situations Figure 4. VIE Model (Nadler and Lawler. 1977) 24 be most highly related to one's effort and less strongly related to either performance or satisfaction. Campbell and Pritchard (1976), in a review of the re- search evidence involving VIE theory, concluded that: While a multiplicative combination of expectancy, instrumentality and outcome valence typically yields a higher correlation than that for the individual components or simpler combination of components, the differences are usually not very great. Expectancy or instrumentality usually accounts for most of the variance that is to be accounted for and multiplying by valence seldom makes much difference. (p. 237) Thus, only research involving E-—9 P or P —> O expectancies will be reviewed. This is also appropriate since E'—9 P and P —9 O expectancies are the VIE components directly re- lated to outcome control. Shuster, Clark and Rogers (1971) found that subjects who were higher in E -9 P expectancies had higher perfor- mance than did subjects with low E —> P expectancies. In an experimental study, Arvey (1972) manipulated E —9 P ex- pectancies and found that subjects in the low E —> P expec- tancy condition performed poorer than did subjects in the high E<—> P expectancy condition. A number of researchers have also demonstrated the importance of PI—é O expectancies. Georgopoulous, Mahoney and Jones (1957) surveyed production employees that were on a work incentive program. The results indicated that em— ployees who perceived a high relationship between perfor- mance and work outcomes had higher levels of productivity. Porter and Lawler (1968) found a positive relationship 25 between P —>»O expectancies and ratings of performance and an even higher correlation between P -> O expectancies and a measure of effort. In a study involving a simulated work organization, Jorgenson, Dunnette and Pritchard (l973) man- ipulated P —> O expectancies by paying participants either on an hourly basis or piece rate basis. Results indicated that participants paid on a piece rate basis (high P —9vO) performed higher than did participants paid on an hourly basis (low P —> 0). Further, participants who shifted from hourly pay to piece rate pay showed an immediate increase in performance. This increased level of performance was maintained for the remainder of this study. In sum, expectancies of control can result in in- creased effort on the job and, to a less extent, improved performance and satisfaction. 26 Activity Control (Autonomy) Activity control has been most frequently defined in the OB literature as one's control over his or her work activities (i.e. autonomy). Most of the research concern- ing autonomy has been in relation to job or task charac- teristics. Herzberg (1966) proposed that jobs could be made more meaningful and satisfying through vertical job loading (i.e. job enrichment). One of Herzberg's princi- ples of enrichment concerned job freedom, which involved providing additional authority to an employee over his or her work activity. Job freedom was hypothesized to involve the "motivators" of responsibility, achievement, and re- cognition (Herzberg, 1966). Turner and Lawrence (1965) proposed that the levels of specific task characteristics present in a job were related to the attitudes and behav- iors of the workers performing that job. The task char— acteristics identified were: autonomy, variety, required job interaction, optional job interaction, knowledge and skills and responsibility. Autonomy was theorized to re- late to an individual's perception of personal responsi- bility for the successes and failures that occur as the re- sult of his or her work. Turner and Lawrence (1965) found that a summary measure of the six task characteristics scores (i.e. Requisite Task Attribute Index) was related to :fiob satisfaction and attendance, although this relation- ship held only for workers from small towns. These results have Spurred a considerable amount of research involving 27 job characteristics, as well as a search for variables that moderate the relationship between task characteristics and worker reactions. Hackman and Oldham (1976) have provided the most com- plete model specifying the relationship between job char- acteristics, individual differences, and work attitudes and behaviors. The Job Characteristics (JC) model pro- poses that the core job dimensions influence the critical psychological states, which in turn cause the personal and work outcomes (see Figure 5). The employee's Growth Need Strength (GNS) moderates the relationships both at the link between job core dimensions and psychological states and between the three psychological states and the personal and work outcomes. While Hackman and Oldham (1976) include five job core dimensions, only one is relevant to personal control—~autonomy. Autonomy is the degree to which the job provides sub- stantial freedom, independence, and discretion in schedul- ing and carrying out the work procedures. The JC model hypothesizes that autonomy leads to the psychological state of experienced responsibility for outcomes of the work. Experienced reSponsibility, in turn, along with the other two psychological states, causes the personal and work out- comes. These outcomes include high satisfaction, work quality, and motivation, as well as low absenteeism and turnover. Hackman and Oldham (1976) conceptualized the core job _28 laka_ .Ew;s_o sew agglomzv Pane: mo_sm_tmrumtm;u now ask .m mtsa_l gsacmtum smmz cuzocw mmXoFQEm Lm>occzp ucmg Liix»w>wpu< x103 on“ mo mupzmmm Emwwmucmmn< so; Fmauu< mcp mo mmumpzocx Am it» Luannmmu x103 web spa; cowuumdepmm gov: A” An x103 any to mmeoupzo com A xBWFwnwmcoammm smocmwtmaxm Am xsocouz< mocmsLOCLma x102 >Lt_m=o amt: mucmuvmvcmam meF x103 to covum>wpoz x103 mmmc_p=mmcwcmwz .AV 1 xuwucmuH xmmp _m:cmucH saw: I vmocmwcwaxm . lL_ r1: sumctm> __w¥m mmsoopzo mmpmpm meowmcmswo xcox ucm Fmowmo_o;oxma now mcou chomtwa Pmowuwtu 29 dimensions as objective characteristics of a job, yet job core dimensions have been most frequently operationalized as workers' perceptions of job core dimensions. The most frequently used instruments to measure job characteristics-- the Job Diagnostic Survey (JDS) (Hackman and Oldham, 1975) and the Job Characteristics Inventory (JCI) (Sims, Szilagyi and Keller, 1977)--assess individuals' perceptions of job characteristics. Thus, the operationalization of autonomy in the OB literature coincides with the conceptualization of activity control defined above as one dimension of per- sonal control. Antecedents of Activity Control The JC model begins with the individual's perception (of task characteristics and, thus, fails to specify the determinants of those perceptions. The assumed antecedents of one's task perception, however, are the actual charac- teristics of the tasks that comprfixaone's job. Hackman and Oldham (1980) have suggested some of the actual task char- acteristics relevant to each job core dimension in their job enrichment Implementing Principles. The Implementing Principles describe the actual steps one would perform to enrich a particular job. Two Implementing Principles are hypothesized to increase the amount of autonomy in a job-- establishing client relations and vertically loading the job. Establishing client relations involves arranging the job in order that the employee has the responsibility to decide how to handle the requests and concerns of the client. Vertically loading a job involves providing the 30 individual with responsibility and authority usually ree served for higher levels of management. Some suggested methods for vertically loading a job include providing the individual with discretion in setting schedules, de- termining work methods, and deciding when and how to check the quality of the work produced, as well as allowing the individual to make decisions concerning work hours, breaks, and work priorities (Hackman and Oldham, 1980). Thus, a job that provides the worker a substantial amount of dis- cretion in work decisions and procedures and responsibility for the outcomes of work will be high in autonomy and should be perceived as such. Evidence exists, however, indicating that different individuals may View the same job differently in regards to level of job characteristics (Schwab and Cummings, 1976). The JC model, unfortunately, does not specify the process by which an individual forms his or her perception of a task. A number of antecedents of task perceptions have been identified by researchers in the past few years. These antecedents include the actual task, individual dif- ferences, social cues, and organizational or situational variables. An individual's perception of autonomy should be most directly influenced by the actual amount of activity con- trol he or she has been able to exercise on the job in the past. Thus, an individual who has been allowed to decide how and when the work will be performed and has been held 31 accountable for the outcomes of work should perceive his or her job as high in activity control. The amount of ac- tivity control one has been able to exercise on the job is a function of job characteristics, type of supervision, and individual differences. The technological requirements of some jobs would prevent the worker from exercising discre- tion in work procedures or work pace. For example, machine- paced jobs provide little discretion to workers with regard to work procedure or work pace. The type of supervision may also limit the extent to which an individual is able to exercise autonomy on the job. Frequent directions and continual checking by a supervisor will result in a job with low autonomy and little employee responsibility for the work outcomes. Certain individual differences variables may also in- fluence the amount of autonomy an individual has been able to experience on the job. Three variables seem likely to influence the amount of activity control one is allowed on the job--tenure, ability, and desire for autonomy. The long- er a competent worker remains on a job, the greater are the chances that he or she will be allowed more discretion and control over the work methods. Also, it is possible that supervisors will provide greater autonomy to those workers with the greatest ability and job skills. This might come about by supervisors rewarding subordinates' good perfor- mance with increased autonomy. Finally, it is likely that an individual with a high need for autonomy will seek 0p- 32 portunities to fulfill these needs. Individuals with a low need for autonomy, on the other hand, may avoid or turn down opportunities for increased autonomy. A number of researchers have recognized the fact that an individual's perception of activity control is influ- enced by factors other than the objectively defined task. Szilagyi and Holland (1981) studied the effects of changes in social density on employees' attitudes and perceptions. A change in social density was defined as an increase or decrease of greater than two employees per 50-foot walking distance. The results indicated that an increase in social density caused a decrease in employee perceptions of au- tonomy, while a decrease in social density resulted in an increase in perceived autonomy (Szilagyi and Holland, 1981). Oldham and Hackman (1981) investigated the relation- ship between organizational structure variables and worker perceptions of task characteristics. The results indicat- ed that workers' perception of autonomy was negatively re- lated to measures of centralization and formalization of the organization. In a study of the relationship between organization structure, job characteristics, and worker satisfaction and performance variables, Brass (1981) found significant positive relationships between criticality of the task, subunit centrality,and perceptions of autonomy. Thus, the task itself, physical work conditions, and or- ganizational variables have been found to be related to an individual's perception of autonomy. It is obvious that 33 these different variables provide information to the indi- vidual which influences his or her formulation of task per- ceptions. Another source of information that is believed to influence task perception involves social cues. Salancik and Pfeffer (1978) proposed that individuals are likely to use social information when developing their perceptions of job characteristics. O'Reilly and Caldwell (1979) found that informational cues indicating whether a task was enriched or not enriched had a stronger influence on job perception than did the actual objective task char- acteristics. O'Connor and Barrett (1980) found that an in- dividual difference variable (i.e. field dependence/inde- pendence) influenced how the participants formed their per- ceptions of the job. Field dependent participants seemed to focus more directly on socially induced informational cues, while field independent participants were more strongly influenced by the physical aspects of the job in formulating their task perceptions (O'Connor and Barrett, 1980). The results are in li'ne‘with Jamesetal.'s (1979) find- ings that subordinates differentially attended to environ- mental cues (i.e. supervisor's behavior) depending on their own needs in their formulation of perceptions of influence. Another individual difference variable that may influ- ence one's perception of autonomy is locus of control. Spector (1982) proposed that internals should perceive their job as offering greater autonomy due to their gener- alized expectancy of control over their environment. Ex- 34 ternals, on the other hand, should perceive their job as offering less autonomy. Kimmons and Greenhaus (1976) found that internals reported having more autonomy than did externals in a sample of 191 managers. A recent study investigating the causality of the task perception-job satisfaction relationship is also rele- vant here. Caldwell and O'Reilly (1982) had participants imagine that they held the job described in a detailed job description and that they were either satisfied or dissat- isfied with that job. Those participants who had imagined that they were satisfied with the job reported higher levels of job characteristics, including autonomy, than did the dissatisfied participants, thus providing some sup- port for Zajonc's (1980) contention that affective judg- ments may precede the cognitive perceptions (e.g. task per- ceptions) in time. Individuals' current mood states have also been found to influence the accuracy of estimates of positive feedback received on a task (Buchward, 1977) and judgments regarding the degree of control exerted over events (Alloy, Abramson and Viscusi, 1981). Thus, an in- dividual's affective evaluation of the job or even his or her current mood state might influence his or her percep- tions of job characteristics. In sum, the manner in which an individual formulates his or her perception of autonomy is not well understood. What is certain is that a number of factors (see Figure 6) in one's work (and probably non-work) environment provide 35 atacm» new xeocous< to» uwmz .>u_ppn< m.szuw>_v:H co_uumwm_umm now pocucoo auw>vuum mo mucmumomucm och .m mtzmwd woo: now on» :o xsocouz< \. mowumwtmuomtmsu umucm_cmaxm Adi xmmp Fog“ ou xpm>wpu<\»20:0p:< mono \\ co_mw>cwa:m um>wmucma Aw _a_oom Ami we waxy mcmxcozuou xu_mcmo x/ 38m Aco_um~w—msgom .cowuun__mcucmuv 838.13 355323.10 \4 do mcowuawotma .mFaaww>_ucH Soc—oczomp use mwauwpoa new mczuuacum Pm:o_um~wcmmco 36 information concerning activity control to the individual. Further, research indicates that individuals selectively attend to the presence of specific types (e.g. opportunit- ies for influences) of information based on their own needs and beliefs and to different sources of information (e.g. social versus task) depending on their personality type (e.g. field dependent/independent). Thus, an individual perceives and weighs the existing autonomy cues depending on his or her personal characteristics in forming percep- tions of autonomy. Consequences of Activity Control The JC model proposes that autonomy leads to the psy- chological state of experienced responsibility for outcomes of the work. This psychological state, together with the other two critical psychological states, is hypothesized to lead to a number of positive work outcomes. These outcomes include high internal work motivation, high "growth" satis- faction, high general job satisfaction, and high work ef- fectiveness. High work effectiveness involves quality and quantity of work, as well as attendance at work (Hackman and Oldham, 1980). The JC model does not hypothesize the relationship of autonomy to the employees' personal and work outcomes inde- pendent of the other job core dimensions. Thus, it is speculative to propose that autonomy is more highly related to one outcome over another. Further, researchers have tended to combine the five job core dimension scales into 37 an overall measure of job complexity, thus further cloud- ing the issue of the independent contributions of perceived autonomy. The theoretical importance of autonomy in relation to the other job core dimensions, however, is indicated in the formula for the Motivating Potential Score (MPS) (Hack- man and Oldham, 1976). The MP8 specifies how the job core dimensions scores should be combined to produce an overall rating of the motivating potential of a particular job. The MP8 is defined as: Skill Task Task Variety Identity Significance 3 MP8 = x Autonomy x Feedback As can be seen by this formula, a very low score on auton- omy would result in an MP5 close to zero. Thus, autonomy is an essential component of an intrinsically motivating job. Research findings, however, have indicated that equal weighting of the individual job core dimension scores has been at least as effective as the MPS formula in explain- ing response variance (Dunham, 1976; Hackman and Oldham, 1976). These findings question the JC model'stheoretical proposition of the noncompensatory nature of the key job core dimensions, as well as the unique importance of au- tonomy. The determination of the consequences of activity con- trol is further hampered by the method with which research- ers report job design research results. Job design re- searchers tend to report only the job complexity score '38 (i.e. combined sum of job core dimension scores) or the MPS score, thus precluding the determination of the indi- vidual contribution of autonomy to the dependent variables studied. This procedure is usually a response to the low reliability (i.e. internal consistency) and high intercor- relations of the job characteristics'subscales. A number of researchers, however, have reported data relevant to the effects of autonomy independent of the other job core dimensions. Only these studies will be reviewed. Hackman and Oldham (1975, 1976) found a significant positive relationship between autonomy and a measure of overall satisfaction. In another study using the Job Di- agnostic Survey (JDS) to measure autonomy, Rousseau (1977) also found a significant positive relationship between au- tonomy and overall job satisfaction. Brief and Aldag (1978) used the Job Characteristics Inventory (JCI) to measure autonomy and replicated these earlier findings. Autonomy has also been found to be related to other types of satisfaction--growth satisfaction (Hackman and Oldham, 1975, 1976), satisfaction with work (Brief, Aldag and Ja- cox, 1978; Keller, Szilagyi and Holland, 1978; Sims and Szilagyi, 1976), satisfaction with pay, promotion, super- vision and co-workers (Sims and Szilagyi, 1976). Thus, considerable research evidence indicates a strong relation- ship between autonomy and satisfaction. The validity of these data, however, is questionable due to the methodolo- gical flaws in some of the studies. Since autonomy and I_J. '39 satisfaction were both measured via the same questionnaire in a large portion of these studies, one must seriously consider the possibility of common method variance (Roberts and Glick, 1981). Autonomy has been demonstrated to be related to a number of other outcome variables. Hackman and Oldham (1975, 1976), in a test of the JC model, have found a sig- nificant positive relationship between perceived autonomy and internal work motivation. Rousseau (1977) found that autonomy was positively related to job involvement and neg- atively related to a measure of alienation. Moch (1980) also found a significant relationship between autonomy and job involvement. Keller, Szilagyi and Holland (1976) found a significant negative relationship between autonomy and measures of role strain (i.e. role ambiguity and role con- flict). Finally, Hackman and Oldham (1976) found autonomy positively related to work effectiveness and negatively re- lated to absenteeism. 40 Perceived Influence Perceived influence concerns the degree to which an individual perceives him or herself as having an influence on the decisions made by their supervisors (James etal.,1979; Vroom, 1960). Research involving psychological influence can be classified into three distinct categories on the basis of how influence was Operationalized in the study. The first classification consists of research that studied the effects of different amounts of objectively defined participation in decision making. These studies tended to be of an experimental nature, comparing conditions of par- ticipation versus nonparticipation (e.g. Coch and French, 1948). The second group of studies involved one's percep- tion of influence and its relationship to satisfaction and performance variables (e.g. Vroom, 1960). The third class- ification of studies involved research that used a "dis- crepancy" measure of influence obtained by subtracting one's ratings of perception of actual influence from his or her ratings of the desired amount of influence in decison making (e.g. Alutto and Belasco, 1972). Major research studies in each of these areas will be reviewed. Objectively Defined Influence One of the earliest studies involving objectively de- fined levels of influence was Lewin, Lippit and White's 1938 comparison of "autocratic" and "democratic" children's activities groups. In the autocratic group the adult de- termined the activities of the children and gave frequent .41 directions. In contrast, the democratic group used group decision making with the adult acting in only an advisory role. A suprising result was that the autocratic group showed a marked decrease in productivity when the adult was not present, while the democratic group showed no change. Further, some children in the autocratic group ex- hibited signs of apathy, dependence, frustration, and ag- gression toward their leader. These symptoms are remark- ably similar to those later labeled as "resistance" in studies involving employees' reactions to organizational change (Coch and French, 1948). Two organizational studies provided further evidence concerning the positive effects of allowing employees to participate in decisions that influence their workplace. Coch and French (1948) studied the effects of different levels of employee participation in production changes in a manufacturing organization. The full and partial partici- pation groups showed faster production recovery rates, higher production rates, lower turnover, and less aggres- sion toward supervisors than the nonparticipation group. Morse and Reimer (1956) introduced planned changes in the level of involvement that groups of clerical workers had in decisions made by their supervisors. In the Autonomy Pro- gram the role of the workers in decision-making operations was increased, while the hierarchical program increased management's role in decision-making operations. The re- sults indicated that the hierarchical groups had a greater 42 increase in productivity than the autonomy groups; however, the autonomy groups had a much higher level of work satis- faction than did the hierarchical groups. A recent organizational study, however, has raised questions regarding these earlier findings. Lischerhorn and Wall (1975) conducted an experimental field study com- paring participation groups (i.e. Action Planning Groups) to nonparticipation control groups. Action Planning Groups (APGs) consisted of 6 to 14 men, their supervisor, and one management representative. Each group met infor- mally every three weeks to discuss grievances, make sug- gestions or to ask questions. Management guaranteed to provide answers to all issues raised by the meetings which followed. A comparison of worker attitudes in the parti- cipation and nonparticipation groups was conducted. The results indicated that the increased opportunity to parti- cipate did not result in a significantincrease in worker satisfaction with the organization, pay, opportunities for promotion, the job itself, immediate supervisors or co- workers. Worker attitudes toward middle management, how- ever, did improve through involvement in APGs. In sum, attempts to increase the actual amount of em- ployee involvement in decision making have not always re- sulted in greater work satisfaction or increased job per- formance. The equivocality of these results may be due, in part, to an insufficient'increase