\~ \ ‘6 ‘ ‘ N K- .\\\‘\\\1~ - ‘\ \‘fi~'.‘.. \\V ‘\:~.Z~-;--~ . K .. ‘ —\—‘ ‘us'. - < “13-fo «.s /‘_’ N' / ..a..‘:' .1- “‘3 ) .o' u I’“ l’ IIIIIIII CC? 293“} “$1295 .2199 ‘ ’ 033.09 1999 (:/K ABSTRACT. EFFECT OF FEEDBACK ON PERFORMANCE AND JOB SATISFACTION IN AN ORGANIZATIONAL SETTING By Jay S. Kim This study examined the effect of feedback on job satisfaction ”and performance in an industrial organization. Basically, this study investigated (1) the degree of job satisfaction before and after the feedback intervention as a reinforcer, (2) the differential effect of the different modes of feedback on workers' job satisfaction, (3) the degree of performance of work groups before and after the intervention of feedback as a reinforcer, and (4) the differential effect of the different modes of feedback on performance in an industrial organization. The nonequivalent control group design was employed to investigate the effect of feedback on performance and job satisfaction in an indus- trial organization. Four separate plants where about 220 employees were performing service type jobs were assigned to three experimental groups and the control group. Three experimental groups received "extrinsic feedback", "intrinsic feedback", and "extrinsic and intrin- sic feedback" respectively, and the control group received "no feed— back" during the experimental period of three months. A pretreatment satisfaction measure which was obtained through an attitude survey taken prior to the experimental intervention was used Jay S. Kim as baseline in analyzing the job satisfaction data of the experimental period. For performance measures, the average performance measures of the previous year were used as baseline in the analysis of three monthly performance measures during the experimental period. The analysis of variance with repeated measures and covariance analysis were employed to investigate the pre-post difference for experimental groups and the differential effect of different modes of feedback on performance and job satisfaction. Results showed that there was statistically significant improvement ‘on three out of four performance measures investigated. Further, the performance measure of the three experimental groups combined was sig- nificantly better than the control group on two objective performance measures. While the combined feedback (i.e., extrinsic and intrinsic feedback) had greater impact on performance than did intrinsic feed~ back or extrinsic feedback, that difference did not produce as signif- icant an effect as the presence or absence of feedback in this study. Findings related to the effect of feedback on job satisfaction showed conflicting results. It was shown that on all four intrinsic job satisfaction dimensions investigated in this study, no pre—post differences were observed. 0n the other hand, workers' job satisfac- tion related to interpersonal dimensions increased after feedback. In contrast, significant decreases in job satisfaction with "present pay" and "job security" were found. Additionally, there was no differential effect of different modes of feedback on job satisfaction. Nhen ini— tial differences among the four groups were adjusted, no statistical difference between the groups was observed. Jay S. Kim In general, the results of this study support one of the basic premises of research based on reinforcement theory, that behavior can. be changed directly without going through the process of attitude change. Feedback given to workers in this study has focused on the ob- servable task performance, not on the workers' job satisfaction. The findings showed that job satisfaction of workers was less affected by feedback while job performance was improved as a result of feedback. EFFECT OF FEEDBACK ON PERFORMANCE AND JOB SATISFACTION IN AN ORGANIZATIONAL SETTING By Jay SinTim A DISSERTATION Submitted To _ Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Management 1974 ' ht rig Copy 3y 'm liy 5. K1 1974 ACKNOWLEDGMENTS I would like to express my deepest appreciation to my dissertation chairman, N. Clay Hamner, two dissertation committee members, Douglas T. Hall and Lawrence w. Foster, and Donald E. Burwell of Michigan Bell for their invaluable support and assistance in the completion of this dissertation. I also thank Richard F. Gonzalez, other faculty members and fellow doctoral students in the