in the participants' per- ception of influence resulting from the experimental manip- 43 ulation. Thus, it is important to review the research findings concerning perceived influence. Perceived Influence Vroom (1960) was the first to study the correlates of individuals' perception of influence (i.e. psychological participation). In a study with 108 supervisors, Vroom found a significant positive relationship between the su- pervisors' perception of psychological participation and a measure of job satisfaction and work performance. Further, Vroom discovered that the magnitude of these relationships WEiS moderated by two personality variables--authoritarian- ism and need for independence. In a replication of this study, Tosi (1970) found a significant positive relation- ship between psychological participation and job satisfac- tion but not with the performance measure. In addition, Tosi was unable to replicate the moderator effects of au- thoritarianism and need for independence. James, etal.,(l979) suggested that psychological participation be renamed psy- chological influence since the construct involves the per- ception of influence in participative decision making. James et al., (1979) examined the relationship between individuals' ratings of psychological influence and three classes of variables--situational, subordinate-person, and subordinate- psychological climate. The results indicated that percep- tions of psychological influence were related to person variables (e.g. anxiety and rigidity) and situational vari- 44 ables (e.g. supervisor behaviors). Discrepancy Measures of Perceived Influence Alutto and Belasco (1972) argue that it is reasonable to assume that not everyone desires an increased involve- ment in company decision-making operations. Thus, the cru- cial variable for determining the effects of participation is the discrepancy between actual and desired opportunities for participating rather than one's perception of or actual involvement in decision making (Alutto and Belasco, 1972). Following this line of thought, Alutto and Belasco (1972) developed a continuum of participation consisting of three conditions: (a) decisional deprivation-actual participa- tion in fewer decisions than one desires, (b) decisional equilibrium-actual participation in as many decisions as one desires, and (c) decisional saturation-actual partici- pation in a greater number of decisions than one desires. The basic assumption of this line of research is that con- gruence between one's desired participation and actual par- ticipation is the desired state. Decisional deprivation or saturation, on the other hand, should lead to more negative job attitudes and higher levels of job tension (Alutto and Acito, 1974). Driscoll (1978), in a study of college faculty, found that the greater the participants' congruence between de- sired and actual participation, the greater was their sat- isfaction with the organization and with the participation 45 itself. Alutto and Vredenburgh (1977) found that decision- ally deprived nurses had higher job tension and greater career dissatisfaction than those nurses with decisional equilibrium. In a study of project engineers, Ivancevich (1979) found a significant relationship between decisional deprivation and measures of physical stress and job ten- sion. Thus, decisional deprivation does appear to be re- lated to negative work attitudes and increased tension on the job. It is important to note, however, that the use of dis- crepancy scores has been questioned by a number of re- searchers (Cronbach and Furby, 1970; Johns, 1981), thus limiting the interpretability of these results. Antecedents of Perceived Influence The most direct antecedent of an individual's percep- tion of influence is his or her actual involvement in par- ticipative decison-making (PDM) activities. PDM, however, is not a unitary well-defined construct. Rather, re- searchers have defined and operationalized PDM in a variety of different ways (Dachler and Wilpert, 1978). Locke and Schweiger (1979), however, have identified a number of di- mensions along which PDM may vary--degree, content, scope, and type, thus providing a basis with which to organize a discussion of the antecedents of influence. Participation can vary in the degree to which an in- dividual is allowed or encouraged to become involved in 46 the decision-making process of his or her supervisor. Tannenbaum and Schmidt (1973) have developed a Leader Be- havior Continuum (LBC) (see Figure 7) which lists the en- tire range of possible leader-subordinate influence rela- tions. The continuum ranges from a boss-centered or au- thoritarian leadership approach in which the manager makes the decsion and then announces it to the subordinates, to a subordinate-centered or participative approach, whereby the subordinates function autonomously within limits de- fined by the manager. Tannenbaum and Schmidt describe a full range of management behavior that falls between these two extremes. Heller, Drenth, KOOpman and Rus (1977) have developed a similar continuum. Their six position Influ- ence and Power Continuum (IPC) describes the varying levels of subordinate influence in decision-making activities: The IPC consists of six levels: 1. No Information No detailed information about the decision is made available. Subordinate-cenle'r‘ed A' I ders i Bmscentercd ‘ H 3‘ leadership ———) Use 0! authority by the manager Area of freedom {or subordinates Manager Majager Manager Manager Manager Manager Manager makes "sells” presents presents presents defines permits decision decision. ideas and tentative problem. limits; subordinates and invites decision gets asks gLoup to. lxnciron u uestions. subiect to suggestions, to ma e wt in It!“ announces q change. makes decision. defined by . decision. superior. Figure 7. Leader Behavior Continuum (Tannenbaum and Schmidt, 1973) (.JI 47 2. Information Fairly detailed information about the decision is made available. 3. Opportunity to Give Advice Before the decision is made, the supervisor explains the problem and asks advice. The supervisor then makes the decision by him or herself. 4. Advice is Taken into Consideration As above, but your superior's final choice usually reflects the advice he or she has received. 5. Joint Decision Making Your superior and his or her subordinate(s) together analyze the problem and come to a decision. Your supervisor usually has as much influence over the final choice as his or her subordinate(s). In fact, one could say everybody in principle has equal influence (one person, one vote). 6. Complete Control You or members of your work group are given the au- thority to deal with this decision on your own. Su- periors would intervene only in exceptional circum- stances. Naturally, every now and then you or the group are expected to account for the action taken (Heller, Drenth, Koopman and Rus, 1977, p. 572). The IPC appears to be a more complete scale of par- ticipation than the LBC since it contains two levels of three hierarchically-arranged factors--information, advice, 48 and decision making, whereas the LBC contains levels with- in only one dimension (i.e. decision making). Little research has empirically assessed the relation- ship between degree of participation and perceived influ- ence. One would assume a positive linear relationship be- tween the amount of involvement one has in PDM and his or her perception of influence, although it is possible that certain personal variables might moderate this relation- ship. For example, an individual with an external locus of control:fimmld.generally perceive less influence than an individual with an internal locus of control exposed to the same level of PDM. A second dimension on which participation may vary is the content of the issue involved in the decision. Locke and Schweiger (1977) have identified four categories of de- cision types: 1. Work Itself - job design, work methods, job pace and production level. 2. Work Conditions - work hours, rest breaks, lighting, and other physical work arrangements. 3. Routine Personnel Functions - selection, compensation, training and performance appraisal issues. 4. Company Policies - including layoffs, fringe benefits, general wage levels, dividend and general policy making. It is important to note that the first two decision cate- gories--the work itself and work conditions--consist pri- 49 marily of decisions involving how one goes about performing his or her job. Thus, these two categories of decisions are related to activity control rather than perceived influence. This distinction is a necessary one in the attempt to maintain conceptual distinction between the con- structs of activity control and perceived influence. Perceived influence has been operationalized by asking employees to rate the amount of influence they have in gen- eral without reference to Specific decisions (e.g. Vroom, 1960; Tosi, 1970). A number of researchers, however, have assessed individuals' perceptions of influence over speci- fic work decisions (e.g. Alutto and Belasco, 1972; Hrebin- iak, 1974). Ivancevich (1979) even conducted extensive in- terviews to determine a relevant and meaningful set of de- cision situations in his study with project engineers. However, no research has attempted to determine the rela- tionship between decision content and perceived influence. Those researchers who have specified the different de- cision situations employed in their study have, for the most part, collapsed the responses on the different deci- sion types into one overall score (usually a discrepancy score determined by subtracting the number of decision sit- uations in which the participant actually participates from the number of decision situations in which the participant wished he or she participated). Thus, information concern- ing the relationship of the actual content of decisions in which one is allowed to participate and the individual's 50 perception of influence has not been established. Such a distinction is made in the operationalizations of activity control and perceived influence in this study. Activity control involves one's perception of control of the acti- vities and decisions concerning the work itself and work conditions. Perceived influence concerns one's perception of involvement in decisions related to routine personnel functions and company policies. The third dimension on which participation may vary concerns the ggppg of the participation. Locke and Schweiger (1979) have defined scope as "the stage of prob- lem solving at which PDM occurs" (p. 276). Although little empirical research has addressed this issue, one study has explored the participation sc0pe--perceived influence re- lationship. Cooper and Wood (1974) compared partial par- ticipation conditions with a complete participation condi- tion in a laboratory study of group decision making. Par- ticipants were assigned to one of the following conditions: (a) generation of alternative solutions, (b) evaluation of alternative solutions, (c) choice of alternative solutions, or (d) generation, evaluation, and choice of alternative solutions (i.e. complete participation). The results indi- cated that those participants in the choice and complete participation conditions had significantly higher percep- tions of participation and influence in the decision-making process, as well as greater levels of task perception than did those participants in either the generation or evalua- 51 tion conditions. Thus, it may be necessary to actually allow subordinates to make some decisions if one wishes to significantly increase their levels of perceived. influence. The final dimension along which participation may vary is Eypg of participation. Locke and Schweiger (1979) have identified three different types of PDM--forced or volun- tary, formal or informal, and direct or indirect. A forced PDM program is one that is mandated by law or union con- tract (e.g., codetermination), while a voluntary program would involve a management-initiated program to which em- ployees agree to join (e.g. Scanlon plan). A formal PDM program involves an officially recognized bargaining com- mittee (e.g. union), whereas an informal PDM relationship is usually based on a personal relationship between the su- pervisor and subordinate. Finally, the distinction be- tween direct and indirect PDM concerns whether one has di- rect involvement in decision making or has a representative participate in his or her behalf. The nature of each PDM type might be related to one's perception of influence. For example, a forced, formal, or direct PDM program may be more likely to be related to a higher level of perceived influence than voluntary, informal, or indirect PDM pro- grams, although no research has directly tested this hy- pothesis. In sum, actual involvement in decision-making activi- ties is the most likely and direct antecedent of perceived 52 influence. Unfortunately, little is known concerning the importance of the different dimensions of participation to one's perception of influence. It is likely, however, that the degree, content, scope, and type of participation will impact on an individual's perception of influence. Fur- ther, one's locus of control might also influence one's 1e- vel of perceived influence. Consequences of Perceived Influence The most frequently studied variable in relation to perceived influence is employee satisfaction. Vroom (1960) found a significant positive relationship between the par- ticipants' perception of influence and a measure of atti- tudes toward the job. In a replication of Vroom's study, Tosi (1970) found similar positive results. Cooper and Wood (1972) assigned participants to experimental condi- tions involving different phases of decision-making activi- ties--generation, evaluation, and choice. The results in- dicated that satisfaction was highest in those conditions (e.g. choice and complete participation) which produced the highest ratings on perceived influence. Wood (1972) has found a significant positive relationship between perceived participation in an experimental task and satisfaction with the decision, leader, method, relations, own role, influ- ence, accomplishments, and overall satisfaction. Morse and Reimer (1956) compared groups of clerical workers after their involvement in decision making had been either in- 53 creased or decreased. A manipulation check showed that perceptions of influence had changed in the expected di- rection for each group. Further, those participants whose involvement in decision making had been increased had sig- nificantly higher job satisfaction than the participants whose involvement was decreased. A number of researchers have found a consistent rela- tionship between perceived decision-making deprivation and worker dissatisfaction. Alutto and Acito (1974) found that decisionally deprived blue collar workers reported lower satisfaction with work, supervision, and promotion than workers who perceived their decision-making involvement as adequate. Ivancevich (1979) found that the greater the re- ported decisional deprivation, the lower was the satisfac- tion with supervision and with work. In a study involving college faculty, Driscoll (1978) found that the greater the congruence between desired and perceived participation in decision making, the greater was the satisfaction with both the organization and the participation itself. Thus, con- siderable research has shown that one's perception of in- fluence is related to satisfaction at work. A second important variable frequently investigated in relation to perceived influence involves employee involve- ment and commitment to his or her work. In a study of the Tennessee Valley Association (TVA), Patchen (1970) found that involvement in decision making was related to an in- crease in individuals' integration into the organization. 54 Siegel and Ruh (1973) found a high positive correlation be- tween employees' level of perceived control and a measure of job involvement. Alutto and Acito (1974) found that de- cisionally deprived employees were less committed to the job and company than were decisionally satisfied employees. Alutto and Belasco (1972), however, found no relationship between decisional deprivation and organizational commit- ment. Hrebiniak (1974) found that decisional deprivation had little impact on employees' organizational commitment. In a study of hOSpital nurses, Alutto and Vredenburgh (1977) found no relationship between decisional deprivation and either organizational or professional commitment. In sum, the relationship between perceived influence and com— mitment appears to be equivocal. A possible explanation for the mixed results is that researchers have failed to suf- ficiently discriminate among the various types of commit- ment or the specific areas of influence. An increased in- volvement in decisions that influence one's work should have a greater impact on one's job involvement than his or her organizational commitment. Following the same line of reasoning, an increase in participative decision making in- volving organizational policies might be more likely to in- crease one's organizational commitment; however, there is no reason to believe that it will significantly affect his or her job involvement. Thus, future research should en- sure a match in the domains of influence and commitment in- vestigated. 55 Another line of research concerns the relationship between perceived influence and role conflict and role am- biguity. Schuler (1977) found that the greater involvement in decision making, the lower were the levels of role con- flict and role ambiguity. In a study involving clerical, blue collar, and professional employees, Morris, Steers and Koch (1979) found that participation was the best pre- dictor of role conflict and the second best predictor of role ambiguity. In a longitudinal field experiment, Jack- son (1983) found that participation had a significant neg- ative effect on role conflict and ambiguity and a positive effect on perceived influence. Schuler (1980) has developed a model relating parti- cipation to role and expectancy perceptions (see Figure 8). According to this model, participation in decision making provides the individual with information concerning his or her work role, job requirements, and work outcomes, thus clarifying his or her work role and outcome expectancies. Schuler hypothesizes that a number of individual and organ- izational variables will influence one's perception of role conflict and role ambiguity and expectancy of outcomes. Finally, Schuler's model proposes a link between employees' perceptions and expectations and their satisfaction with both work and supervision. Schuler (1980) tested this model in a study involving three different organizational levels-~upper level management, middle level management, and clerical and blue collar workers of two different or- 56 comw>cmaam cum; cowuomtmwumm xtoz cow; corpummmwpmm \ An. Aommp .cmpagomv mcwxms cowmwomu cw comumamowurmq do mmmcmumWLQOLQQm map can cowuummmwpmm new mcombamULma socwpomaxm new mFoc new mm—nmwrm> chowpm~wcmmto new sznw>wncw do mawsmcowumch omemmcpoaxz .w mt:m_m sb1_wn< quFmCOmtwa mm_nmwcm> szuv>mvcH Fm>mL mem mmfinmwcm> Pm:0wumn+cmmto N aucmpomaxm ss_=a_ssa m_oe BUEFECOU w~om mcwxmz cowmwumo A" cw cowumawu_utma 57 ganizations. The results indicated that participation in decision making was negatively related to role conflict and ambiguity and positively related to the performance-reward expectancy, thus providing initial support for Schuler's model. A third line of research has investigated the rela- tionship of perceived influence to physical and psycholo- gical strain. Alutto and Vredenburgh (1977) found a posi- tive relationship between decisional deprivation and job tension. Ivancevich (1979) found that the greater the amount of decisional deprivation, the greater was the phy- sical stress and job tension. In a study involving two different samples of employees, Caplan, Cobb, French, Har- rison and Pinneau (1975) investigated the relationship be- tween perceived participation and several measures of stress. The results indicated that perceived participation had a significant negative relationship with depression in both samples and a significant negative relationship with anxiety and somatic complaints in only one of the samples. Margolis, Droes and Quinn (1974) studied the relation- ship of six potential stressors with ten measures of phys- ical and psychological strain. The six stressors used were nonparticipation in decisions affecting one's job, role am- biguity, underutilization of abilities, overload, resource inadequacy, and insecurity about future employment. The ten measures of strain included overall physical health, depressed mood, self-esteem, and job satisfaction. Non- 58 participation was significantly related to all ten of the strain measures. Further, nonparticipation had a higher correlation than any of the other stressors studied with eight of the ten strain measures. Zaleznik, Kets de Vries and Howard (1977) investigated the determinants of job stress reported by members of three different occupational groups--operations, staff, and man- agement--in a large Canadian service organization. The op- erations and staff groups both reported higher levels of stress than the management group. An analysis of the self- reports of job experience of the three groups of employees indicated that the operations and staff groups felt frus- trated by their lack of influence on decisions that affect- ed their work. The group of managers, in contrast, did not share these low perceptions of influence since they had considerable authority over the decision-making operations. Thus, the inability to influence decisions that affect one's worklife can have a considerable impact on one's physical and mental health. A final variable investigated in relation to perceived influence is attitude toward unions. The basic premise un- derlying this research is that an individual with little influence in organizational decision making would view unionization as a possible means of increasing his or her control over organizational decisions. Alutto and Belasco (1972) collected information concerning the attitudinal dispositions of school teachers toward unions, strikes, and 59 collective bargaining. They also measured the teachers' perceived level of decisional deprivation. The results indicated a strong positive relationship between decisional deprivation and attitudes toward unions, strikes and col- 1ective bargaining. In a study involving hospital employ- ees, Hrebiniak (1974) had respondents rate the amount of influence that they would like to have, as well as the - amount of influence they would like their unit leader to have. Respondents' perceptions of involvement in depart- mental decision making were also measured. The study found that decisionally-deprived individuals wanted greater in- fluence for themselves and less influence for their unit leaders. Thus, the desire for increased influence and positive attitudes toward unions has been found to be re- lated to decisional deprivation. In sum, considerable research evidence exists that in- dicates a strong positive relationship between perceived influence and job satisfaction, and measures of physical and psychological strain. Role conflict and ambiguity have also been consistently found to be related to perceived in- fluence. The relationship between perceived influence and job involvement, organizational commitment and attitudes toward unions appears to be more equivocal. 60 Summary and Research Plan The literature involving the three dimensions of per- sonal control--outcome control, activity control, and per- ceived influence--has been reviewed to determine how each construct has been operationalized by organizational be- havior researchers and to identify the important antecedents and consequences of each dimension of personal control. Table 1 summarizes the literature review presented in this paper and describes how each personal control dimen- sion has been operationalized. Outcome control has been most frequently operationalized using a self-report measure of individuals' perceptions of Effort to Performance or Performance to Outcome expectancies. Neither the Effort to Performance nor the Performance to Outcome expectancy scale, however, seems fully appropriate to measure outcome control. Outcome control, as defined in this paper, involves an in- dividual's perception of causing, controlling, or influenc- ing the outcomes that he or she receives on the job. Effort to Performance expectancies provide an indication of one's perception of control over only one aSpect of his or her work environment (i.e. performance). Performance to Out- come expectancies, on the other hand, are more indicative of the predictability with which one is rewarded for per- formance rather than one's perception of control over those outcomes. To overcome these limitations, outcome control was operationalized in this study using a self-report meas- ure of individuals' perceptions of Effort to Outcome eXpec- ‘6]. Hams .ccaaao: can asmmdanm whoa .msoncwmuo can mCOEEAz omma .uuouumm can u0::00.0 Ahma .muomom can xumao .uoum iszom “mmma .uma3oa can umuuom «bead .mmcon can xoconmz .msoHsoaomuomo Huma .mumm tom can xumHU .uoumsnom “mead .uoa3ma can nouuom momma .uouuom can :nExomm mama .meHm 6cm Hammaaum “mama .msoncomuo can mcoeewx “Shad .aumauoaua can coda omma .aou loam new anammno .Hwnomflz .anochBma ubhma .cOm tench new Hamuuoe .caaoo umhma .COmEmund can >0Ha¢ mumnoummmom Cu Ou Ou ca 00 Cu Ou xuamcmp HoHUOm poumHou >Ho>qummoc Houucoo mo msooH pmumHou >Ho>wuwmom oocopcommp paoww pouoamu >Ho>wuamom mocoEuOuuom pmumeu >Hw>auwmom upOmuo poumaou xam>wuwmom Houucoo mo msooH pmumamu >Hm>wuwmom coos commoummc pmuoamu >Hw>flumooc muasmmm concomwm mamom anc0usc Hun no man Houucou >ua>auo< .N mmflucmuuomxo 0 ml m HO O +|.m m Aaouucoo no scamsaaw .o.av Houucoo mo wocmuoodxm pmuoamcw Houucou mEoouso .H pouflaocoHumummo 30m coflmcoefio Houucou anaemumm .mcofimcoefio Houucou Anaconda on» can moanmwum> mEouuzo can ucmpououc< cooZuom awnmcoaumamm on» onwamw>om ousuououflq on» no >uneesm a wanna (52 mmmd .ouwnz can uwdqu .cfi3oq “mama .nocmum new 2000 mva .cocmum pcm £000 mama .aocmua can nooo «had .6003 can uoaoou umhma .uomooo whma whma .sommmsom “omma .mhma .Eonpao new coExomm .Emnpao was :mExomm whma .GCMHHOm can Hammafium .uoaaox hhma mhma .Emnpao pan :mExom: mama .asmmaanm paw mEHm “mhma .mhma .Eonpao cam :nExom: reams .omnsa 6cm madam MMQSUHMOMOK .nooz >ua>auospoum ob poundou >Ho>wufimoa uo>ocusu ou ooumamu >Ho>wummoc xnuomo no“ Ou pmumaou >Ho>fiuommc :oMuooumwuom Houucoo mcwonICOAmwoop mcwmmouocw noduoasmwcme ou poumawu >Ho>auflmom mmoco>fiuoouuo xuos ou pmuoamu >Ho>wuwmom Emfiomucmmnm ou pmumamu aao>wumuoc cannun maou Ou poumamu >Ho>aumomc ucmso>ao>cw n0n ou pouoaou >Ho>wuoomc :oHum>HuOE xuo3 Hocuoucw ou pmumamu >H0>wuamom coHuoowmauom Ou ooumamu >Hm>wuam0d muasmom nouowmwm Apmscflucouv H manna pwnaamcoflumummm 30m oucwaaucH ©m>flooumm .m Apmscflucoov Houucou >uw>fluo< .m concoan Houucoo anaconda 6Z3 mhma .oomoamm pan Ouusaa puma .cmusncopmu> can Ouuaam thH .OUWMHOQ UCfi OUUSH< vhma .Ouaod Ucm OuuSad mhma .noa> Imucm>H hmhma .aaOUmauo “whma .Ouaod can Ouusam coma .Eoou> vmma .ccaso can moouo .mmaaomuoz ummma .smoccam pan comauunm .nocwum .nnou .coammo enma .nsm 6cm ammmam rosma .cmnouaa omma .Eoou> uonma .amoe mumnoumwmmm mopsuauum scan: 0» poumaou >a0>auamoa camuum ado iamancoxmm can aooamxnm o» pounaou aao>auamoa ucoEuaEEou aocOaunuacnmuo Ou concaou mao>aummoc ucwso>ao>Ca non ou coumamu >ad>auomoc ceauonwmauom now 0» pouMaou >ao>aummoc oucnfiu0wuom QOn ou poumaou mam>auamom canuum anoamOa iosohmm can anoamxnm ou pmuMamu mam>aummoc ucoEo>ao>ca non Cu poumamu >ad>auamom cOHuoowmaumm com o» nonnamu >am>auamom muasmom soummmom Aconcaucoov ceaum>aumop anCOamaoop mo ousmmoe mocmoawca om>aoouom mo ousmmms apnoeaucoov mUCQSauca po>aoouom .m pouaamcoauoummo ceamcman 30m aouucou anaemuom a manna 64 tancies. A measure of Effort to Outcome expectancies pro- vides an indication of one's belief that his or her effort will lead to certain outcomes. Thus, Effort to Outcome ex- pectancies more closely approximate outcome control than either Effort to Performance or Performance to Outcome ex- pectancies. The JDS and JCI autonomy scales were the most commonly used measures of activity control. Both the JDS and JCI autonomy scales are self—report measures of individuals' perceptions of the level of autonomy in their jobs. Auton- omy has been defined as one's perception of freedom in scheduling the work and determining the procedures to carry it out (Hackman and Oldham, 1975). The JDS and JCI auton- omy scales assess one's perception of a generalized notion of freedom or independence 0n the job and, as such, may not fully capture the perception of activity control. There- fore, a scale was designed to assess an individual's per- ception of control over specific work activities related to how an individual determines and carries out his or her work. Two categories of Locke and Schweiger's (1977) cate- gorization of work activities are relevant to activity con- trol--the work itself and working conditions. The work de- cisions selected for this scale were derived from these categories and include: 1. the Speed with which you do your work 2. the setting of work deadlines 65 3. the selection of work tasks that you perform 4. when you take your rest breaks 5. the choice of methods to do your work 6. the layout of your workspace 7. the setting of performance goals 8. the choice of equipment to do your work 9. determining the order in which you will do your work 10. the Specific hours you work each day. Subjects were asked to rate the amount of control they have over these ten work activities on a five-Option scale rang- ing from no control to complete control. Thus, activity control was operationalized in this study by assessing in- dividuals' perceptions of control over specific activities and decisions related to how one goes about performing his or her job. Perceived influence has been Operationalized in several different ways in the research literature. Researchers have manipulated the actual amount of influence individuals had in organizational decision-making operations and ob- served the results (e.g. Coch and French, 1948). Perceived influence has also been operationalized using a self-report measure of perceived influence (e.g. Vroom, 1960), as well as discrepancy scores of an individual's actual-versus- desired involvement in organizational decision-making opera- tions (e.g. Alutto and Belasco, 1972). Several researchers have warned of the reliability and validity problems associ- ated with the use of discrepancy measures (Cronbach and 66 Furby, 1971; Johns, 1981). Further, since personal control has been defined in this paper as one's perception of con- trol over his or her immediate environment, it was appro- priate that perceived influence be operationalized as a perceptual measure in this study. The most commonly used perceptual measure of influence was Vroom's (1960) measure of psychological participation. The psychological participation scale assesses one's per- ception of influence over the decisions made by his or her supervisor. The scale, however, does not include specific decisions in which one might be involved. Therefore, a measure of one's perception of involvement in and control over Specific work decisions was developed for the same rea- sons that the activity control scale was developed. Basi- cally, it was believed that the existing scale was too gen- eral and that assessing one's perception of influence by sum- ming his or her perceptions of influence over a variety of specific work decisions would be a more accurate measure of perceived influence. Further, individual items in the new scale can be used to determine one's perception of influence over a particular decision area. This could be used to focus management's effort in attempting to increase employee involvement in decision-making operations to those areas that are most deficient. The decision areas selected for use in the perceived influence scale were derived from the remaining two cate- gories of Locke and Schweiger's (1976) list--routine per- 67 sonnel functions and company policies. All of the decision areas selected affect one's worklife, however, decisions related to how one performs his or her job were excluded to maintain a conceptual distinction with the activity control scale. The decision areas selected were: 1. hiring new employees 2. your promotion 3. your performance appraisal 4. training new employees 5. your pay raise 6. discipline procedures 7. evaluation of other personnel 8. allocation of department budget 9. assignment of personnel 10. department layoff policy 11. department policy making 12. department wage level 13. department promotion procedures 14. department performance appraisal procedures. The perceived influence instrument had subjects rate their past level of involvement in the 14 decison areas using the following scale developed by Heller, Drenth, Koopman and Rus (1977): 1. No advance information was provided to you concerning the decision. 2. You were informed in advance of the decision to be made. 68 3. You were able to voice your opinion concerning the decision. 4. Your opinion concerning the decision was taken into account in the decision-making process. 5. The decision was made jointly with equal authority be- tween yourself and your supervisor. 6. The decision was entirely your own with no involvement from anyone else. The Heller et a1. scale provides a decided advantage over the response format utilized in Vroom's psychological par- ticipation scale. The Vroom measure uses a five-point scale ranging from very little to very much. In contrast, the Heller et al. scale provides specific behavioral indicators to anchor each scale value. In sum, the three dimensions of personal control will be operationalized using perceptual self—report measures. Outcome control will be assessed using an Effort to Outcome expectancies scale derived from Lawler's (1980) scale. Activity control and perceived influence will be measured using newly developed scales that assess control or influ- ence over specific work activities or decisions. A major hypothesis of this study was that the three organizational behavior variables--expectancy of control, autonomy, and perceived influence--are much more similar than one would expect given the different theoretical orien- tations and practical applications associated with each con- struct. It was proposed that these constructs are related 69 since they each involve an important aSpect of one's per- ception of control in a work organization. The research literature involving these constructs was examined to iden- tify the important antecedents and consequences of each of these dimensions of personal control. An examination of the relationships between each personal control construct and its antecedent and outcome variables was made to deter- mine the similarities and differences among the personal control constructs and to guide the development of a model of personal control. It is important to note that no study has examined the relationships between the different personal control vari- ables. Thus, little is known regarding the interrelation- ship of these three variables. Evidence concerning the similarities and differences among the personal control di- mensions, however, can be inferred on the basis of their re- lationships with other variables. Nunnally (1967) proposed that a test of how similar different measures are is the ex- tent to which they have a similar pattern of relationships with external variables. Table 2 provides a summary of the relationships be- tween antecedent and outcome variables and each of the per- sonal control variables. An "X" on this chart indicates that a relationship between that particular personal control variable and that antecedent or outcome variable has been suggested by theory or demonstrated through research results. Table 2 shows different patterns of relationships for each Summary of the Literature Reviewing the 70 Table 2 Relationship Between Antecedent and Outcome Variables and the Personal Control Dimensions. ANTECEDENT VARIABLES Mood Locus of Control Field Dependence E —9 P Perceptions Valence of Outcomes Self-Esteem Physical Environment Organizational Structure PDM Dimensions 1 Type of Supervision Communication Social Density OUTCOME VARIABLES Satisfaction Job Involvement PERSONAL CONTROL DIMENSIONS Outcome Control X X Organizational Commitment Union Attitudes Stress Related Physical Strain Emotional Strain Effort/Motivation Turnover Intention Attendance Behavior Productivity Activity Control X X Perceived Influence ><><><><>< 71 personal control variable with the antecedent and outcome variables, thus providing some evidence that the personal control dimensions are different constructs. On the other hand, the overlapping relationships the personal control variables have with some of the external variables provides evidence of the similarity among the personal control di- mensions. In sum, the results of the literature review pro- vides some support that expectancy of control, autonomy, and perceived influence are related variables. Figure 9 illustrates a mediational model of personal con- trol that the present study will empirically test in a field setting. The antecedent variables consist of two types of variables-- personality and situational variables. The personality variables include mood and locus of control, while job sta- tus (i.e. high control or low control) constituted the sit- uational variable. Each of the antecedent variables are proposed to affect the mediating variables (i.e. outcome control, activity control, and perceived influence). The personal control variables, in turn, are related to the out- come variables: intrinsic and extrinsic satisfaction, job involvement, organizational commitment, physical and psycho- logical strain, work effort, union attitudes, and turnover intentions. The personal control constructs are proposed to be re- lated but not identical constructs. Therefore, they should have similar patterns of relationships with the antecedent and outcome variables but not identical ones. Table 3 il- 72 aouucoo amcomumm mo moEoouso paw mucopmomucm ocu mo aopoz .m madmam mSOausmucH um>ocusm74 mopsuauud coacs pnommm camuum amoamoaonommm paw amoammnm msumum non ucosuaEEou mucoSamcH po>amouwm amaoaumNacmmuo aMdOaumsuam ucoEm>ao>cH non m1 aouucoo mua>apo¢ m coauomwmaumm aouucoo mo msooa oamcauuxm Ugo oamcauucH aouucou mEoouso p002 NwaaMCOmHom moanmaum> oEoopso moanmaum> aouucou anaemuwm moanmaum> ucmpmomusa 73 Table 3 Hypothesized Relationships Between Antecedent, Personal Control and Outcome Variables Personal Control Variables Outcome Activity Perceived Control Control Influence Antecedent Variables Mood ‘ ' - Locus of Control — ’ ' Job Status - ' - Outcome Variables Intrinsic Satisfaction NR + NR Extrinisc Satisfaction + NR + Job Involvement + + NR Organizational Commitment + + + Effort/Motivation + + NR Physical/Psychological Strain - - - Union Attitudes - NR - Turnover Intention NR NR - 74 lustrates the hypothesized relationships of each personal control construct with the antecedent and outcome variables employed in this study. A "+" symbol on this chart indi- cates that a positive relationship is hypothesized between that antecedent or outcome variable and the personal con- trol dimension, a "-" symbol signifies that a negative rela— tionship is believed to exist, and an "NR" indicates that no relationship is hypothesized to exist. An explanation of the rationale used in determining the hypothesized rela- tionship between the personal control dimensions and each antecedent and outcome variable is given below. Mood Considerable research indicates that one's mood state is highly related to his or her perception of control (e.g. Alloy and Abramson, 1979). In a laboratory study, naturally depressed students who were temporarily made elated in the laboratory exhibited an illusion of control when judging the amount of control that they had over uncontrollable events (Alloy, Abramson and Viscusi, 1981). In contrast, natrually nondepressed students who were temporarily made depressed showed no illusion of control and accurately judged their personal control over the event. Thus, the extent to which an individual feels depressed will lower his or her perceptions of outcome control. Allen and Greenberger (1980) have suggested an explana- tion for this relationship: "It is probable that positive affect and high control have been frequently associated in 75 an individual's past experience. Therefore, when a person experiences positive affect, it is likely that he or she will also perceive greater personal control than when in a negative mood." (p. 89.) Thus, a negative relationship is proposed between (depressed) mood and outcome control. It is likely that mood is also related to activity con- trol and perceived influence since they both involve one's perception of control. Caldwell and O'Reilly (1982) found an indicator of positive affect (i.e. job satisfaction) to be causally related to perceptions of autonomy. Thus, mood is pr0posed to be negatively related to all three dimensions of control. Locus of Control Locus of control involves the generalized expectancy of control over one's environment and, as such, should be positively related to all three dimensions of control. A number of studies have provided empirical evidence of posi- tive relationships between locus of control and both acti- vity control (e.g. Szilagyi and Sims, 1975) and outcome control (Kimmons and Greenhaus, 1976). No study has ex- amined the relationship between locus of control and per- ceived influence; however, such a relationship seems likely. An "internal" individual would probably see him or herself as having greater influence in decision-making operations than would an "external" person. Therefore, locus of con- trol is hypothesized to be positively related to all three dimensions of control. 76 Job Status An important antecedent of one's perception of control in an organization is the actual amount of control provided him or her on the job. The amount of control afforded in- dividuals in different jobs varies according to the type of job and the status associated with that position in the or- ganization. The two jobs chosen for this study--faculty members and clerical workers--were selected because of the large discrepancy in the amount of control individuals in each group have over their work lives. In general, faculty members have a great deal of freedom and control over many aspects of their job. In addition, faculty members have some influence in departmental decision-making operations by serving on committees and voting at department meetings. Clerical workers, in contrast, are characterized by strict work rules, close supervision, and little personal discre- tion in how to perform their jobs. Job status (coded l for faculty members and 2 for clerical workers) should be nega- tively related with all three dimensions of control. Satisfaction Job satisfaction has been theoretically or empirically linked to each of the personal control variables. Thus, it would be difficult to prOpose which dimensions of personal control would be more highly related to satisfaction.- Con- ceptualization and measurement of both intrinsic and extrin- sic satisfaction, however, may provide a useful means of differentiating between the personal control dimensions. 77 Intrinsic satisfaction is theorized to result primarily from satisfying work activities, while extrinsic satisfac- tion is related to one's contentment with factors external to work activities (e.g. pay and supervision). Control over outcomes or influence in decisions, therefore, should not be related to feelings of intrinsic satisfaction. Ac- tivity control, on the other hand, should be highly related to intrinsic job satisfaction since activity control implies control over intrinsic job factors. Outcome control and perceived influence, however, should be positively related to extrinsic satisfaction since they imply control over at least some of these ex- trinsic job factors. Outcome control, however, only in- volves control over outcomes related to one's job, while perceived influence includes influence over factors unre- lated to one's job (e.g. company policy). Therefore, out- come control should be more highly correlated with extrin- sic satisfaction than would perceived influence. Job Involvement Job involvement involves one's psychological identi- fication with a particular job and is determined, to some extent, by the individual's perceptions of the need- satisfying potentialities (both intrinsic and extrinsic) of the job (Kanungo, 1982). An individual with a high level of activity control is responsible for his or her work and has the freedom to perform the job as he or she 78 desires. Such an individual should, therefore, be highly involved in his or her job. Moch (1977) found a high positive relationship between activity control and job involvement. No research has in- vestigated the link between outcome control and job involve- ment, although such a relationship seems likely. An indi- vidual with high outcome control should be involved in his or her job since it fulfills certain needs. Perceived in- fluence should have less of a relationship with job involve- ment since influence involves control over decisions related to one's worklife rather than his or her job itself. Thus, job involvement should be most highly related to activity control, somewhat less highly related to outcome control, and not related to perceived influence. Organizational Commitment Organizational commitment involves acceptance of the organization's goals and values, a willingness to exert effort for the organization, and a desire to retain member- ship in the organization (Porter, Steers, Mowday and Boulian, 1974). In essence, commitment involves attachment to the organization. Perceived influence should have the greatest impact on one's attachment to an organization since involve- ment in decision making should lead an individual to feel greater ownership and acceptance of those decisions. Fur— ther, involvement in decision making implies that one has influence over the actual decision, and thus the decision, to some extent, is representative of his or her values and 79 beliefs. Alutto and Belasco (1972) have found a strong negative relationship between decisional deprivation and organizational commitment. High levels of outcome control and activity control might lead an individual to enjoy working for the organi- zation, thus increasing his or her desire to remain in the organization. Therefore, perceived influence is hypothe- sized to have the highest correlation with organizational commitment. Outcome control and activity control are ex- pected to have lower, but positive, relationships with or- ganizational commitment. Effort/Motivation Considerable theoretical and empirical research sug- gests a positive relationship between effort and both out- come and activity control. According to expectancy theor- ists, a high expectancy of receiving a highly valent reward for one's work will lead to a high level of work motivation. Job design theorists have proposed that activity control leads to perceived responsibility, which brings about high internal work motivation. Activity control increases one's level of effort through intrinsic satisfaction factors (i.e. autonomy). Outcome control, in contrast, motivates indi- viduals through extrinsic satisfaction factors (e.g. pay). Therefore, an employee who prefers intrinsic factors would have higher correlations between activity control and effort than employees who more highly desire extrinsic factors. Regardless, a positive relationship is hypothesized between 80 effort and both outcome and activity control. There is no theoretical or empirical evidence to prOpose a relationship between perceived influence and effort. Physical and Psychological Strain The three dimensions of control are similar in that they all impact on one's general feeling of personal con- trol. Considerable theoretical and empirical research suggests that this feeling of control is important for one's physical and mental well-being (Blauner, 1964; May, 1972). Thus, a negative relationship between the measures of strain and each dimension of control is possible. Given the popu- lations chosen in this study (i.e. faculty and clerical workers), however, it seems unlikely that either activity control or outcome control levels would be sufficiently low to cause excessive strain. Therefore, a low positive relationship is expected between the measures of strain and both activity control and outcome control. In contrast, levels of perceived influence could be quite low, particu- larly among clerical workers. The inability to influence decisions that affect one's worklife can be quite stressful. Researchers have found a strong negative relationship be- tween perceived influence and several different indicators of strain (Caplan, Cobb, French, Harrison and Pinneau, 1975). Alutto and Vredenburgh (1977) found a positive relationship between decisional deprivation and job tension. Therefore, strain should be most highly related to perceived influence and less so to activity or outcome control. 