Department of Management at Michigan State University for their continuous support and encouragement. Finally, I would like to tell my wife, Mary, how much I appreciu ated her help. TABLE OF CONTENTS LIST OF TABLES . . . .................. . LIST OF FIGURES ..... . . . . . . . . . . . . . . . . . Chapter I. INTRODUCTION AND REVIEW OF LITERATURE . . . . . . Review of Literature . . ..... . . . Reinforcement Theory and Its Industrial Application ......... . . . . . . . Effect of Feedback on Attitudes . . . . . . . Effect of Feedback on Behavior ...... . . Effect of Ex rinsic and Intrinsic Feedbacl< on Performance . . . . . . . . . . Summ ary Oi Literature Review . . . . . . . . . Hypotheses . . . . . . . . . . . . . . . . . . . II. META ODLOGY ........ . . Experimental Design . . . . . . . ..... . . Independent Variables ......... . . . . Dependeni: variables . . . . . . . . . Measures, Data Collection, and Scoring . . . Analysis of Data ........ . . . . . . Summary . . . . . . . . . . . . . . . . . . . . III. RESULTS Performance . ........... . . . . . . Cost Performance . . . . . . . . . . . . . . . Safety . . o . . . . . ............ Absenteeism . . . ......... . . . . . SCIVICQ ............. . . . o a Job Satisfaction ....... . . . . . . . . . Job Security . . . . . . . . . . . . . . . Present Pay . . . . . . . . . . . . Supervision on Present Assignment . . . . . Cxi,.s1te Job Sat.is faction . . . . . . . . . . Job Satisfaction~ Performance Relationship . . . iii . I C Q C O Page viii d .4 Now w 55 Chapter IV. DISCUSSION AND CONCLUSION ............ Effect of Feedback on Performance . . . . . . . . . . Effect of Feedback on Job Satisfaction . . . . . Implications of the Findings ....... . . . Limitations of the Present Study and Directions for Future Research . . . . . . iv Table 10. ll. l2. l3. l4. LIST OF TABLES Test~Retest Reliability and Internal Consistency of Job Satisfaction Scores . . . . . . . . . . . . . . . Actual Means, Standard Deviations, and the Adjusted Means of "Cost Per ormance". . . . . . . . . . . . . . . Analysis of Variance Mith Repeated Measures Table on "Cost Performance" for Mode of Feedback and Period . . . Planned Comparisons Summary of Covariance Analysis on ”Cost Performance" for the Posttreatment Periods . . . . Actual Means, Standard Deviations and Adjusted Means of "Safety" ....... . . . . . . . . . . . . . . . . Analysis of Variance with Repeated Measures Table on “Safety“ for Mode of Feedback and Period . . . . . . . . Planned Comparisons Summary of Covariance Analysis on "Safety” for the Posttreatment Periods . . . . . . . . . Means and Standard Deviations of "Absenteeism" . . . . . . Analysis of Variance with Repeated Measures Table on the Logarithmic Transformed Absenteeism for Mode of Feedback and Period . . . . . . . . . . . . . . . . . . Planned Comparisons Summary of Analysis of Variance with Repeated Measures on Logarithmic Transformed Absenteeism. Planned Comparisons Summary of Covariance Analysis on Absenteeism for the Posttreatment Periods . . . . . . . 'Actual Means, Standard Deviations and the Adjusted Means of “Service“ . . . . ........ . . . . . . . Analysis of Variance with Repeated Measures Table on Service for Mode of Feedback and Period . . . . . . . . Planned Comparisons Summary of Covariance Analysis of Service for the Posttreatment Periods. . . . . . . . . . 43 44 46 48 54 54 Table Page l5. Summary of Findings on Performance Measures by Analysis of Variance with Repeated Measures and Covariance Analysis ............ . . . . . . . . . . . . 6O l6. Summary of Findings on Job Satisfaction by Analysis of Variance with Repeated Measures and Covariance Analysis . . . . . . . . . . . . . . . . . . . . . . . . 62 l7. Means, Standard Deviations and Adjusted Means of Job Satisfaction on ”Job Security“. . . . . . . . . . . . . . 64 l8. Analysis of Variance with Repeated Measures Table for "Job Security" ..... . . . . . . . . . . . . . . . . . 65 19. Planned Comparisons Summary of Covariance Analysis on “Job Security“ for the Posttreatment Period . . . . . . . 66 20. Means and Standard Deviations of Job Satisfaction for "Present Pay" . . . . . . ....... . . . . . . . . . 67 2l. Analysis of Variance with Repeated Measures Table for “Present Pay“ . . . . . . . . . . . . . . . . . . . . . . 68 22. Planned Comparisons Summary of Covariance Analysis on “Present Pay“ for the Posttreatment Period . . . . . . . 