81 Union Attitudes Unions are one means with which employees can increase the amount of control they have in an organization. While other factors also influence one's attitudes toward unions, the amount of control the individual is able to exercise within the organization should also impact on his or her union attitudes. Alutto and Belasco (1972) found a strong positive relationship between decisional deprivation and union attitudes. Therefore, a negative relationship is hy- pothesized between union attitudes and perceived influence. No research has investigated the relationship between out- come control or activity control and attitudes toward unions. For individuals with low levels of activity con- trol, it is more likely that one would attempt to increase control through his or her supervisor rather than desiring union intervention. Thus, no relationship between activity control and union attitudes is hypothesized. Individuals with little outcome control, however, might see unions as a viable means of increasing the likelihood of their receiving certain desired outcomes from the organization (e.g. pay raises, promotions), since these outcomes are often part of collective bargaining agreements. Thus, a negative rela- tionship is proposed between both outcome control and per- ceived influence and union attitudes, and no relationship is expected between activity control and union attitudes. Turnover'Intention Turnover intention is most frequently a result of one's 82 dissatisfaction with some aspect of his or her job. As such, low levels of any of the personal control variables might lead one to think about changing his or her job. Little empirical research has investigated the relationship between any of the personal control variables and turnover intention. It is proposed that perceived influence will have the highest correlations with individuals' desires to leave an organization because it is more indicative of one's relationship to the organization than either activityr or outcome control. Further, perceived influence is hy- pothesized to have the highest relationship with organiza- tional commitment, which is also a predictor of turnover intention (Steers, 1977). The purpose of the present study is to empirically ex- amine the multidimensional conceptualization of personal control and to test the mediational model of personal con- trol in a field setting. The psychometric properties of the newly developed instruments used to measure the per- sonal control dimensions will be examined. In addition, the hypothesized relationships between each personal con- trol dimension and the antecedent and outcome variables will be tested. CHAPTER III METHOD Subjects The sample consisted of two distinct populations of sub- jects--the faculty and clerical staff of a large midwestern university. The faculty sample included all full-time tenure track faculty members. The clerical sample consisted of all full-time clerical employees. Faculty or clerical staff mem- bers with formal supervisory or administrative duties were excluded from this study. Procedure Questionnaire packets containing a cover letter explain- ing the purpose of the study, the questionnaire, and a computer- scan answer sheet were mailed to 1,768 faculty and 1,624 cleri- cal staff members. Questionnaire packets were mailed to each subject's work location and completed questionnaires were to be returned directly to the psychology department, thus ensur- ing the confidentiality of the responses. Anonymity of sub- jects was maintained since completed questionnaires did not contain their names or any identifying numbers. A follow-up letter was mailed to all subjects two weeks later to remind them to complete and return the questionnaires. Instruments Demographic Variables (See Appendix A) The demographic variables consisted of length of time employed by the organization, sex, education and job level. 83 84 Length of time employed was measured using a six-option re- sponse scale: (1) less than 6 months, (2) 6 months to 1 year, (3) l to 5 years, (4) 6 to 10 years, (5) 11 to 20 years, (6) over 20 years. Respondents were asked to indi- cate their ggx by responding with a 1 if male and a 2 if female. Education was measured on a scale with five re- sponse options: (1) high school graduate or less, (2) some college--no degree, (3) two-year college degree, (4) four- year college degree, (5) graduate degree. To determine jgp 13331 it was necessary to ask different questions of the faculty and clerical staff members. Clerical positions in this organization are ordered in a civil service-type hier- archy of levels ranging from level 4 to level 12. The lower level positions consist primarily of highly super- vised clerical positions. The middle levels include secre- taries and administrative assistants, while the highest level positions consist of technical and highly responsible administrative jobs. Clerical employees were asked to in- dicate their job level on a scale of 4 to 12. Faculty mem- bers, on the other hand, were asked to indicate their £23k on a four-option scale: (1) professor, (2) associate prof- essor, (3) assistant professor, (4) instructor. Antecedent Variables (See Appendix B) Locus of control was assessed using an ll-item form of Rotter's original measure (1966) used previously by Schmitt, Coyle, Rauschenberger and White (1979) who reported an in- ternal consistency reliability of .70. This short form con- 85 sists of the more adult and work-oriented items from the original form and utilizes five-point Likert-type scales ranging from strongly agree to strongly disagree. High scores on this scale indicate an external locus of control, while a low score is indicative of an internal locus of control. Participants' current mood state was assessed us- ing a modified version of the Depression/Dejection scale of the Profile of Mood States (POMS) (McNair, Lorr, and Droppleman, 1971). This instrument asks subjects to rate the extent to which each of 15 adjectives describe how they currently feel using a five-point scale ranging from not at all descriptive to extremely descriptive. The items used included: "unhappy,"'hdserable" land "guilty." The modifi- cation to the scale involved changing the directions to read "how are you feeling today" rather than "how have you been feeling during the past week, including today." This modification was made in an attempt to assess one's current mood as he or she worked on the questionnaire rather than his or her personality. McNair, Lorr and Droppleman (1971) reported an internal consistency reliability of .95 for the POMS Depression/Dejection scale. The final antecedent variable utilized in this study was job status. The two populations chosen for this study--faculty members and clerical workers-—were selected because of the large dis- crepancy in the amount of control individuals in each group have over their work life. In general, faculty members have a great deal of freedom and control over many aspects 86 of their job. In addition, faculty members have some in- fluence in departmental decision-making operations by serv- ing on committees and voting at department meetings. Cleri- cal positions, in contrast, are characterized by strict work rules, close supervision and little personal discre- tion in how to perform their job. Respondents' job status was coded 1 for faculty members and 2 for clerical workers. Personal Control Variables (See Appendix C) Three different aspects of personal control were assessed: (1) outcome control, (2) activity control, and (3) perceived influence. Outcome control was measured us- ing a modified version of Lawler's (1981) ll-item Perfor- mance to Outcome expectancy scale. The instructions were modified to change the scale to an Effort to Outcome expec- tancy measure. The revised instructions read, "Listed be- low are some things that could happen to people if they work hard at their job. How likely is it that each of these things would happen if you worked hard at your job?" Subjects indicated the likelihood of receiving each of 11 different work outcomes on a seven-option scale: (1) not at all likely, (2) unlikely, (3) somewhat likely, (4) likely, (5) quite likely, (6) very likely, (7) extremely likely. Activity control was measured using three different scales: The Job Diagnostic tSurvey (JDS)-Autonomy scale (Hackman and Oldham, 1975), The Job Characteristics Inventory (JCI)- Autonomy scale (Sims, Szilagyi and Keller, 1976), and a 87 newly develOped Activity Control scale. The JDS and JCI autonomy scales are each part of different multi-scale in- struments designed to measure an individual's perception of job characteristics. The JDS autonomy scale contains three items, each employing a seven-point scale. Pierce and Dun- ham (1978) reported an internal consistency reliability of .79 for the JDS autonomy scale. The JCI autonomy scale consists of six items using a five-point scale. Pierce and Dunham (1978) reported an internal consistency reliability of .85 for the JCI autonomy scale. The activity control scale consisted of ten items assessing respondents percep- tions of control over specific work activities. Respon- dents were asked to rate the amount of control they have using a five-option scale ranging from no control to com- plete control for each of ten work activities. These work activities included: "your pay raise," "discipline pro~ cedures" and "department wage level." Perceived influence was assessed using two different instruments: Vroom's (1960) measure of psychological par- ticipation and a new measure of perceived influence. The psychological participation scale consists of four items, each employing a five-point scale. The scale assesses one's perception of influence over the decisions made by his or her immediate supervisor. James, Hater and Jones (1981) reported an internal consistency reliability of .82 for the psychological participation scale. The perceived influence instrument had subjects rate their past level of 88 involvement in 14 decision areas using a six-point scale developed by Heller, Drenth, Koopman and Rus (1977) rang- ing from "no advance information was provided to you con- cerning the decision" to "the decision was entirely your own with no involvement by your supervisor." The decision areas used were derived from Locke and Schweiger's (1977) categorization of work decisions and included: "your pay raise," "discipline procedures" and "department policy mak- ing." Outcome Variables (See Appendix D) The short form of the Minnesota Satisfaction Question- naire (MSQ) was used to measure job satisfaction. This 20- item scale uses a five-option Likert-type scale that ranges from strongly agree to strongly disagree. The MSQ short form produces two subscales—-a 12-item intrinsic satisfac- tion scale and an eight-item extrinsic satisfaction scale. Weiss, Dawis, England and Lofquist (1967) reported median coefficient alphas of .90 for the overall MSQ short form, .86 for the intrinsic satisfaction subscale, and .80 for the extrinsic satisfaction subscale across a variety of different samples. Job involvement was measured using a ten-item scale developed by Kanungo (1981). Kanungo reported an internal consistency reliability of .90 for his job involvement scale. Organizational commitment was assessed using the lS-item Organizational Commitment Questionnaire (OCQ) de- 89 veloped by Porter, Steers, Mowday and Boulian (1974). Mowday, Steers and Porter (1979) reported a median internal consistency reliability of .90 for the OCQ across eight samples. Physical and psychological strain were also assessed. The Physical Strain Index (PSI) asked subjects to indicate how frequently they were bothered by four physical problems (i.e. upset stomach, backache, headache and fatigue) on a five-option scale ranging from not at all to every day. The short form of the General Health Questionnaire (GHQ) (Goldberg, 1972) was used to assess psychological strain. This instrument asks subjects to respond to 12 questions using a five-option scale which ranges from not at all to much more than usual. Some examples of the questions asked are: "Lost much sleep over worry?," "Felt constantly under strain?," and "Been able to face up to your problems?". Banks, Clegg, Jackson, Kemp, Stafford and Well (1980) re- ported coefficient alphas ranging from .82 to .90 for six samples using the short form of the GHQ. Attitudes toward union in general were assessed using the 20-item unionism-in-general scale of the Institute for Social Research Union Attitude Scale (Uphoff and Dunnette, 1956). This instrument asks subjects to respond to a num- ber of positive and negative statements concerning unions on a five-point Likert-type scale ranging from strongly agree to strongly disagree. A coefficient alpha of .88 was reported for this scale by Schriesheim (1978). Effort/ 90 motivation was measured using the four—item Job Motivation Index (Patchen, 1965). One additional item was added ask- ing subjects to indicate the amount of uncompensated time that they spend at work on a five-option response scale ranging from almost every day to about once a month or less. Finally, turnover intention was measured using a single-item scale which asked subjects to indicate how they felt about leaving or staying with the organization on a five-option scale: (1) strongly inclined to leave, (2) in- clined to leave, (3) don't know whether I want to stay or leave, (4) inclined to stay, (5) strongly inclined to stay. Data Analysis Chi Square tests were used to determine whether or not the sample obtained in this study is representative of the pOpulation from which it was drawn. Specifically, Chi Square tests determined whether the respondents differed significantly from the population from which they were drawn in terms of level of education, sex, time employed in the organization, and job level. Separate analyses were performed for the faculty and clerical samples. The psychometric properties of the personal control scales were examined. Coefficient alphas were computed to determine the internal consistengy reliability of each per- sonal control scale. The comparability/distinctiveness of the new personal control scales were then examined in rela- tion to the existing scales using several different analy- 91 ses. First, the intercorrelations among the six personal control scales were examined for evidence of convergent and discriminant validity. Second, the item-scale correla- tions of all the items comprising the six personal control scales were examined to assess the empirical distinctive- ness between the new and existing scales. Third, the ex- ternal consistency of the personal control scales was assessed by examining the pattern of correlation each per- sonal control scale had with a set of relevant organiza- tional behavior variables. Finally, the extent to which the new scales explained additional variance in the depen- dent variables (i.e. antecedent and outcome variables), beyond that accounted for by the existing scales, was de- termined using a series of hierarchical multiple regression analyses in which the existing scale was entered first and the new personal control scale was entered second. The zero-order correlations between the personal con- trol variables and each antecedent and outcome variable were examined to test the hypothesized relationships among these variables. The mediational model of personal con- trol was then tested using a series of hierarchical multi- ple regression analyses. In these analyses, two possibly confounding demographic variables--sex and educational level--were entered into the regression equations first. The personal control variables were entered in the second step, and the antecedent variables were entered last. This analysis was performed for each of the outcome variables. 92 To test the mediating hypothesis, this hierarchical re- gression was compared with one in which the demographic variables were entered first, the antecedent variables were entered second and the personal control variables were entered last. Finally, the relationships between the personal con- trol variables and each of the outcome variables were ex- amined while statistically controlling for the effects of the demographic and antecedent variables. These analyses assessed the degree to which the personal control variables explained additional variance in the outcome variables be- yond that accounted for by the demographic and antecedent variables. CHAPTER IV RESULTS AND DISCUSSION Response Rate Of the 3,392 questionnaires mailed to the combined faculty and clerical worker samples, 1,078 usable question- naires were returned, yielding a response rate of 32 per- cent. A considerable difference, however, exists between the reSponse rates of the two sub-samples--faculty and clerical staff members. Of the 1,768 questionnaires sent to faculty members, 423 questionnaires were returned in usable condition, producing a response rate of 24 percent. In contrast, the reSponse rate among the clerical sample was 40 percent, with 655 usable questionnaires returned from the 1,624 that were mailed. A possible explanation for this discrepancy in response rates is that clerical workers may have found time to complete the questionnaire while at work. Because clerical workers are required to put in eight hours a day at work, those workers who filled out the survey during working hours may have felt they were doing it on company time. Faculty members, on the other hand, have far greater discretion over the amount of time they spend at work and may have felt that the survey infringed on their personal time. Representativeness of the Sample An important consideration in the interpretation of survey results involves determining whether or not the re- 93 94 spondents differ in a significant way from those individ- uals who did not return the questionnaire. Tables 4 and 5 show comparisons of sample and population demographic characteristics for faculty and clerical staff members re~ spectively. Percent comparisons of the samples and popula- tions on the demographic characteristics of sex, education, length of time employed, and job level generally indicated that the samples reflected their respective populations. One exception to this pattern involves the educational le- vel of clerical workers. Twenty percent of the clerical sample reported their educational status as high school di- ploma or less, and 47 percent indicated that they had some college but no degree. In contrast, the population data indicates that 43 percent of the clerical workers were at high-school-diploma-or-less level and only 28 percent have some college education. It is quite likely that this dis- crepancy is due to an error in the population records ra- ther than a sampling bias. Educational data for the popula- tion of clerical workers were obtained from university files. These data are collected at the time an employee is hired and is not updated unless the employee earns a degree. Thus, an employee who has a high school diploma at the time of his or her hiring and subsequently takes a few college courses would not have this change in educational status listed on his or her personnel file. Chi Square tests were performed to determine whether the faculty and clerical worker samples differed signifi- Table 4. 95 Characteristics for Faculty 5 Sex Male Female Education High School College - no degree Associate's degree Bachelor's degree Graduate degree Length of Time Employed 0-6 months 7 months-1 year 1-5 yearss 6-10 years 11-20 years over 20 years Job Level Instructor Assistant Professor Associate Professor Full Professor Ignure Yes No Comparison of Sample and Population Demographic Population Percentage Sample Percentage 1,792 100 423 23.6 1,522 85 341 81 270 15 82 19.4 0 0 1 .2 4 .2 0 0 0 0 0 0 6 .3 2 .5 1,782 99.4 416 99.3 9 .5 2 5 51 2.8 15 3.5 363 20.2 63 15 295 16.5 59 14.1 749 41.8 179 42.7 325 18.1 101 24.1 4 .2 3 .7 315 17.6 81 19.6 494 27.6 94 22.2 979 54.6 245 57.9 1,500 83.4 339 81 292 16.3 80 19 96 Table 5. Comparison of Sample and Population Demographic Characteristics for Clerical Workers Population Percentage Sample Percentage N ‘ 1,586 100 655 41.3 Sex __Male 38 24 22 3.4 Female 1,548 98 630 96.6 Education High School 461 43 133 20.4 College - no degree 303 28.2 306 46.9 Associate's degree 56 5.2 49 7.5 Bachelor's degree 223 20.7 131 20.1 Graduate degree 32 3 33 5 Missing cases 512 -- -- -- Length of Time Employed 0-6 months 110 6 9 34 5.2 7 months—1 year 44 2.8 24 3.7 1-5 years 652 41.1 255 38.9 6-10 years 409 25.8 180 27.5 11-20 years 316 19.9 132 20.1 over 20 years 55 3.5 30 4.6 Job Level Level 4 49 3.1 14 2.2 Level 5 331 20.9 130 20.1 Level 6 337 21.2 124 19.2 Level 7 484 30.5 211 32.7 Level 8 205 12.9 94 14.6 Level 9 152 9.6 60 9.3 Level 10 28 A 8 l3 2 97 cantly from the populations from which they were drawn in terms of sex, education, length of time employed, and job level. Table 6 shows the Chi Square results for the facul- ty sample. The results indicate that the faculty sample differed significantly from its population in several char- acteristics: sex (p<:.05), length of time employed (p<:.01) and job level (p<1.05). Specifically, females, faculty mem- bers with over 20 years of tenure, and full professors were somewhat overrepresented in the sample. While these dif- ferences are statistically significant, it does not appear that the differences are large enough to represent any practical significance. The Chi Square results for the clerical sample are sum- marized in Table 7. The clerical worker sample differed significantly from its population values only in regards to educational level. As discussed earlier, this difference may be artifactual in nature (i.e. inaccurate pOpulation data records). In sum, both the faculty and clerical worker samples appear to adequately reflect the demographic char- acteristics of the population from which they were drawn. Psychometric Properties of the New Personal Control Scales Before testing the model of personal control described earlier, the psychometric properties of the new or revised personal control scales were examined. Specifically, the internal consistency reliability of each personal control scale was assessed and the comparability/distinctiveness of the new personal control scales was examined in relation to 98 Ho. Va «.4 mo. VQ s mmm amm mow om mm 02 am. a madame mam amm MOmmmmoum aasm mm baa HOmmwmonm onwaoommd mm mm nommmmoum ucmumammd pan uouosuumcH amv.m N ao>ma now aoa mm mung» om Hm>o aha maa mamas omiaa mm mm whom» oaim mm mm mung» mia ha va How» a Hoods sama.wa v pomoamEm mEaB mam has mmummp mumspmuo m m mmumop mumsooum oz oo.a a coaumospm mm mm mamamm awm mmm mam: «mm.m a xmm mumsqm ano mm m5am> moam> whommumu manmmmmw pmbummno pmuowmxm coaumasmom muH paw mamEmm muasomm on» cmmSumm moaumaumuomumnu oanmmumOEwo ca mmocmummwao mg“ no umma mumswm anO .m manna 99 ao.uvm as mommo mcammae Nam mos dump GOaumasmom .m Nw.m mm.b skma.mma am.~ mundwm acu ma om vm Ham vma oma va om Nma oma mmm vm mm mm ama ma mom mma 0mm NN mp moam> pm>uwmno mamEmm menoz amoaumau may cmmzumm moaumaumuomumnu oanmmuonmo ca mmocmuwmmao on» mo umwB mumswm H30 aa mm vm hma hma mma om vm oma mma mom ma mo ma mma vm «ma 0mm mmm ma msam> pmuommxm ao>ma am>ma am>wa am>wa am>ma am>ma am>ma vmmhoomu—i mung» om Hm>o mung» omiaa mummh oaim mummm mia Ham» aimnucoe n Hmmw aimnucoe m mmumwp mumspmuw mmummp m.H0amnomm mmummp m.mumaoomm¢ mmummp mmmaaou mmwa no moummp aoonom swam mamEmm mam: mnommumu coaumasmom mUH pan am>ma nos pmNOaQEm mEaB msoaUMUSpm xwm magmaum> .k magma 100 the existing scales. Internal Consistengy Internal consistency involves an estimate of the re- liability of a measure based on the average correlation/covari- ance among the items in a scale (Nunnally, 1967) . Stone (1978) proposed that an internal consistency estimate should be used whenever "the researcher wishes to assess the degree to which the items in a measure are homogenous (i.e. indices of a common construct)." (p. 49.) Nunnally (1967) de- scribed internal consistency as a necessary, although not sufficient, condition of the construct validity of scale. Table 8 contains the coefficient alphas, as well as the number of items for each personal control scale: the JDS Autonomy Scale (JDS), the JCI Autonomy Scale (JCI), the Ac- tivity Control Scale (AC), the Perceived Influence Scale (PI), Vroom's Psychological Influence Scale (Vroom), and the Outcome Control Scale (OC). The coefficient alphas range from .74 to .90, indicating that the internal consistency reliability is adequate for all the personal control mea- sures. Further, the alphas for the three personal control Scales (i.e. outcome control, activity control, and per- ceived influence) developed or revised for this study were among the highest, ranging from .87 to .90. Comparability/Distinctiveness of the New Personal Control Scales Two new scales were developed for this study--the ac- 101‘ Table 8 Coefficient Alphas for the Personal Control Scales. S2119. JDS Autonomy JCI Autonomy Activity Control Vroom's Participation Perceived Influence Outcome Control No. of Items 3 6 10 14 11 Coefficient Alpha .74 .87 .87 .83 .90 .88 a. Vroom's Participation Scale ordinarily contains four items, however, one item was inadvertently omitted on the questionnaire. 102 tivity control scale and the perceived influence scale--be- cause it was believed that the existing scales were too gen- eral to adequately capture the underlying constructs. The following analyses compared the new scales with the ex- isting scales in order to determine their comparability and distinctiveness. Ideally, the new scales will be highly related (i.e. comparable) to the existing scales since they were intended to measure the same constructs. The new scales, however, should be better or different from the ex- isting scales in some way (i.e. distinctive) if the new scales are to be of any practical value. Several different analyses were performed to compare the new with the existing personal control scales. First, the intercorrelations among the six personal control scales were examined for evidence of convergent and discriminant validity. Second, the similarities and differences of the personal control scales were further assessed by examining the pattern of correlations each personal control scale had with a set of relevent organizational behavior variables (i.e. external consistency). Third, the item-scale corre- lations of all the items comprising the six personal con- trol scales were examined to assess the empirical distinc- tiveness between the personal control scales. Finally, the extent to which the new scales explained additional variance in the dependent variables (i.e. ante- cedent and outcome variables), beyond that accounted for by the existing scales,was determined using a series of hier- 103 archical multiple regression analyses in which the existing scale was entered first and the new personal control scale was entered second. The preceding analyses also provided evidence of the similarities and differences between the three personal control dimensions--outcome control, acti- vity control, and perceived influence. Personal Control Scales/Dimensions Intercorrelations The intercorrelations of the personal control scales were examined for evidence of convergent and discriminant validity. Table 9 contains the intercorrelations of the six personal control scales. The highest intercorrelations were among the three scales used to measure the autonomy/ activity control dimension: JDS Autonomy, JCI Autonomy, and Activity Control Scales. These intercorrelations ranged from .64 to .73. In contrast, the correlation between the two scales used to assess the participation/influence scales was a more moderate .44. The results suggest that the autonomy/activity control scales have a high degree of convergent validity, while the participation/influence scales Show much less convergence. This indicates that the activity control scale is tapping the same construct as the JDS and JCI autonomy scales, while the perceived influence scale and the Vroom psychological participation scale represent related but different con- structs. The intercorrelations between scales measuring dif- ..S. 104 ——S SA--. SN- S. .N- SN. .NN. NS- .SS. W. a..- m— aa- Nv .SS. ma- Sm- m~ on- m.- .NS. cw- co So- 0—- No .mm. m_- cm- on NN .SS. .SoeaSSoe eta. o_mc.S oSoca to. poundEOS on uaaou SSSS.NSo SSSSNS.Scou Sucrose. o: .u .Seuacos .Su.ca—u NOS ~ tea Stones: SN_:SSS to. _ vocou SS: SSNSSS new .a .So—So to. w an. Swansea to. _ uauou SS: Sum .S ._u>o_ .c. was NS SSSS...=m.S SSS mo. cogs tuuoocm Sec—No.0Ncou .SoSS. So. oz. NS “coo...em.S eta co. cogs Nouoocm Sco.NS_occou .u_SSS on» sat. voNS.eo coon use: Sac—on .Sa.uoo .SogS—S aco_u.uaoou ScSSSNSoc .acomn.u was co Su:_o> No.02 SN- S.- S. .S SN- SN .N- SS- SN- SN- S.- S.- SSSS..SSS SS.¢S .SN SS- N.- NN- S.- SS SS S.- SN- SN- SN- SN- SN- SS.SSSSS. Lo>ocuah .a— SS .. SS- NS- SS NS- . SS S. SS SS SN .S ScoecS .S. SS- SN- NS SS SS- NS .S- .N- S.- SN- SN- .N- =.S.SS .Su_mo_o;uSSS .S— SS- N.- SS SN SN- SN S.- S.- SN- .S- SN- SN- S.S.SS .SS.SNSS .S. SS S. SS- S.- SS N.- SS SS SS SS SS .S SSSSS.SSSS .SSS.S.N.SSSLS .S. SS SS S. SS- SS SS- SS S. SS SS SS .S 29.9.2... 8.. .S. SS S. SS- S.- S. SN- NS SS .S SS SS SS SS.SSSSS.SSS S.S=.SSNS .S. .SS. SN SS- NS- NN NS- SS SS SS SS SS SS =S.SSSSS.SSS u_Sc.cu:_ .~. .NS. S. NS NS.. SS NN- S.- S.- SN- N.- S.- .SSSSSS No SauoL .__ .NS. SS SS- SS SN- SN- S.- SN- SN- SN- SSS: .S. 0.--. SS- SN SN- SS- SS- NS- NS- SS- SSSSSSS SSS .S 6.--. SS- S. SS N. SS SS SS .Sso. .SSS.SSSSSS .S 0.--. S.- SS- SS- SS- NN- .S- .SSS .N .an. m. an an on we accueou 988.5 .0 gem. co mo MS _m mucus—Sc— oo>.oucoa .m .SS. SS SS NS SS.S.S.S.SNSS soot, .S .Nm. me So .0eucou SS.>.SSS .S .NS. SN SaocSSSS .SS .N .SN. NsocoSSS SSS .. Ma aa ma w m a S w S a M a .SS.SS.NS> acousao can .ocScou .SSSSNSS .Scovouo.:< .u.gaocooeoc No SSNSSz co—No_otcoutuuc. m 0—Oc— 105 ferent personal control dimensions were moderate, ranging from .38 to .51. The activity control scale, for the most part, had lower correlations with the outcome control and perceived influence scales than did the other two autonomy/ activity control scales, thus providing some evidence of the discriminant validity of the activity control scale. The perceived influence scale, however, shows little evi- dence of discriminant validity since it correlated higher with the JDS autonomy and the activity control scales than it did with the Vroom scale. Further, the perceived influ- ence scale had somewhat higher correlations with the out- come control and autonomy/activity control scales than did the Vroom scale. The intercorrelations among the personal control scales also provides evidence of the interrelationships among the personal control dimensions--outcome control, activity con- trol, and perceived influence. The correlations among these three scales are much higher than one would expect, given the different theoretical orientations and practical appli- cations associated with each. The high intercorrelations among these scales provides some support for this study's hypothesis that these scales tap a highly related construct (i.e. personal control). External Consistency Additional evidence of the similarity and differences of a variable or construct can be acquired through examina- tion of the pattern of correlations each variable has with 106 several different (i.e. external) variables. Similar scales should have a similar pattern of relationships with other variables, while dissimilar scales should show un- related patterns of correlations with the external vari- ables. The external variables employed in this study included current mood state, locus of control, job status, intrinsic satisfaction, extrinsic satisfaction, job involvement, or- ganizational commitment, physical strain, psychological strain, effort/motivation, turnover intention, and atti- tudes toward unions. Table 9 shows the internal consistency reliabilities (i.e. coefficient alphas) of the external variables, as well as the intercorrelations between the per- sonal control and external variables. The coefficient a1- phas for the external variables were quite high, ranging from .67 to .92, indicating an acceptable level of internal consistency for all of these scales. The pattern of correlations between each personal con- trol scale and the set of external variables are very simi- lar. As expected, the pattern of correlation for the JDS Autonomy, JCI Autonomy and Activity Control are most simi- lar. For example, their correlations with mood range from -.23 to -.24, with locus of control they range from .16 to .20, and with turnover intention they range from .24 to .28. The similarity of these patterns provides further evidence that the activity control scale is tapping the same con- struct as the autonomy scales. 107 The two scales used to assess the participation/influ ence construct have the most dissimilar pattern of correla- tions, once again suggesting that these two scales are tap- ping separate constructs. The Vroom Participation scale also has the most dissimilar pattern of correlations with the ex- ternal variables of all the personal control scales. This suggests that the Vroom scale rather than the perceived in- fluence scale is the less appropriate measure of perceived influence. The patterns of correlations of five of the personal control scales are very similar. This provides additional support for the hypothesis that these scales are measuring similar constructs. In sum, based on both their intercor- relations and pattern of correlations with external vari- ables, it appears that the six personal control scales yield three highly related factors: JDS/JCI/AC scales, per- ceived influence and outcome control and 1 factor (i.e. Vroom Participation Scale) that is less highly related. Item-Scale Correlations Further evidence of the comparability and distinctive- ness of the personal control scales was obtained by examin- ing the item-scale correlations. Table 10 shows the item- scale correlations for all of the items comprising the six personal control scales. To ensure an unbiased correlation between an item and its own scale, the item was removed be- fore the correlation was computed. For the most part, items correlated higher with their 108 Table.10 Item-Scale Correlations for Personal Control Measures £33m Personal Control Measures £32 £91 4.9 VROOM 21 92 JDSl 60 11 67 4o 48 34 JSDZ 55 51 44 36 38 37 JDS3 g1 55 47 35 38 31 JCIl 51 £2 42 32 26 27 JCIZ 56 11 45 4o 34 27 JCIB 39 53 36 25 14 17 JCI4 67 18 63 42 43 35 JCIS 74 g9 55 47 49 43 JCI6 50 60 El 34 29 29 AC1 37 47 £5 24 20 23 AC2 46 48 fig 29 32 29 AC3 51 53 54 31 42 31 AC4 34 40 g; 23 27 19 AC5 58 60 12 35 39 30 AC6 35 35 53 26 31 19 AC7 48 54 E; 40 3o 31 AC8 39 39 §2 25 37 28 AC9 51 59 67 31 23 25 AC10 41 36 Z; 14 42 24 Vroom l 43 47 3; Z; 42 37 Vroom 2 45 47 4o 13 43 37 Vroom 3 33 32 32 ‘ 59 32 28 PIl 47 38 43 33 £1 32 P12 21 19 20 17 £1 25 P13 21 16 13 23 2g 22 PI4 35 3o 31 38 pg 29 P15 23 21 21 24 £1 27 P16 30 28 30 34 62 25 917 42 34 4o 35 72 31 109 TableIU)(continued) Item-Scale Correlations for Personal Control Measures Item PI8 PI9 P110 PI11 PI12 PI13 PI14 0C1 0C2 0C3 0C4 0C5 0C6 0C7 0C8 0C9 OClO OC11 Note: JDS 31 37 34 45 26 44 39 35 23 39 13 34 20 35 25 31 28 20 Personal Control Measures JCI 27 29 30 37 24 36 34 29 20 39 12 31 16 30 28 31 28 22 42 31 34 34 46 27 43 39 37 20 33 12 31 20 33 20 28 24 16 VROOM 29 37 25 35 25 28 28 22 12 20 18 27 24 19 30 30 54 25 corrected item-total correlations. ON (.0 mm ual-4 Correlations between an item and its own scale are 110 own scale than with any of the other scales, indicating a high degree of homogeneity of items in each of the scales. All items comprising the autonomy/activity control scales had high correlations with each of the autonomy/activity control scales (i.e. JDS, JCI and AC). This provides addi- tional evidence concerning the similarity of content of these three scales. Once again, however, the participation/influence scales showed marked differences. Items from the Vroom scale had very low correlations with the perceived influ- ence scale, and perceived influence scale items had low cor- relations with the Vroom scale. One of the outcome control items (i.e. 0C 10) had its highest correlations with the Vroom scale. This is inter- esting since the three Vroom scale items and 0C 10 refer to one's supervisor. This suggests that the Vroom scale may assess one's relationship with his of her supervisor rather than one's perception of influence in decision-making opera- tions, as previously believed. It was unfortunate that the fourth item in the Vroom scale (i.e. "In general how much say or influence do you have on what goes on in your sta- tion?") was inadvertently excluded on the questionnaire used in this study. Because it does not refer to one's su- pervisor, it would have been interesting to see on which scale it had its highest correlation--the Vroom scale or the perceived influence scale. 111 Variance in the Dependent Variables Accounted for by the New Scales A final set of analyses were performed to examine the distinctiveness of the new personal control scales--activity control and perceived influence. These analyses assessed the degree to which the new personal control scales ex- plained additional variance in the dependent variables be- yond that accounted for by the existing scales: JDS auton- omy, JCI autonomy, and Vroom participation. Specifically, hierarchical multiple regression analyses were performed on each of the antecedent and outcome variables with the exist- ing scale entered into the regression equation first and the new scale entered second. Tables 11 and 12 contain the simple correlations, standardized regression coefficients, and the multiple squared correlations for eaCh independent variable in these analyses. F-tests for the change in multiple R2 caused by the entry of each independent variable into the regression equation are also reported. This F-test de- termines whether a particular independent variable explains a significant amount of additional variance (of the depen- dent variable) beyond that accounted for by the other inde- pendent variables. The formula used in computing this stat- istic was described by Nie, Hull, Jenkins, Steinbrenner and Bent (1975, p. 336). Table 11 shows the results of the hierarchical regres- sion analyses of the participation/influence scales--per- 112 «tom.mm Nommm. «amm.mmh Nommm. «*Hm.Hm mOHmN. *«mw.mmm mmmom. ¥«hv.vma.a omnom. «*0m.ma mmHHo. *«mH.m hvmvo. *3mm.om vmmmo. hv. vommo. *twm.hv ommmo. mm AN mm HOW .m ma. Hm. om. Hm.l ma. Ha. «a. mo.l NN.I ml mv. ow. mv. mv. mm.| HH.I ha. ma. MH.I «N.I HI Hm Eoon> Hm Eoou> Hm Eoou> Hm Eoou> Hm Eoou> mmum moum hammv GOwuommmHumm camcfluuxm mmum mmum Ammmv cofluomwmflumm oflmcfluucH mmum mmum mmum mmum mmum mmum Commucm magmaum> moanmflum> accumuxm map so AHmv mamum mocmdamcH Um>flmoumm map can AvaV mmsumum non Amamv Houucoo mo msooq Aaomv U002 AZV mHQMHHm> pampcwmmo AEOOH>V mamom coflummfloflanmm HMUHmOHonowmm Eoou> Ilmmamom mocmsHmcH\coHummHoHuHmm mo mommamca coflmmmummm mamfluadz Hobasonmumflm mo muHSmwm .HH magma 113 «4mm.mma ¥*NO.HN mm.H ««ON.NV tivH.hN «*mm.mH *xmh.mv *4Nw.mmm *«mm.wma «*mm.mm Mwmnu .HOM rm thom. nmomo. mammo. Nmmmo. Hoamo. hmmao. ovmmm. mmmmm. mw¢HN. mwamo. N 0.1 mv. wo.l mo.l NN.I mm. ow. mv. Ho. ml N H N H mmum mmum Ammmv uuommm mmum mmum . Ammnv cannum HCUflmoHonohmm dmum mwum Ammsv cflmuum Hmoflmmnm mmum mmum Amomv ucwEuHEEou HmCOHumNHcmmuo dmum mmum Amomv ucmsm>ao>cH non mv. Hm va. Eoou> mH.| Hm vm.l Eoou> NN.I Hm wa.u Eoou> av. Hm om. Eoou> ow. Hm mm. Eoou> m Apmscflucouv Ha magma Commucm manmflum> AZV manmflum> ucmosmmma mo.uvm 3 Ho. Va 5; .umxuo3 Hmofiumao now N can umnEmE xuasomm How H Cmpoo ma muumum now .8 M .mmmmnpcmumm may CH CmumoHCCH mNHm mHmEmm map so Comma mos mammamcm conmmuomu l mamfluasa may .coflumamp mmfl3umwa mo mm: may cam mumo mcflmmHE on» NO mmsmomm "muoz 48mm.va *tma.mh itmm.HN *«mm.HH mmHAu you n mwmoa. vommo. vmmmo. mhmao. «a. MN. mH.I mo.l ml mN. Hm N mmum om. Eoou> H mmum ON.I Hm N mmum NH.I EOOH> H Qmum m pmumucm manmflum> Aamscflucoov HH magma Amway coflucmucH uw>ocuse Ammmv moosuwuud.coflcs sz mHQMHHm> ucmpcmmmo 115 ceived influence and Vroom participation with each of the dependent variables. The perceived influence scale ex- plained a significant amount of additional variance (p<:.01) in all but two of the dependent variables (i.e. mood and psychological strain) beyond that accounted for by the Vroom participation scale. The perceived influence scale account- ed for more variance in a number of dependent variables (i.e. job status, job involvement, physical strain, effort and union attitudes) than did the Vroom participation scale. These results indicate that the perceived influence scale is not redundant with the Vroom participation scale. Ra- ther, the perceived influence scale explains a significant amount of additional variance for many of the dependent variables and in a number of cases accounted for more vari- ance than did the Vroom scale. Table 12 shows the results of the hierarchical multiple regression analyses of the autonomy/activity control scales: JCI autonomy, JDS autonomy and activity control. irhe activity control scale explained a significant amount of additional external variable variance beyond that accounted for by the JCI and JDS autonomy scales for all of the variables ex— amined. Again, these results indicate that the activity control scale, while highly related to the two autonomy scales, 15.not merely a redundant scale. The activity con- trol scale explained a significant amount of additional variance beyond that accounted for by the two autonomy scales of a set of relevant organizational behavior vari- 116 *«mn.mv *«Nh.mmo 4*Nm.maa «*vm.HHN ¥«nm.HH xth.Hv *mH.m «*mw.mm mmum mmum Aommv coauomwmwumm UHmcHHuCH mmum mmum mmum mwum mmum mmum omHmv. wN. hm. Ud mmNov. ON. mm. won II om. mm. How HONVN. Nv.l 5v.l Ufi ommma. mH. Nm.l mDh II MN.I mm.l HUb moamo. ma. NN. U< mmmmo. No. ma. mnh II mo. ON. HOW mwmwo. HH.I mN.I Um Nomho. vo.l NN.I mob II mH.I mN.I HUh m .m. m mmHnmaum> Hmcumuxm map so Umumucm mHQMHHm> AHOOHV mmsumum now Avmmv Houucou mo msooq Avhmv U002 sz manmflum> unmocwmmn loav mHmum Houuaoo mufl>fluoa map mam Amobv mamom msoaous¢ man .AHUbv mamom >Eocou5¢ Hoolummamom Houucou mufl>fluo¢\>Eocou5¢ mo mmmhamcm scammmnmmm mamfluasz Hmoflnoumuwflm mo muadmmm .NH magma 117 «*vv.Hm mmooa. vN.I Hm.l Ufi N mmum «*wv.mh mmano. mo.| mN.I wow I: vo.u mm.n Hon H mmum 166mb chuum Hmonmam .4m4.~H ommmH. mH. mm. o4 m mmum «*mN.mom mthH. mo. mm. moo u- hm. H8. Hon H mmum Ammmv ucmEuHEEOU HmcoflumNHcmmuo «460.64 nmmHm. mm. m6. oa N mmum 44NH.nHN mmomH. «0.: mm. moo 1: mm. m4. Hos H mmum Hmmmv ucwEm>Ho>cH 90b 8«om.mm avmmN. vN. «v. Dd N mmum ¥¥om.VBN MONHN. NH. NV. mOb nu om. N4. Hos H mmum Hmmmv COHuommmHumm owmcflnuxm Nm 4 Mm .mu m omnmpcm mHQmHHmS sz magmaumg Mow m ucmpcwmmo lamscHuaoov NH mHgma 118 mo. VQ « HO.VQ «a .umxuos HmoHuwHo How N can HmnEmE wuHsomm new H Cocoa mH msumum now .0 .mwmmcucmnmm may CH wouCOHCCH mNHm mHmEmm on» :0 Comma mm3 mmeHmcm :onmwummu mHmHuHDE may .coHumep mmH3umHH mo mm: map can mump mchmHE may mo mmsmomm ”wuoz 45>.v vONmo. 0H. vN. Dd N mmum «4mm.Hm mvhho. mo. mN. Hun - II mH. mN. moo H omum Amway mcoHucmucH uw>ocusa «.o~.m~ mmmmo. m~.u v~.n o4 m mmum «4mo.mm mwNmo. mo. mH.I mob nu mo.n 6H.u Hon H mmum Hmmmv COHCD «3mm.Nv mmth. 5N. um. um N mmum 34mm.va NBHNH. mo.n mN. moo um. pm. How H mmum Hmmmv uMOwwm «4vm.mH NmHmo. mH.I NN.I Ud N mmum ««mm.mm mmmmo. mo. mH.I moo u: mH.u m~.n Hon H mmum Annoy GHmuum HMUHmoHocommm m AN mm m m Commucm anmHum> sz mHQMHum> mew m uswvawmwo AcmscHucoov NH mHnme 119 ables. In one case--job status--the activity control scale explained more variance than did the JDS and JCI au— tonomy scales combined. In sum, the new personal control scales--activity con- trol and perceived influence--appear to be psychometrically sound. The activity control scale demonstrated a high in- ternal consistency reliability, as well as a certain degree of convergent and discriminant validity. An examination of this scale's item-scale correlations and its pattern of correlations with the other personal control scales and ex- ternal variables revealed a strong relationship with the two autonomy scales--JDS autonomy and JCI autonomy. This indicates that the activity control scale is tapping the same construct as the autonomy scales. The activity con- trol scale, however, is not redundant with the autonomy scales in terms of explaining the variance of external variables. In fact, the activity control scale was able to explain a significant amount of additional variance in a set of relevant organizational behavior variables beyond that accounted for by the two autonomy scales. Thus, the activity control scale appeared to be a sound measure and was used in the subsequent analyses as the measure of ac- tivity control. The perceived influence scale also had a high degree of internal consistency reliability, however, it demon- strated little convergence with the other participation scale (i.e. Vroom participation). An examination of item- 120 scale correlations and correlations with the other personal control scales and a set of external variables revealed that the perceived influence scale and the Vroom scale were quite different measures. Considerable evidence sug- gests that the perceived influence scale is the better measure of the influence/participation construct than the Vroom scale. Of the personal control scales, the Vroom participa- tion scale had the least similar pattern of relationships with the external variables. This indicates that the Vroom scale is less related to the other personal control scales than perceived influence and may be measuring some different construct. An examination of the item-scale correlations of the personal control scales revealed that one of the outcome control scale items (i.e. CC 10) had its highest correlation with the Vroom scale. This pro- vides additional evidence that the Vroom scale may actually be assessing one's relationship with his or her supervisor rather than one's perception of influence in decison- making operations. A further indication of the usefulness of the per- ceived influence scale was revealed in a series of hier- archical multiple regression analyses. The perceived in- fluence scale eXplained a significant amount of additional variance (p<:.01) in all but two of the dependent variables beyond that accounted for by the Vroom scale. Further, the perceived influence scale accounted for more variance in a 121 mnMxnrof dependent variables (i.e. job status, job involve- ment, physical strain, effort, and union attitudes) than did the Vroom scale. This is especially significant since the Vroom scale entered the regression equation first and picked up the common variance shared between the dependent variable and both the Vroom scale and the perceived influ- ence scale. For these reasons, the perceived influence scale was used as the measure of perceived influence in the subsequent analyses. The preceding analyses also provided evidence of the relationships between the underlying constructs of the per- sonal control scales. A major hypothesis of this study is that the three variables--autonomy, expectancy of control, and perceived influence--are nuch more similar than one muld expect, given the different theoretical orientations and practical applications associated with each. It has been proposed in this study that what unifies these three vari- ables is that each is related to one's perception ofpersonal control in an organization. An examination of the personal control scale intercorrelations and their pattern of corre- lations with a set of relevant organizational behavior variables provides support for this hypothesis. The moderate to high intercorrelations among the scales representing different personal control dimensions suggest highly similar yet dis- tinct underlying constructs. Further, the similarity among the pattern of correlations between the personal control scales and a set of relevant organizational behavior vari- 122 ables provides additional evidence of the similarity among these constructs. Test of Hypothesized Relationships Between Personal Control Dimensions and Antecedent and Outcome Variables In an attempt to increase our understanding of per- sonal control, a literature review was performed to identi- fy the important antecedents and outcomes of each dimension of control--outcome control, activity control and perceived influence. The hypothesized relationships between the con- trol dimensions and the antecedent and outcome variables were summarized in Table 13. Correlations were computed to test these hypothesized relationships . Because of missing data on various items, these correlations were based on a sample of 983 respondents. Table 13 includes the correlations obtained between the personal control scales and each external variable, as well as the hypothesized relationships between those exter- nal variables and each personal control dimension. A "+" indicates that a positive relationship between that person- al control dimension and the external variable has been hy- pothesized. A "-" indicates a negative relationship, and an "NR" signifies that no relationship was believed to exist. Mood was hypothesized to have a negative relationship with each of the personal control dimensions. The results indicated that mood did indeed have a significant negative relationship with each of the personal control variables (p<’.01), with outcome control having the strongest rela- 123 .Hm>mH Ho. may um ucmoHMHcmHm mum mo. swap nmummum mcoHumenHou .Hm>mH mo. man no unmoHMHc ImHm mum mo. swap umvmmum mcoHumHmuuou .pmem on ©m>mHHma mH mHamcoHuMHmu on umau mmHMHommm :mz: cam .mHamcoHpmeH m>Hummmc m mwumoHUaH :1: m .mHam ICOHUMHCH m>HuHmom m mwumoHCCH =+= 4 .mwmmaucmnmm CH mum mmHQMHum> Hmcuwuxw man cam concmEHo Houucoo Hmcomuma aomm :mw3uma meamcoHumHmH cmNHmmauomhm "wuoz mm.u any a~.u xmzv mm.u Hmzv coHuamuaH Am>oaune mH.I HIV mN.I Hmzv HN.I HIV mopsuHuua COHCD mH.u HIV mN.u any Hm.| any chuum HMOHmoHoaUMmm mm.. 1.2 Hm.u any mH.- Hue aHmuum Hmonmam we. Hmzv mm. A+v om. A+v buommm ow. A+v mm. H+v om. H+v ucmEuHEsoo HMCOHumNHcmmuo we. Hmzv ow. A+V mm. A+v ucmEm>Ho>cH aoo Ha. H+v we. Amzv Nm. H+v coauommmHumm oncHHuxm 84. Amzv mm. A+v mm. Amzo :OHuommeumm UHmHHHHnFHH am.u luv Ba.u Any om.u luv maumum non aH.| Any om.u All n~.u Ase Houuaoo mo msooq oH.u HIV ~.mm.u HIV mN.n mane poo: mocmsHmcH cm>Hmoumm Houucou muH>Huom Honucou mEouudO mconcmEHo Houucou HMCOmHmm mmHQMHHm> wEoouso Cam ucmomumuaa paw mCOHmcmEHQ Houuaoo HMCOmHmm ammzumm mcoHumHmHuou HMOHHHQEm Cam mQHamaoHuMHmm UmNHmmapomhm NH mHQMB 124 tionship. A negative relationship was also hypothesized between locus of control and each of the personal control scales. The correlations between locus of control and the personal scales were negative and significant (p<:.01). Thus, depressed mood and an external locus of control were negatively associated with individuals' perception of con- trol at work. The final antecedent variable--job status--was also proposed to be negatively related to the personal control scales. Job status had been coded l for faculty member and 2 for clerical worker, so a high score on job status is in- dicative of a lower control position. The results indicated that each of the personal control scales was negatively re- lated to job status. Activity control and perceived influ- ence had eSpecially high correlations with job status. The high correlations between the personal control scales and job status also provide some evidence of the con- struct validity of these scales. According to Nunnally (1967), an important source of proof of the construct valid- ity of a measure is the extent to which the measure "be- haves as expected." Nunnally, (1967) has described an ex- ample of how a measure should behave as expected: "If, for example, a particular measure is thought to relate to the construct of anxiety, common sense would suggest many find- ings that should be obtained with the measure. Higher scores (higher anxiety) should be found for: (1) patients classified as anxiety neurotics than for unselected nonpa- 125 tients, (2) subjects in an experiment who are kept threaten- ed with a painful electric shock than for subjects not so threatened, and (3) graduate students waiting to undergo a final oral examination for the Ph.D. than for the same stu- dents after passing the examination" (p. 92). A measure of personal control in organization should be capable of distinguishing between incumbents in low con- trol versus high control jobs. The populations chosen for this study were faculty and clerical staff members at a large midwestern university. In general, faculty members have a great deal of freedom and control over many aspects of their job. Clerical positions, in contrast, are charac- terized by strict work rules, close supervision, and little personal discretion in how the job is performed. The high correlations between the personal control scales and job status demonstrate the ability of the personal scales to distinguish between members of a high control versus a low control position in an organization, thus providing some evidence of the construct validity of these scales. Intrinsic satisfaction was hypothesized to be positive- ly related to activity control and to be unrelated to both outcome control and perceived influence. Activity control had the highest correlation with intrinsic satisfaction (r.= 56), however, its correlations with outcome control and perceived influence were also high (r.s = .55 and .44). Extrinsic satisfaction was proposed to be related to out- come control and perceived influence and unrelated to acti- 126 vity control. The results indicated that extrinsic satis- faction had a high positive relationship with all three scales. Thus, all three personal control scales had high positive relationships with both intrinsic and extrinsic satisfaction, supporting several hypothesized relationships. Significant correlations were found between satisfac- tion variables and personal control scales that were hypo- thesized to be unrelated (e.g. intrinsic satisfaction and outcome control). This can be partly explained by the high correlation (.64) between intrinsic satisfaction and extrin- sic satisfaction. Also, it is likely that the effects of common method variance acted to inflate the correlatiOns between these variables. A positive relationship was hypothesized between job involvement and both outcome control and activity control, while no relationship was believed to exist between job in- volvement and perceived influence. The results indicated, however, that perceived influence had the highest correla- tion with job involvement (r. = .44). Once again, variables proposed to be unrelated turned out to be highly related. Activity control and outcome control also had high positive correlations with job involvement, supporting their hypo— thesized relationships. Organizational commitment was hypothesized to be posi- tively related to all three personal control scales. The results supported these hypothesized relationships. Effort, however, had its highest correlation with perceived influ- 127 ence, to which it was hypothesized to be unrelated. Effort was positively correlated to both outcome control and per- ceived influence, as hypothesized. Physical and psychological strain were hypothesized to be negatively related to all three personal control dimen- sions. The results supported these hypotheses, although the magnitude of the correlations were lower (althoughstill significant--p<:.01) than the correlation with the other outcome variables. Union attitudes was hypothesized to be negatively re- lated to both outcome control and perceived influence and unrelated to activity control. Turnover intention was pro- posed to be negatively related to perceived influence and unrelated to both outcome control and activity control. Significant negative correlations were found between the three personal control scales and both union attitudes and turnover intention, once again supporting the hypothesized relationships and also finding relationships where none were believed to exist. In sum, every hypothesized relationship between the personal control scales and the antecedent and outcomexmri- ables was confirmed. However, significant correlations were found between personal control and those variables that were be- lieved to be unrelated on a priori conceptual bases . In fact, every correlation between a personal control scale and an antecedent or outcome variable was significantly greater than zero (p < . 01) . labile these significant correlations provide sorre evidence of the im- 128 portance of the personal control scales, it also signals the presence of common method bias. Common method variance was also indicated by the high correlations among the out- come variables. Each of the three personal control scales had their highest correlations with intrinsic satisfaction, extrinsic satisfaction, job involvement and organizational commit- ment. Thus, the dimensions of personal control appear to be most highly related to one's satisfaction and identity with and commitment to his or her job and organization. Test of the Personal Control Model This paper presented a model of personal control of the form x —9 m —9 y, whereby the "x" represents the ante- cedent variables (i.e. job status, mood, and locus of con- trol), the "m" symbolizes the personal control dimensions (i.e. outcome control, activity control, and perceived in- fluence), and the "y" represents the outcome variables (i.e. intrinsic satisfaction, extrinsic satisfaction, job involvement, organizational commitment, physical strain, psychological strain, effort, union attitudes, and turn- over intention). This mediational model proposes that the antecedent variables transmit their effects on the outcome variables through the personal control variables. If the mediational model of personal control is cor- rect as hypothesized, then the relationship between the antecedent variables and the outcome variables should vanish if the personal control variables are held constant 129 (James and Brett, 1984). This model was tested using a series of hierarchical multiple regressions, whereby each outcome variable was regressed on the set of personal con- trol variables and the set of antecedent variables. In these analyses the personal control variables were entered into the regression equation,first and the antecedent vari— ables were entered second. To test the mediating hypothe- sis, this hierarchical regression was compared with one in which the antecedent variables were entered first and the personal control variables were entered second. If the mediational hypotheses are supported, we would not expect the antecedent variables to add significantly to the re- gression analyses in which they are entered after the per- sonal control variables. However, when antecedent vari- ables are entered first, we would expect significant addi- tional variance accounted for when the personal control variables are added. It was anticipated that one of the antecedent variables (i.e. job status) might have been confounded by two demo- graphic variables--sex and educational level. Clerical workers in this sample were predominantly female (97%), while the faculty members were mostly male (81%). In re- gards to educational level, 97% of the faculty members had a graduate degree, while over 70% of the clerical workers did not even have a bachelor's degree. Although these findings are not surprising and are actually quite typical of these occupational groups, it was felt that the effects 130 of these variables should be controlled in the regression analyses. Therefore, the demographic variables--sex and educational level-—were entered into the regression equa- tion first, the personal control variables were then entered second, and the antecedent variables were entered last. Table 14 summarizes the results of the hierarchical multiple regression analyses. Because of missing data and the use of listwise deletion, the analyses were based on samples ranging from 758 to 777. The results indicate that the antecedent variables explained a significant amount of additional variance (p<:.01) beyond that explained by the personal control scales. These results provide only partial support for the mediational model of personal control for satisfaction, job involvement, organizational commitment and turnover intention, which proposed that the antecedent vari- ables were related to the outcome variables only through their effects on the personal control variables. This clearly is not the case since, with the personal control variables controlled, the antecedent variables were signifi- cantly related to the outcome variables. With very large sample sizes, however, small increments in R2 are significant. Given the sample size in the pre- sent study, an increase in R2 of only .003 would be signifi- cant (p<:.05). If one compares the change in R2 for ante- cedent variables when they are entered second with similar changes when they enter the regression first, one observes a pattern of relationships consistent with the mediation 131 .HHmumU once CH mConmmummu mmmap qummum aoHa3 .MN ou mH meama Eoum Cwuomuuxm mum meHm> wmmaa .meQMHum> HouuCoo HCCOmuwm Na pm30HH0m umuHm Conmmummu may Commqu mmHamHum> quCmoqum Cmas omuusooo aoHa3 Nm CH mmmCmao mum mommaquumm CH mmCHm> .mmmmHmCC mmmau CH UmHHouuCoo wHHCOHpmHumum mumzluHm>mH HCCoHumoCCm CCC xmmuummHamHnm> oHammum 10586 was mau Ho mmsmowm .mmmmaquumm map CH owumoHCCH mNHm mHmEmm may Co Comma mm3 mHmem ICC Conmmummu wHQHbHCE map .CoHumHmp mmH3umHH mo mm: may CCC mumc mCHmmHE .HHo.uvmv uCCOHMHCmHm mum mHamH mHau CH mmCHm> m mau mo HHm “wuoz mm.HH Ammo.v mmo. no.mm HNHH.V mmH. Hubby CoHquuCH uw>0CHCB mH.m Ammo.v Hmo. mm.MH HHHo.C «we. lemme mmasuHuua coHca HH.4¢ HHMH.V Hmo. mm.mm HHHo.v Hmo. Hmmsv uuowmm mm.~vH HmHH.V «Hm. mm.em Ammo.v mmH. Ammac chuum HmonoHoaommm «H.~H HmmH.c CNH. mm.mm 1800.2 moH. Amuse .aHmuum Hmonsam om.mH HHMH.V ovo. 60.04H qum.v mmm. AHmHV uzmeuHaaoo HMCOHumNHCmmHO mH.om AHOH.V who. mm.om Awmo.o mmH. HHBHV pamem>Ho>cH nos mm.om AmmH.o mmo. o~.HMH AQHN.V mom. AHHHV aoHuommmHumm onaHuuxm mH.q~ AmmH.v mac. mm.mmH AmH~.c 86m. AHHHV aoHuommmHumm onaHuuaH MMImW. meamHum> MMImW_ mmHamHum> Houu Hmfl mmHamHum> quUCmmwo How m quUmomuC¢ How m ICoo HMComuwm Mom Nm Q How Nm AV mmHamHum> mEoouCO CCm quUmumuca .HouuCou HMCOmuwm .oHammumoEmo wau mo mmmeMCa COHmmmummm mHmHuHCz HCUHaonumHm may no muHCmmm vH mHQMB 132 hypothesis. That is, while all antecedent variables are significantly related independently of other variables, personal control variables may be mediating part of the re- lationship. For example, results for intrinsic satisfaction indicate that change in R2 due to antecedent variables was only .046 when personal control variables were controlled. Reversing the order of entry, we see that antecedent vari- ables account for much greater portions of variance. On the other hand, personal control explained substan- tial amounts of additional variance in several of the de- pendent variables (i.e. satisfaction, job involvement, or- ganizational commitment and turnover intention), with the antecedent variables controlled. The personal control vari- ables accounted for an additional 21% of variance in satis- faction, 24% in organizational commitment, 11% in turnover intention and 8% in job involvement beyond that explained by the antecedent variables. This provides additional sup- port for the mediational model of personal control, at least for this reduced set of dependent variables. A possible explanation for the somewhat negative re- sults is common method variance. As was discussed earlier, the correlations among the variables used in this study may have been inflated because they were measured with similar instruments on the same questionnaire. The correlations between the antecedent and outcome variables, therefore, may be due in part to common method variance. In Tables 15 to 23, the results of the hierarchical re- 133 gressions testing the unique relationship of personal con- trol to various dependent measures are presented. In these regressions, the demographic variables were entered into the regression equations first, the antecedent variables second, and the personal control variables were entered last. These analyses assessed the degree to which the per- sonal control variables eXplained additional variance in the outcome variables beyond that accounted for by the demo- graphic and antecedent variables. Tables 15 to 23 contain the simple correlations,stand- ardized regression coefficients, and the multiple squared correlations for each independent variable in the regression analyses. F-tests for the change in multiple R2 caused by the regression equation are also reported. Table 15 shows the results of the hierarchical regres- sion analyses of the demographic, antecedent and personal control variables on intrinsic satisfaction. The analysis was based on a sample of 774 respondents because of missing data and the use of listwise deletion. The multiple squared correlation of the demographic, antecedent and personal con- trol variables on intrinsic satisfaction was .514. Of the demographic variables, only sex was significantly related to intrinsic satisfaction. All of the antecedent variables accounted for significant amounts of additional variance in intrinsic satisfaction beyond that accounted for by the demographic variables. More importantly, each of the per- sonal control scales explained a significant amount of addi- Results of Hierarchical Regression Analysis of Demographic, Antecedent and Personal Control Step 1. Step 2. Step 3. 134 Table 15 Variables on Intrinsic Satisfactiona Variable Entered Demographic Sexb Education Level Antecedent Job StatusC Mood Locus of Control Personal Control Outcome Control Activity Control Perceived Influence IH .19 .56 .57 .44 Iw .02 .17 .20 .18 .07 .30 .30 .08 R2 .103 .103 .159 .267 .301 .440 .510 .514 F for A122 162.05** .55 87.19** 170.11** 53.72** 218.37** 111.46** 5.94* a. Because of missing data and the use of listwise deletion, the multiple regression analysis was based on a sample of 774. b. Sex is coded l for male and 2 for female. c. Job status is coded l for faculty member and 2 for clerical worker. ** p <1.01 * p <:.05 Results of Hierarchical Regression Analysis of Demographic, Antecedent and Personal Control Step 1. Step 2. Step 3. 135 Table 16 Variables on Extrinsic Satisfactiona Variable Entered Demographic Sex Education Level Antecedent Job Status Mood Locus of Control Personal Control Outcome Control Activity Control Perceived Influence l H -.22 .11 -.20 -.41 -.20 .51 .45 .42 IUJ -.09 -.06 .12 -.27 .01 .28 .22 .20 R2 .046 .048 .054 .205 .216 .355 .403 .426 61.72** 2.31 8.05** 200.29** 15.32** 184.04** 63.41** 31.86** a. Because of the missing data and the use of listwise deletion, the multiple regression analysis was based on a sample of 771. Results of Hierarchical Regression Analysis of Demographic, Antecedent and Personal Control Variables on Job Involvementa 136 Table 17 Variable F for Entered r B 33 .A R Step 1. Demographic Sex -.45 .04 .203 260.73** Education Level .40 -.09 .222 24.40** Step 2. Antecedent Job Status -.57 -.53 .326 132.77** Mood -.12 .01 .329 3.79 Locus of Control -.08 .06 .329 .38 Step 3. Personal Control Outcome Control .40 .24 .401 91.50** Activity Control .41 .10 .410 12.00** Perceived Influence .45 .08 .413 4.34* a. Because of the missing data and the use of listwise deletion, the multiple regression analysis was based on a sample of 761. 137 Table 18 Results of Hierarchical Regression Analysis of Demographic, Antecedent and Personal Control Variables on Organizational Commitmenta Variable Entered Step 1. Demographic Sex Education Level Step 2. Antecedent Job Status Mood Locus of Control Step 3. Personal Control Outcome Control Perceived Influence Activity Control .1: E -.12 .00 .05 -.06 -.12 .14 -.34 -.20 -.22 -.02 .52 .33 .41 .26 .39 .16 R2 .015 .016 .022 .128 .150 .324 .379 .395 18.75** 1.38 7.83** l3l.69** 26.96** 215.76** 68.44** 20.28** Because of the missing data and the use of listwise deletion, the multiple regression analysis was based on a sample of 761. p‘< .01 :><;.05 138 Table 19 Results of Hierarchical Regression Analysis of Demographic, Antecedent and Personal Control Variables on Physical Straina Variable Entered Step 1. Demographic Sex Education Level Step 2. Antecedent Job Status Mood Locus of Control Step 3. Personal Control Activity Control Outcome Control Perceived Influence _r_ E .26 .08 -.22 -.08 .26 .05 .41 .35 .19 .09 -.29 -.10 -.19 .02 -.21 -.02 R2 .066 .071 .076 .224 .234 .242 .242 .243 F for ARZ 67.07** 5.20* 4.95* 149.59** 10.29** 8.38** .10 .13 ** * Because of the missing data and the use of listwise deletion, the multiple regression analysis was based on a sample of 776. p <1.01 p4<..05 Results of Hierarchical Regression Analysis of Demographic, Antecedent and Personal Control Variables on Psychological Strain 139 Table 20 a deletion, on a sample of 763. Variable F for Entered 1: _B 33 A R2 Step 1. Demographic Sex .07 .02 .005 6 86** Education Level -.03 .02 .006 91 Step 2. Antecedent Job Status .06 .07 .006 .94 Mood .63 .57 .400 536.38** Locus of Control .26 .11 .421 28.37** Step 3. Personal Control Outcome Control -.31 .12 .438 23.33** Activity Control -.27 .11 .446 9.73** Perceived Influence -.13 .02 .446 .33 a. Because of the missing data and the use of listwise the multiple regression analysis was based 140 Table 21 Results of Hierarchical Regression Analysis of Demographic, Antecedent and Personal Control Variables on Efforta Variable F for Entered 5 B 3% A R2 Step 1. Demographic Sex -.51 .05 .264 364.50** Education Level .49 -.04 .308 59.86** Step 2. Antecedent Job Status -.66 -.66 .432 l70.44** Mood -.07 .04 .432 0.00 Locus of Control -.14 -.06 .439 9.94** Step 3. Personal Control Outcome Control .29 .10 .452 l9.01** Perceived Influence .44 .08 .456 ll.30** Activity Control .35 .00 .456 0.00 a. ,Because of the missing data and the use of listwise deletion, the multiple regression analysis was based on a sample of 758. ** p4.01 * p<..05 Results of Hierarchical Regression Analysis of Demographic, Antecedent and Personal Control Variables on Union Attitudesa Step 1. Step 2. Step 3. 141 Table 22. Variable Entered Demographic Sex Education Level Antecedent Job Status Mood Locus of Control Personal Control Outcome Control Activity Control Perceived Influence IH -.25 -.24 .31 .16 .17 -.25 [(11 -.02 -.03 .25 .07 .09 -.12 -.07 .04 .062 .072 .095 .112 .127 .141 .143 .144 53.91** 9.52** 20.42** 14.56** 12.90** 12.02** 2.02 .94 Because of the missing deletion, the multiple regression analysis was based on a sample of 757. data and the use of listwise Results of Hierarchical Regression Analysis of Demographic, Antecedent and Personal Control Variables on Turnover Intentiona Step 1. Step 2. Step 3. 142 Tabley23_ Variable Entered Demographic Sex Education Level Antecedent Job Status Mood Locus of Control Personal Control Outcome Control Perceived Influence Activity Control IH -.10 -.03 -.11 -.27 -.17 .39 .26 .26 ICD -.02 -.24 -.10 -.16 -.03 .25 .13 .08 R2 .010 .027 .047 .108 .112 .206 .220 .224 F for ARZ 10.41** 16.25** 19.80** 60.63** 11.74** 84.84** 14.65** 4.05* a. -Because of the missing data and the use of listwise deletion, the multiple regression analysis was based on a sample of 777. 143 tional variance in intrinsic satisfaction beyond that ac- counted for by the demographic and antecedent variables. Table 16 shows the results of the multiple regression analysis of the demographic, antecedent and personal con- trol scales on extrinsic satisfaction. The results are very similar to those involving intrinsic satisfaction, al- though the R2 was somewhat lower (i.e. .412). Each of the personal control scales accounted for significant amounts of additional variance in the dependent variable (i.e. ex- trinsic satisfaction) beyond that explained by the demo- graphic and antecedent variables. These results provide evidence of the importance of the personal control scales in terms of their explanatory power of the satisfaction variables. The multiple R2 of the antecedent, demographic and personal control variables on job involvement was .413. Both demographc variables were significant predictors, while job status was the only significant predictor among the antecedent variables. Once again, all three personal control scales explained a significant amount of additional variance of the dependent variable beyond that explained by the demographic and antecedent variables. The regression analyses with the other dependent vari- ables provided additional evidence of the explanatory power of the personal control scales (see Tables 17 through 23). Each of the personal control scales explained significant increments of explained variance in organizational commit- 144 ment. Only activity control, however, was a significant predictor of physical strain. Both activity control and outcome control accounted for a significant amount of addi- tional variance in psychological strain beyond that ex- plained by the demographic and antecedent variables. Out- come control and perceived influence explained a signifi- cant amount of additional variance in effort and turnover intentions, while only outcome control was a significant predictor of union attitudes. In sum, the regression analyses provided evidence of the importance of outcome control, activity control and perceived influence as predictors of important organiza- tional behavior variables (i.e. the outcome variables). Further, including more than one personal control scale in the regression equations increased the amount of variance explained in many of the dependent variables: intrinsic satisfaction, extrinsic satisfaction, job involvement, or- ganizational commitment, psychological strain, effort and turnover intentions. Thus, the three-dimensional concep- tualization of personal control provides greater explana- tory power of important organizational behavior variables than any of the personal control scales alone. CHAPTER V SUMMARY AND CONCLUSIONS Summary and Conclusions Theorists and researchers have long proposed that per- sonal control is an important human need. Laboratory re- searchers have demonstrated the negative effects of lack of control and the positive effects of perceived control over aversive environmental events. The purpose of this study was to increase our understanding of the construct of per- sonal control in organizations. Personal control was defined as one's perception of freedom in and control over work activities, events and out- comes. Bazerman's (1982) multidimensional conceptualiza- tion of personal control was used to provide a framework for the operationalization of personal control in this study. It was necessary, however, to include the construct of perceived influence along with Bazerman's two dimensions of control to conform to this study's definition of per- sonal control derived from the work of psychological theor- ists (i.e. Tannenbaum, 1962; White, 1959) and laboratory researchers (e.g. Glass and Singer, 1972). Thus, personal control was defined as consisting of the following dimen- sions: outcome control, activity control and perceived influence. Although each of these constructs of personal control has been operationalized and studied by organizational be- havior researchers, no researcher has studied more than one 145 146 dimension of control at the same time. Thus, our under- standing of the interrelationships among the different di- mensions of control is limited. Further, researchers studying these constructs have not explicitly related them to personal control. Expectancy of control has been studied in the context of motivation theory, autonomy as a job characteristic, and perceived influence in regards to participative management. The present study prOposed that these three constructs are much more similar than one would expect, given the different theoretical orientations and practical applications associated with each. It was fur- ther proposed that what unifies these three constructs is their relationship to one's perception of control. A review of the research literature involving the three constructs of control revealed how each has been 0p- erationalized and the important antecedents and outcome variables associated with each. A mediational model of personal control was developed based on the antecedent and outcome variables identified in the literature review. Further, the relationship between the personal control di- mensions and each antecedent and outcome variable was hy- pothesized on the basis of the research literature. The three dimensions of control were Operationalized in the present study using existing scales and newly de- veloped scales to test the multidimensional conceptualiza- tion of personal control and the mediational model of con- trol in a field setting. 147 Questionnaires were mailed to 1,768 faculty and 1,624 clerical staff members of a large midwestern university. Usable questionnaires were returned by 423 faculty members and 655 clerical workers, yielding response rates of 24 and 40 percent reSpectively. A comparison of the sample and population demographic characteristics indicated that the sample was representative of the population from which it was drawn. The results indicated that the new personal control scales had a high degree of internal consistency, as indi- cated by their coefficient alphas and item-scale correla- tions. Each of the new personal control scales had high correlations with their respective existing scales, indi- cating a certain degree of convergent validity. Further, the new and existing personal control scales had similar patterns of correlations with the antecedent and outcome variables. Thus, it was concluded that the new personal scales were comparable to the existing scales. The new scales were also sufficiently distinctive from the existing scales to demonstrate their practical value. In a series of hierarchical regression analyses in which the existing scale was entered into the regression equation first and the new scale entered second, the new scale ex- plained a significant amount of additional variance in each of the antecedent and outcome variables beyond that ex- plained by the existing scale. The results supported the multidimensional conceptual- 148 ization of personal control. An examination of the personal control scale intercorrelations and their pattern of cor- relationswith the antecedent and outcome variables indicated that the personal control dimensions are indeed highly re- lated but distinct variables. Further, all of the hypothesized relationships between the personal control scales and the antecedent and outcome variables were confirmed. However, variables that were theorized to be unrelated were also significantly cor- related. In fact, each personal control scale was signifi- cantly correlated with all of the antecedent and outcome variables. It is likely that method variance was somewhat responsible for these higher-than-expected correlations. The results, however, did not fully support the media- tional model of personal contorl. Although the antecedent variables were significantly related to the mediating vari- ables (i.e. personal control) and the mediating variables were significantly related to the outcome variables, the antecedent variables were significantly related to the out- come variables with the mediating variables controlled. In other words, the antecedent variables were significantly re- lated to the outcome variables independent of the mediating variables. This does not support a complete mediational model which proposes that the antecedent variables affect the outcome variables completely through its effects on the mediating variables. While the results did not support a complete media- 149 tional model of personal control, there was support for a partial mediational model. Given the large sample size on which these analyses were computed, even small changes in R2 are significant. A comparison of the change in R2 for the antecedent variables when they are entered second with similar changes when they enter the regression equations first, reveals a pattern of relationships consistent with a partial mediational model. For example, with the~penxnal control variables controlled, the.antecedent variable860nly accounted for 4.6% of the additional variance in intrinsic satisfaction, 6.8% in extrinsic satisfaction, 4% in organi- zational commitment, 7.2% in job involvement and 3.5% in turnover intentions. These are substantially lower changes in R2 as compared to similar changes in R2 when the antece- dent variables enter the regression equation first. Further, the personal control variables explained sub- stantial amounts of additional varianCe in several of the dependent variables with the antecedent variables con- trolled--24% of the variance in organizational commitment, 21% in satisfaction, 11% in turnover intention and 8% in job involvement. This provides some support for the media- tional model of personal control, at least for this reduced set of dependent variables. The results also provided evidence of the importance of outcome control, activity control and perceived influ- ence as predictors of important organizational behavior variables (i.e. the outcome variables). Further, including 150 more than one personal control scale as predictors of the outcome variables significantly increased the amount of variance explained in most of the outcome variables: in- trinsic satisfaction, extrinsic satisfaction, job involve- ment, organizational commitment, psychological strain, effort and turnover intentions. Thus, the three-dimensional conceptualization of personal control provides greater ex- planatory power of the antecedent and outcome variables than any of the personal control scales alone. The Personal Control Model This study empirically tested a mediational model of a multidimensional conceptualization of personal control in a field setting. While the results did not fully support a mediational model of personal control, evidence of partial mediation was found. There are several possible explana- tions for the somewhat negative results. First, it is possible that common method variance was somewhat responsi- ble for the higher-than-expected correlations between the antecedent and outcome variables. Unfortunately, method variance cannot be statistically controlled in this study because of the methodology used to collect the data. The specific antecedent variables utilized in this study may also have been somewhat responsible for the par- tial failure of the mediational model. The antecedent vari- able, job status, although not susceptible to method vari- ance, may have been both deficient and contaminated as an indicator of the level of control in a job. Job status was 151 selected as an antecedent variable because it was believed that a vast difference exists between the amount of control provided incumbents in faculty versus clerical positions. It was also assumed that faculty members would have higher levels of all three dimensions of control as compared to clerical workers. The results indicated that job status was highly cor- related with activity control (r. = -.47) and perceived in- fluence (r. = -.54) and much less related to outcome con- trol (r. = -.20). Apparently, faculty and clerical workers did not differ as much in their perceptions of outcome con- trol as they did with activity control and perceived influ- ence. It would have been preferable for job status, as an antecedent variable, to be highly correlated with 211 three dimensions of personal control. Of greater concern is the possible contamination of job status with extraneous factors (i.e. unrelated to per- sonal control) that were related to the outcome variables. Job status was very highly related to certain demographic characteristics--sex (r. = .79) and educational level (r. = -.80). Although these variables were statistically controlled in the analyses, faculty and clerical workers may also differ on factors that were not measured and sub- sequently not controlled. For example, faculty earn higher salaries and enjoy greater prestige in their jobs than do clerical workers. These factors, while unrelated to per- sonal control, would likely impact on some of the outcome 152 variables (e.g. job satisfaction). The results indicated that job status had high correlations with several outcome variables: effort (r. = -.67), job involvement (r. = -.56), and intrinsic satisfaction (r. = -.37). Future research needs to identify specific antecedents that directly im- pact on each of these dimensions of personal control. On a more positive note, the results of this study supported a multidimensional conceptualization of the con- struct of personal control. Further, the construct of per- sonal control provides an integration of aSpects of three separate research literatures--expectancy theory, job design and participative decision making. These three variables-- outcome control, activity control and perceived influence-- are clearly highly related yet distinct constructs. The importance of personal control in organizatins was also demonstrated in this study in terms of the high cor- relations found between the personal control scales and a set of important organizational behavior variables (i.e. the outcome variables). Further, the multidimensional conceptu- alization of personal control explained more variance in the outcome variables than any of the personal control scales alone. Researchers investigating any of these individual dimensions of control would be advised to measure all three constructs for a more complete understanding of the phenom- ena at hand. 153 Limitations of Study The results obtained in this study, however, are limited in several important ways. The two samples em- ployed in this study--faculty and clerical workers-- dif- fered in ways other than their occupationsl status. The faculty members were predominantly male, while the clerical workers were mostly female. Also, the faculty members were much more highly educated than the clerical workers. Al- though the analyses statistically controlled for the effects of sex and educational level, the results may be somewhat limited to samples of male faculty and female clerical workers. Further, the study was conducted in a non-profit, state-run, educational organization. Thus, the generaliza- tion of these results to private-sector, non-educational organizations should be done with caution. A more serious problem was the presence of common method variance. It was apparent that a large general fac- tor existed, affecting both the independent and dependent variables. This common factor was likely a method bias and provided an alternative explanation of the relationships between the variables studied. Thus, it cannot be concluded with certainty that Specific variables are related, since the relationship could be due, at least in part, to common method variance. 154 Future Research The present model of personal control is limited by its assumption that personal contorl is a basic human need desired by all. It is likely that not everyone desires or is capable of using greater amounts of control at work. Schneider, Reichers and Mitchell (1982) have warned that attempts at job enrichment (i.e. increased autonomy) may fail because these changes can increase the requirements of the job beyond the job incumbent's ability. Similar problems are likely to result from programs designed to in- crease employees' participation in organizational decision- making operations. Increased personal control would likely be associated with greater levels of responsibility,and in- creased responsibility at work would not be welcomed by all. Bazerman (1982) prOposed that the Optimal control state is one in which an individual's ability to use con- trol is congruent with the amount of control provided him or her by the organization. In a laboratory study using college students, Bazerman found that performance was higher in the congruent condition than either the under- control or over-control conditions. Future research should examine this congruence model in a field setting. An important issue that warrants a great deal of fu- ture research involves the change in one's level of per- sonal control. In a series of studies, Brehm (1966, 1972) found that eXperimental participants reacted very nega- 155 tively (i.e. reactance) when their choice of rewards for participating in the study was restricted. Thus, loss of control over important work outcomes could produce serious problems to an organization. Future research should in- vestigate the range of negative outcomes that might result from workers' loss of personal control. Of particular im- portance in this line of research is the relationship be- tween loss of personal control and stress reactions in or- ganizational members. Finally, future research should address the issue of how to facilitate organizational change with a minimum of disruption to its organizational members. It is likely that personal control would be a central variable in this line of research. This line of research would be especial- ly important, given the current state of rapid organiza- tional changes in American companies brought about by in- creased foreign competition and continually evolving tech- nologies. Practical Implications The measures of personal control developed for this study have practical value for organizations in several functions. The primary use of these instruments is as a diagnostic tool to assess individuals' perceptions of con- trol or influence over important aspects of their worklife. The present study has demonstrated the importance of per- ception of personal control in terms of its relationship 1156 with other variables (e.g. satisfaction and reduced strain). Low scores on these scales may signal the need for improvement in some aspect of the organization. For ex- ample, low levels of activity control suggest a need for job design. The perceived influence scale can identify a need for a more participative style of management, while scores on the outcome control scale have implications for the organization's reward and control systems. Further, the individual items in each of the scales can be used separately as a single-item measure of control or influ- ence over specific aspects of work (e.g. job security, work deadlines or promotions). The results obtained from the individual items can allow practitioners to focus their change programs on the specific deficient area. The personal control scales might also be used as part of an evaluation effort to determine the efficacy of an organizational change intervention related to job de- sign or participative decision making. Any manipulation that attempts to increase employees' participation in de- cision making or involvement in other management functions (e.g. quality circles) or enrich their jobs through job design night impact on an employee's perception of per- sonal control. Thus, personal control may be the direct result of a wide variety of organizational develOpment pro- grams. In turn, increases in personal control may influ- ence the more frequently used measures of the effective- 157 ness of these programs such as behavioral and affective reactions. Personal control, then, may serve as an impor- tant "barometer" of employees' reactions to organizational change programs, much like the function that job satis- faction measures have played in the past. APPENDICES APPENDIX A DEMOGRAPHIC ITEMS 158 Demographic Items Answer each of the following questions using the scale provided: 1. How long have you been an employee of MSU? . less than 6 months . 6 months to 1 year . l to 5 years 6 to 10 years 11 to 20 years over 20 years mmwat-t 2. Please indicate your gender. 1. male 2. female 3. Please indicate your educational status. 1. high school graduate or less 2. some college--no degree 3. two-year college degree 4. four-year college degree 5. graduate degree 4. Please indicate your job level (clerical only). 1. level level level level level level level 10 level 11 & 12 other QQQONU'IDUJN 0.. \DQOU‘lb 5. Please indicate your rank (faculty only). 1. professor 2. associate professor 3. assistant professor 4. instructor 5. other 6. Have you received tenure (faculty only)? 1. no 2. yes APPENDIX B ANTECEDENT VARIABLES 159 Profile of Mood States - Depression/Dejection Scale (McNair, Lorr and Droppleman, 1971) The list of words below describes feelings people have. Please read each item and rate how you are feeling today using the following scale: 1 2 3 4 5 not at all ‘a little moderately quite extremely l. unhappy 2. sorry 3. sad 4. blue 5. hopeless 6. unworthy 7. discouraged 8. lonely 9. miserable 10. gloomy 11. desperate 12. helpless 13. worthless l4. terrified 5. guilty 160 Locus of Control (Rotter, 1966) How much do you agree or disagree with each of the following statements? Mark your response on the answer sheet using the scale below: 10. 11. 1. strongly agree 2. agree 3. neither agree nor disagree 4. disagree 5 . strongly disagree Many of the unhappy things in people's lives are partly due to bad luck. Who gets to be boss often depends on who was lucky enough to be in the right place first. In my case, getting what I want has little or nothing to do with luck. In the long run, people get the reSpect they deserve in this world. When I make plans, I am almost certain that I can make them work. Without the right breaks, one cannot be a good leader. In the long run, the bad things that happen to us are balanced by the good ones. What happens to me is my own doing. Many times I feel that I have little influence over the things that happen to me. Most people don't realize the extent to which their lives are controlled by accidental happenings. Becoming a success is a matter of hard work; luck has little or nothing to do with it. APPENDIX C PERSONAL CONTROL MEASURES 161 OUTCOME CONTROL (Lawler, 1981)* Here are some things that could happen to people if they work hard at their job. How likely is it that each of these things would happen if you worked hard at your job? Use the follow- ing scale to answer: 1. not at all likely 5. quite likely 2. unlikely 6. very likely 3. somewhat likely 7. extremely likely 4. likely 1. You will get a bonus or pay increase. 2. You will feel better about yourself as a person. 3. You will have an opportunity to develop your skills and abilities. 4. You will have better job security. 5. You will be given chances to learn new things. 6. You will be promoted or get a better job. 7. You will get a feeling that you've accomplished something worthwhile. 8. You will have more freedom on your job. 9. You will be respected by the people you work with. 10. Your supervisor will praise you. 11. The people you work with will be friendly with you. * Instructions modified by the author. .162 Job Diagnostic Survey - Autonomy Scale (Hackman & Oldham, 1975) How much autonomy is there in your job? own how to go about doing the work? 1 . . . 2 . . . 3 . . . .4 . . . .5 . very little: the moderate autonomy; job gives me almost many things are no personal "say" standardized and about how and when not under my con- the work is done trol, but I can make some decisions about the work That is, to what extent does your job permit you to decide on your .6 . . . 7 very much; the job gives me almost complete responsibility for deciding how and when the work is done Indicate the accuracy of each of the following statements con- cerning your job using the scale below: very inaccurate mostly inaccurate slightly inaccurate uncertain slightly accurate mostly accurate . very accurate \IOAUWDOJNH The job denies me any chance to use my personal initiative or judgment in carrying out the work. The job gives a person considerable opportunity for inde- pendence and freedom in how he or she does the work. 163 Job Characteristics Inventory - Autonomy Scale (Sims, Szilagyi and Keller, 1976) Use the following scale to answer these questions: 1. . . . . 2 . . . . . . 3 . . . . .‘4 . . . . 5 very little a moderate amount very much 1. How much are you left on your own to do your own work? 2. To what extent are you able to act independently of your supervisor in performing your task? 3. To what extent are you able to do your job independently of others? Use the following scale to answer these questions: 1. O O O O O 2 O O O O I 3 O O O O 4 O O O O .5 a minimum amount a moderate amount a maximum amount 4. The freedom to do pretty much what I want on my job. 5. The opportunity for independent thought and action. 6. The control I have over the pace of my work. 164 Activity Control Rate the amount of control that you have over each of the following aspects of your job using the following scale: 1 . . . . . 2 . . . . 3 . . . . 4 . . . 5 complete control moderate control no control 1. The Speed with which you do your work. 2. The setting of work deadlines. 3. The selection of work tasks that you perform. 4. When you take your rest breaks. 5. The choice of methods to do your work. 6. The layout of your workspace. 7. The setting of performance goals. 8. The choice of equipment to do your work. 9. Determining the order in which you will do your work. 10. The specific hours you work each day. .165 Psychological Participation (Vroom, 1960) If you have a suggestion for improving the job or chang- ing the setup in some way, how easy is it for you to get your ideas across to your immediate supervisor? 1 . . . . . 2 . . . . 3 . . . . 4 . . . . . 5 very easy average very difficult Do you feel you can influence the decisions of your im- mediate supervisor regarding things about which you are concerned? 1 . . . . 2 . . . . . 3 . . . . . . 4 . . . . 5 very little a moderate amount very much Does your immediate supervisor ask your opinion when a problem comes up which involves your work? 1 . . . . 2,. . . . . 3 . . . . . . 4 . . . . 5 very little a moderate amount _ very much In general, how much say or influence do you have on what goes on in your station? 1 . . . . 2 . . . . . 3 . . . . . . 4 . . . . 5 very little a moderate amount very much 166 Perceived Influence Rate the level of your past involvement in each of the 14 de- cision areas listed below using the following rating scale: 10. 11. 12. 13. 14. l. 2. No advance information was provided to you concerning the decision. You were informed in advance of the decision to be made. You were able to voice your opinion concerning the decision. Your opinion concerning the decision was taken into account in the decision-making process. The decision was made jointly with equal authority between yourself and someone else. The decision was entirely your own with no involve- ment by anyone else. Hiring new employees. Your promotion. Your performance appraisal. Training new employees. Your pay raise. Discipline procedures. Evaluation of other personnel. Allocation of department budget. Assignment of personnel. Department layoff policy. Department policy making. Department wage level. Department promotion procedures. Department performance appraisal procedures. APPENDIX D OUTCOME VARIABLES MEASURES 167 Minnesota Satisfaction Questionnaire (Weiss, Dawis, England and Lofquist, 1967) The questions in this part ask youtx>describe your job or how you feel about your job. Use the scale below to indi- cate your answer: 1. very satisfied 2. satisfied 3. neutral 4. dissatisfied 5. very dissatisfied On my present job this is how I feel about: 1. Being able to keep busy all the time. 2. The 3. The 4. The 5. The 6. The chance to work alone on the job. chance to do different things from time to time. chance to be "somebody" in the community. way my supervisor handles his or her employees. competence of my supervisor in making decisions. 7. Being able to do things that don't go against my conscience. 8. The 9. The 10. The 11. The 12. The way my job provides for steady employment. chance to do things for other people. chance to tell people what to do. chance to do something that makes use of my abilities. way company policies are put into practice. 13. My pay and the amount of work I do. 14. The 15. The 16. The 17. The 18. The chances for advancement on this job. freedom to use my own judgment. chance to try my own methods of doing the job. working conditions. way my co-workers get along with each other. 168 Job Involvement (Kanungo, 1981) How much do you agree or disagree with each of the following statements? Mark your response on the answer sheet using the scale below: . 9. 10. 1. strongly agree 2. agree 3. neither agree nor disagree 4. disagree 5 . strongly disagree The most important things that happen to me involve my present job. To me, my job is only a small part of who I am. I am very much involved personally in my job. I live, eat, and breathe my job. Mostcfifmy interests are centered around my job. I have very strong ties to my present job which would be very difficult to break. Usually, I feel detached from my job. Most of my personal life goals are job oriented. I consider my job to be very central to my existence. I like to be absorbed in my job most of the time. 169 Organizational Commitment Questionnaire (Porter, Steers, Mowday_and Boulian, 1974) How much do you agree or disagree with each of the following state- ments? Mark your response on the answer sheet using the scale below: 1. *3. *9. 10. *11. *12. 13. 14. *15. 1. strongly disagree 5. slightly agree 2. moderately disagree 6. moderately agree 3. slightly disagree 7. strongly agree 4. neither agree nor disagree I am willing to put in a great deal of effort beyond that normally expected in order to help this organization be successful. I talk up this organization to my friends as a great organiza- tion to work for. I feel very little loyalty to this organization. I would accept almost any type of job assignment in order to keep working for this organization. I find that my values and the organization's values are very similar. I am proud to tell others that I am part of this organization. I could just as well be working for a different organization as long as the type of work was similar. This organization really inspires the very best in me in the way of job performance. It would take very little change in my present circumstances to cause me to leave this organization. I am extremely glad that I chose this organization to work for over others I was considering at the time I joined. There's not too much to be gained by sticking with this organi- zation indefinitely. Often, I find it difficult to agree with this organization's policies on important matters relating to its employees. I really care about the fate of this organization. For me, this is the best of all possible organizations for which to work. Deciding to work for this organization was a definite mistake on my part. *Indicates reverse scoring of item. 170 The Physical Strain Index Use the following scale to answer these questions: not at all less than once a week 1-2 times a week 3-4 times a week every day mwal-J Listed below are some physical problems that often bother people. How often does each of them happen to you at work? 1. upset stomach 2. backache 3. headache 4. fatigue 171 General Health Questionnaire (Goldberg, 1972) Use the following scale to answer these questions: Compared *1. Been 2. Lost *3. Felt 4. Felt 5. Felt 6. Felt *7. Been *8. Been 9. Been 10. Been 11. Been *12. Been *Indicates . not at all . less than usual . no more than usual . a little more than usual . much more than usual U'l-bOJNH to usual, have you recently: able to concentrate on whatever you're doing? much sleep over worry? that you are playing a useful part in things? capable of making decisions about things? constantly under strain? you couldn't overcome your difficulties? able to enjoy your normal day-to-day activities? able to face up to your problems? feeling unhappy and depressed? losing confidence in yourself? thinking of yourself as a worthless person? feeling reasonably happy, all things considered? reverse scoring of item. 172 Effort/Job Motivation (Patchen, 1965) Answer the following questions using the scales provided: 1. *3. *4. *5. On most days on your job, how often does time seem to drag for MACON)“ you? about half the day or more about one—third of the day about one-quarter of the day about one-eighth of the day time never seems to drag Some people are completely involved in their job--they are job you 1. 2. 3. 4. 5. How isn' 1. 2 3 4 5 absorbed in it night and day. For other people their is simply one of several interests. How involved do feel in your job? very little involved: my other interests are more absorbing slightly involved moderately involved; my job and my other interests are equally absorbing to me strongly involved very strongly involved; my work is the most absorbing interest in my life. often do you do some extra work for your job which t really required? almost every day several times a week about once a week once every few weeks about once a month or less Would you say you work harder, less hard, or about the same as other people doing your type of work at (name of organization)? (11.33me How ing 1. 2. 3. 4. 5. much harder than most others a little harder than most others about the same as most others a little less hard than most others much less hard than most others often do you work through lunch or after regular work- hours without getting paid to do so? almost every day several times a week about once a week once every few weeks about once a month or less *Indicates reverse scoring. Item number 5 was added by the author. Note: 173 Attitudes Towards Unions (Uphoff and Dunnette, 1956) How much do you agree or disagree with each of the following statements? Mark your response on the answer sheet using the scale below: 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 1. strongly agree 2. agree 3. neither agree nor disagree 4 disagree 5 . strongly disagree If it were not for unions, we'd have little protection against favoritism on the job. I think the best person should be kept on the job re- gardless of seniority. Unions impose too many restrictions on employers. Charges of "racketeering " in unions are greatly exaggerated. Employees of a firm have better wages and working conditions when all of them belong to unions. Unions should have something to say about whom the employer hires. A nonunion shOp usually pays lower wages than a union shop. Union rules often interfere with the efficient running of the employer's business. Every worker should be expected to join the union where he/ she works. We need more laws to limit the power of labor unions. Labor unions hold back progress. The high wage demands of unions reduce chances for employment. The growth of unions has made our democracy stronger. The selfishness of employers can be fought only by strong unions. Workers should not have to join a union in order to hold a job. Labor unions should be regulated to a greater extent by the federal government. Every labor union should be required to take out a li- cense from the U.S. government. In a factory where there is a union, workers who are not members should be required to pay the regular union fees if they are getting union rates of pay. Most unions gain their membership by forcing workers to join with threats of violence. If the majority of workers in a plant vote to have a union, the others should be required to join. 174 Turnover Intention Please indicate how strongly you feel at present about leav- ing or staying at MSU. strongly inclined to leave inclined to leave don't know whether I want to stay or leave inclined to stay strongly inclined to stay UlubWNI-J O BIBLIOGRAPHY Allen, V.L. & Greenberger, D.B. Destruction and perceived control. In A. Baum and J.E. Singer (Eds.),Advances in environmental‘psychology: Volume 2--App11cations ofgpersonal control, Hillsdale, J.J.: Lawrence Erl- baum Associates, Inc., 1980. Alloy, L.B. & Abramson, L.Y. Judgment of contingency in depressed and nondepressed students: Sadder but wiser? Journal of Experimental Psychology: General, 1979, 198, 441-485. Alloy, L.B., Abramson, L.Y., & Viscusi, D. Induced mood and the illusion of control. Journal of Personality and Social Psychology, 1981, 41, 1129-1140. Alutto, J. & Acito, F. Decisional participation and sources of job satisfaction: A study of manufactur- ing personnel. Academy of Management Journal, 1974, 11, 160-167. Alutto, J. & Belasco, J. A typology for participation in organizational decision making. Administrative Sci- ence Quarterly, 1972, 11, 117-125. Alutto, J.A. & Vredenburgh, D.J. Characteristics of de- cisional participation by nurses. Academy of Manage- ment Journal, 1977, 10, 341-347. Angyal, A. Foundations for a science of personality, New York: Commonwealth Fund, 1941. Arvey, R.D. Task performance as a function of perceived effort-performance and performance-reward contingen- cies. Organizational Behavior and Human Performance, 1972, 8, 423-433. Banks, M.H., Clegg, C.W., Jackson, P.R., Kemp, N.J., Staf- ford, E.M., & Wall, T.D. The use of the General Health Questionnaire as an indicator of mental health in occupational studies. Journal of Occupational Psychology, 1980, 51, 187-194. Bazerman, M.H. Impact of personal control on performance: Is added control always beneficial? Journal of Ap- plied Psychology, 1982, 61, 472-479. - Blauner, R. Alienation and freedom: The factory Worker and his industry, Chicago: University of Chicago Press, 1964. 175 176 Brass, D.J. Structural relationships, job characteristics, and worker satisfaction and performance. Administra- tive Science Quarter1y, 1981, 36, 331-348. Brehm, J.W. A theory_of psychological reactance. New York: Academic Press, 1966. Brehm, J.W. ReSponses to loss of freedom: A theory of psychological reactance. Morristown, N.J.: General Learning Press, 1972. Brief, A.P. & Aldag, R.J. The job characteristics inven- tory: An examination. Academy of Management Journal, 19781 Zip 659-670. Brief, A.P., Aldag, R.J. & Jacox, A. The impact of task characterisitcs on employee responses in hOSpital nursing. Nursing Administration Quarterly, 1978, 1, 107-114. Buchwald, A.M. Depressed mood and estimates of reinforce- ment frequency. Journal of Abnormal Psychology, 1977, 66, 443-446. Caldwell, D.F. & O'Reilly, C.A. Task perceptions and job satisfaction: A question of causality. Journal of Applied Psychology, 1982, 61, 361-369. Campbell, J.P. & Pritchard, R.D. Research evidence per- taining to Expectancy-Instrumentality-Valence Theory. In M.D. Dunnette (Ed.), Handbook of industrial and gpganizational psychology, Chicago: Rand McNally, 1976, 84-95. Caplan, R.D., Cobb, 5., French, J.R.P., Jr., Harrison, R.V., & Pinneau, S.R., Jr. Job demands and worker health. Washington, D.C.: HEW Publication No. (NIOSH) 75-160, 1975. Coch, L. & French, R.P. Overcoming resistance to change. Human Relations, 1948, 1, 512-532. Cooper, M.R. & Wood, M.T. Effects of member participation and commitment in group decision making on influence, satisfaction and decision riskiness. Journal of Ap- p1ied Psychology, 1974, 62, 127-134. Cronbach, L.J. Coefficient alpha and the internal struc- ture of tests. Psychometrika, 1951, 16, 297-334. Cronbach, L.J. & Furby, L. How should we measure 'change'-- or should we? Psychological Bulletin, 1970'.Z£J 68-80. 177 Dachler, H.P. & Wilpert, B. Conceptual dimensions and boundaries of participation in organizations: A critical evaluation. Administrative Science Quarter1y, 1978, 16, 1-39. Driscoll, J.W. Trust and participation in organizational decision making as predictors of satisfaction. Academy of Manggement Journal, 1978, 11, 44-56. Dunham, R.B. The measurement and dimensionality of job characteristics. Journal of Applied Psychology, 1976, 61, 404-409. Fosco, E. & Geer, J.H. Effects of gaining control over aversive stimuli after differing amounts of no con- trol. Psychology Reports, 1971, 12, 1153-1154. French, W.L. & Bell, C.H. Organizational deve10pment: Be- havioral science interventions for organizational im- provement. Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1984. French, J.R.P., Kay, E., & Meyer, H.H. Participation and the appraisal system. Human Relations, 1966, 13, 3-19. Geer, J.H., Davidson, G.C., & Gatchel, R.I. Reduction of stress in humans through nonveridical perceived con- trol of aversive stimulation. Journal of Personality and Social Psychology, 1970, 16, 731-738. Georgopoulous, B.S., Mahoney, G.M., & Jones, N.W. A path- goal approach to productivity. Journal of Applied Psychology, 1957, 11, 345-353. Glass, D.C. & Singer, J.E. Urban stress. New York: Aca- demic Press, 1972. Goldberg, D.P. The detection of_psychiatric illness py questionnaire. Oxford, England: Oxford University Press, 1972. Golin, S., Terrell, F., & Johnson, B. Depression and the illusion of control. Journal of Abnormal Psychology, 1977, 66, 440-442. Hackman, J.R. & Oldham, G.R. Development of the Job Diag- nostic Survey. Journal of Applied PsychologY: 1975, 66, 159-170. Hackman, J.R. & Oldham, G.R. Motivation through the design of work: Test of a theory. "Organizational Behavior and Human Performance, 1976, 16, 250-279. 178 Hackman, J.R. & Oldham, G.R. Work redesign. Reading, Mass.: Addison-Wesley, 1980. Hackman, J.R. & Porter, L.W. Expectancy theory predic- tions of work effectiveness. Organizational Behavior and Human Performance, 1968, 1, 417-426. Heller, F.A., Drenth, P.J.P., K00pman, P., & Rus, V. A longitudinal study in participative decision making. Human Relations, 1977, 66, 567-587. Henslin, J.M. Craps and magic. American Journal of So- ciology, 1967, 16, 316-330. Herzberg, F. Work and the nature of man. Cleveland: World, 1966. Hrebiniak, L.G. Effects of job level and participation on employee attitudes and perceptions of influence. Acad- emy of Management Journal, 1974, 11, 649-662. Ivancevich, J.M. An analysis of participation in decision making among project engineers. Academy of Management Journal, 1979, 11, 253-269. Jackson, S.E. Participation in decision making as a strat- egy for reducing job-related strain. Journal of Ap- p1ied Psychology, 1983, 66, 3-19. James, L.R. & Brett, J.M. Mediators, moderators and tests for mediation. Journal of Applied Psychology: 1984, 62, 307-321. . James, L.R., Gent, M.J., Hater, J.J., & Coray, K.E. Cor- relates of psychological influence: An illustration of the psychological climate approach to work environ- ment perceptions. Personnel Psychology, 1979, 11, 563-588. 1 James, L.R., Hater, J.J., & Jones, A. Perceptions of psy- chological influence: A cognitive information pro- cessing approach for explaining moderated relation- ships. Personnel Psychology, 1981, 66, 453-477. Jenkins, D. Job power. New York: Doubleday, 1973. Johns, G. Difference score measures of organizational be- havior variables: A critique. Organizational Be- havior and Human Performance, 1981, 11, 443-463. 179 Jorgenson, D.O., Dunnette, M.D., & Pritchard, R.D. Effects of the manipulation of a performance-reward contin- gency of behavior in a simulated work setting. Journal of Applied Psychology, 1973, 21, 271-280. Kanungo, R.N. Measurement of job and work involvement. Journal of Applied Psychology, 1982, 61, 341-349. Keller, R.T., Szilagyi, A., & Holland, W. Boundary span- ning activity and employee reactions: An empirical study. Human Relations, 1976, 22, 699-710. Kimmons, G. & Greenhaus, J.H. Relationship between locus of control and reactions to work characteristics. Psychological Reports, 1976, 22, 815-820. Krantz, D.S., Glass, D.C., & Snyder, M.L. Helplessness, stress level and coronary prone behavior pattern. Journal of Experimental Social Psyghology, 1974, 16, 284-300. Langer, E.J. The illusion of control. Journal of Person- ality and Social Psychology, 1975, 22, 311-328. Lawler, E.E. Motivation in work organizations. Monterey, CA.: Brooks/Cole Publishing Co., 1973. Lawler, E.E. Pay and organizational development. Reading, Mass.: Addison-Wesley Publishing Co., 1981. Lawler, E.E. Pay and organizational effectiveness: A psychological View. New York: McGraw-Hill, 1971. Lawler, E.E. Should the quality of worklife be legislated? The Personnel Administrator, 1976, 61, l7-21. Lawler, E.E. & Porter, L.W. Antecedent attitudes of ef- fective managerial performance. Organizational Be- havior and Human Performance, 1967, 2, 122-142. Lewin, K., Lippitt, R., & White, R.K. Patterns of aggres- sive behavior in experimentally created "social cli- mates." In W.E. Natemeyer (Ed.L.Classics of organi- zational behavior, Oak Park, IL.: Moore Publishing Co., 1978. Lewinsohn, P.M., Mischel, W., Chaplin, W., & Barton, R. Social competence and depression: The role of il- lusory self-perceptions. Journal of Abnormal Psychol- og_, 1980, 62, 203-212. 180 Lied, T.R. & Pritchard, R.D. Relationships between per- sonality variables and components of the expectancy- valence model. Journal of Applied Psychology, 1976, 61, 463-467. Likert, R. New patterns of management. New York: McGraw- Hill, 1961. Lischeron, J. & Wall, T.D. Employee participation: An ex- perimental field study. Human Relations, 1975, 28, 863-884. Locke, E. & Schweiger, D. Participation in decision mak- ing: One more look. In B. Staw (Ed.L Research in organizational behavior, Vol. 1, Greenwich, Conn.: JAI Press, 1979. Margolis, B.L., Kroes, W.H., & Quinn, R.P. Job stress: An unlisted occupational hazard. Journal of Occupational Medicine, 1974, 16, 659-661. Martinko, M.J. & Gardner, W.L. Learned helplessness: An alternative explanation for performance defecits. Academy of Management Review, 1982, 1, 195-204. May, R. Power and innocence. New York: Norton, 1972. McGregor, D.M. The human side of enterprise. New York: McGraw-Hill, 1960. McNair, D.M., Lorr, M., & Droppleman, L.F. Profile of mood states. San Diego, CA.: Educational and Industrial Testing Service, 1971. Mineka, S. & Kihlstrom, J.E. Unpredictable and uncontrol- ‘1able events: A new perspective on experimental neurosis. Journal of Abnormal Psychology, 1978, 81, 256-271. Moch, M.K. Job involvement, internal motivation, and em- ployees' integration into networks of work relation- ships. Organizational Behavior and Human Performance, 1980, gg, 15—31. Morris, J.H., Steers, R.M., & Koch, J.L. Influence of or- ganizational structure on role conflict and ambiguity for three occupational groupings. Academy of Manage- ment Journal, 1979, 22, 58-71. Morse, N.C. & Reimer, E. The experimental change of a major organizational variable. Journal of Abnormal and Social Psychology, 1956, 52, 120-129. 181 Mowday, R.T., Steers, R.M., & Porter, L.W. The measure- ment of organizational commitment. Journal of Voca- tional Behaviour, 1979, 14, 224-247. Nadler, D.A. & Lawler, E.E. Motivation: A diagnostic approach. In J.J. Hackman, E.E. Lawler and L.W. Porter (Eds.), Perspectives on behavior in organiza- tions. New York: McGraw-Hill, 1977. Nie, N.H., Hull, C.H., Jenkins, J.G., Steinbrenner, K., & Bent, D.H. Statistical package for the social sci- ences. New York: McGraw-Hill, 1975. Nunnally, J.C. Psychometric theory. New York: McGraw- Hill, 1975. O'Connor, E.J. & Barrett, G.V. Informational cues and in- dividual differences as determinants of subjective perceptions of task enrichment. Academy of Management Journal, 1980, 2;, 697-716. Oldham, G.R. & Hackman, J.R. Relationships between organi- zational structure and employee reactions: Comparing alternative frameworks. Administrative Science Quarterly, 1981, 26, 66-83. O'Reilly, C.A. & Caldwell, D. Informational influence as a determinant of task characteristics and job satis- faction. Journal of Applied Psychology, 1976, 64, 157-165. Ouchi, W.G. & Maguire, M.A. Organizational control: Two functions. Administrative Science Quarterly, 1975, 29, 559-569. Overmier, J.B. & Seligman, M.E.P. Effects of inescapable shock upon subsequent escape and avoidance learning. Journal of Comparative and Physiological Psychology, 1967, 6;, 23-33. Patchen, M. Participation, achievement and involvement on the job. Englewood Cliff, N.J.: Prentice-Hall, 1970. Patchen, M. Some questionnaire measures of employee moti- vation and morale. Monograph No. 41, Ann Arbor, Michigan: Institute for Social Research, 1965. Pennebaker, J.W., Burnam, M.A., Schaeffer, M.A., & Harper, D.C. Lack of control as a determinant of perceived physical symptoms. Journal of Personality and Social Psychology, 1977, 35, 167-174. 182 Pierce, J.L. & Dunham, R.B. The measurement of perceived job characteristics: The Job Diagnostic Survey ver- sus the Job Characteristics Inventory. Academy of Management Journal, 1978, 22, 123-128. Porter, L.W. & Lawler, E.E. Managerial attitudes and per- formance. Homewood, 111.: Dorsey Press, 1968. Porter, L.W., Steers, R.M., Mowday, R.T., & Boulian, P.V. Organizational commitment, job satisfaction and turn- over among psychiatric technicians. Journal of Ap- plied Psyphology, 1974, 22, 603-609. Roberts, K.H. & Glick, W. The Job Characteristics Approach to task design: A critical review. Journal of Ap- plied Psychology, 1981, 22, 193-217. Roth, 8. & Bootzin, R.R. The effects of experimentally induced expectancies of external control: An investi- gation of learned helplessness. Journal of Person- ality and Social Psychology, 1974, 22, 253-264. Rotter, J.E. Generalized expectancies for internal versus external control of reinforcement. Psychological Monographs, 1966, 22 (whole No. 609) Rousseau, D.M. Technological differences in job character- istics, employee satisfaction and motivation: A syn- thesis of job design, research and sociotechnical systems theory. Organizational Behavior and Human Performance, 1977, 22, 18-42. Salancik, G. & Pfeffer, J.A. A social information process- ing approach to job attitudes and task design. Admin- istrative Science Quarterly, 1978, 22, 224-253. Schmitt, N., Coyle, B.W., White, J.K., & Rauschenberger, J. Background, needs, job perceptions and job satisfac- tion. Personnel Psychology, 1978, 22, 889-901. Schneider, B., Reichers, A.E., & Mitchell, T.M. A note on some relationships between the aptitude requirements and reward attributes of tasks. Academy of Management Journal, 1982, 22, 567-574. Schriesheim, C.A. Job satisfaction, attitudes toward unions, and voting in a union representation election. Journal of Applied Psychology, 1978, 22, 548-552. Schuler, R.S. A role and expectancy model of participa- tion in decision making. Academy_of Management Journal, 1980, 22, 331-340. 183 Schuler, R.S. Role conflict and ambiguity as a function of the task—structure-technology interaction. Organi— zational Behavior and Human Performance, 1977, 22, 66-74. Schuster, J.R., Clark, 3., & Rogers, M. Testing portions of the Porter and Lawler model regarding the motiva- tional role of pay. Journal of Applied Psychology, 1971, 22, 187-195. Schwab, D.P. & Cummings, L.L. A theoretical analysis of the impact of task scope on employee performance. Academy of Management Review, 1976, 2, 23-25. Siegel, A.L. & Ruh, R.A. Job involvement, participation in decision making, personal background and job behavior. Organizational Behavior and Human Performance, 1973, 2, 318-327. Sims, H.P. & SzilagY, A.D. Job characteristics relation- ships: Individual and structural moderators. Organi- zational Behavior and Human Behavior, 1976, 22, 211- 230. Sims, H.P., Szilagyi, A.D., & Keller, R.T. The measure- ment of job characteristics. Academy of Management Journal, 1976, 22, 195-212. Spector, P.E. Behavior in organizations as a function of employee's locus of control. Psychological Bulletin, 1982, 22, 482-497. Steers, R.M. Antecedents and outcomes of organizational commitment. Administrative Science Quarterly, 1977, 22, 46-56. Steiner, I.D. Perceived freedom. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 51, New York: Academic Press, 1970. Stone, E.F. Research methods in organizational behavior. Santa Monica, CA.: Goodyear Publishing Company, 1978. Szilagyi, A.D. & Holland, W.E. Changes in social density: Relationships with functional interaction and per- ceptions of job characteristics, role stress and work satisfaction. Journal of Applied Psychology, 1980, 65, 28-33. Szilagyi, A.D. & Sims, H.P. Locus of control and expectan- cies across multiple occupational levels. 'Journal of Applied Psychology, 1975, 22, 638-640. 184 Tannenbaum, A.S. Control in organizations: Individual adjustment and organizational performance. Admin- istrative Science Quarterly, 1962, 1, 236-257. Tannenbaum, A.S. Social psychologylof the work organiza- tion. Belmont, Ca1if.: Brooks/Cole Publishing Co., 1966. Tannenbaum, R. & Schmidt, W.H. How to choose a leadership pattern. In W.E. Natemeyer (Ed.), Classics of organi- zational behavior. Oak Park, 111.: Moore Publishing Co., 1978. Tosi, H. A reexamination of personality as a determinant of the effects of participation. Personnel Psychol- ogy, 1970, 22, 91-99. Turner, A.N. & Lawrence, P.R. Industrial jobs and the worker. Boston: Harvard University, Division of Research,Graduate School of Business Administration, 1965. Uphoff, W.H. & Dunnette, M.D. Understanding the union member. Minneapolis: University of Minnesota, In- dustrial Relations Center, 1956. Vroom, V.H. Some personality determinants of the effects of participation. Journal of Abnormal and Social Psychology, 1959, 22, 322-327. Vroom, V.H. Somelpersonality determinants of the effect of participation. Englewood Cliffs, N.J.: Prentice- Hall, 1960. Vroom, V.H., Work and motivation. New York: John Wiley & Sons Inc., 1964. Ward, W.C. & Jenkins, H.M. The display of information and“ the judgment of contingency. Canadian Journal of Psy- chology, 1965, l2, 231-241. Weiss, D.J., Dawis, R.V., England, G.W., & Lofquist, L.H. Minnesota studies in vocational rehabilitation: 22, Manual for the Minnesota Satisfaction QuestiOnnaire. Minneapolis, Minn.: University of Minnesota, 1967. White, R.w. Motivation reconsidered: The concept of com- petence. PsycholOgical Review, 1959, 22, 297-333. Wood, M.T. Participation, influence and satisfaction in group decision making.’ Journal of V0cationa1 Be- havior, 1972, 2, 389-399. 185 Woodworth, R.S. Dynamics of behavior. New York: Holt, 1958. Zajonc, R. Feeling and thinking: Preferences need no in-fi ferences. American Psychologist, 1980, 22, 151-175. Zalesnik, A., Kets de Vries, M.F.R., & Howard, J. Stress reactions in organizations: Ayndromes, causes and consequences. Behavioral Science, 1977, 22, 151-162.