69 23. Means and Standard Deviations of Job Satisfaction for ”Peeple on Present Assignment“. . . . . . . . . . . . . . .70 24. Analysis of Variance with Repeated Measures Table for ”People on Present Assignment”. . . . . . . . . . . . . . 7l 25. Planned Comparisons Summary of Covariance Analysis on “People on Present Assignment” for the Posttreatment Per-IOd o ooooooo o o o v o o o o I o o n g o o I o :12 26. Means and Standard Deviations of Job Satisfaction for “Supervision of Present Assignment” . . . . . . . . . . . 73 27. Analysis of Variance with Repeated Measures Table for “Supervision of Present Assignment" . . . . . . . . . - - 74 28. Planned Comparisons Summary of Covariance Analysis on “Supervision on Present Assignment” for the Posttreatment Period . . . . . .., . . . . . . . . . . . 74 29. Summary of Analysis of Variance for “Composite Job Satisfaction Score" . . . ...... . . . . . . . . . . 76 vi Table Page 30. Analysis of Covariance Table for the "Composite Job Satisfaction" .......... . . . . . . . . . . 76 31. Kendall Tau Coefficients Between Job Satisfaction and Performance at the End of Experimental Period . . . . 77 vii figure l. Operant Conditioning Summary ........... . . . 2. Experimental Design for the Study of the Effect of Feedback on Job Satisfaction . . . . . . 3. Experimental Design for the Study of the Effect of Feedback on Performance . . . . . . . . . 4. Group Feedback Sheet .......... . . . . . . . 5. Job Descriptions of the Subjects ......... . . LIST OF FIGURES viii 27 28 29 34 CHAPTER I INTRODUCTION AND REVIEW OF LITERATURE Introduction Ever since Skinner stated that "behavior is a function of its consequences," there have been both a wide variety of controversy among behavioral scientists and valuable theoretical development by his followers. In recent years a concerted effort has been made by social scientists to apply the reinforcement principles and their implications for describing and modifying the behavior of human beings in various organizational settings. While several field research studies on reinforcement principles and behavior modification tech- niques have been conducted in educational and in mental hospital set- tings, until recently it had been largely neglected in the industrial setting. With the rising disillusionment with cognitive models in pre- dicting and controlling human behavior, a number of theorists in the area of organizational behavior have begun to focus on the acog- nitive models as bases for research for the last several years. One of the basic premises of the research based on reinforcement theory is that if one attempts to change the behavior of others, he should focus on the behavior itself. It is assumed that behavior can be changed directly without going through the process of attitude change. According to this premise, once behavior is changed, attitude will become consistent with that behavior. Thus, research studies based on reinforcement theory concentrate on the observable behavior and its relationships with the organizational contingencies rather than on the internal state of the worker which purportedly determines one's behavior. Most of the research applying the principles of reinforcement to industry has focused on the monetary reward as reinforcer in the labo— ratory setting. Research focusing on the effect of feedback as a reinforcer on the performance of work groups within the reinforcement- based framework is virtually non-existent. The purpose of this research is two-fold. The first objective is to investigate (l) the degree of job satisfaction before and after the feedback intervention as a reinforcer, and (2) the differential effect, if any, of the different modes of feedback on worker's job satisfaction in an industrial organization. The second objective is to investigate (l) the degree of performance before and after the intervention of feedback as a reinforcer, and (2) the differential effect, if any, of the different modes of feedback on worker's performance in an indus- trial organization. Thus, this study will attempt to expand our knowledge pertaining to the effect of feedback on job satisfaction and performance as they are manifested by the result of feedback in an organization. Prior to discussing the hypotheses and methodology to achieve these objectives, a review of the relevant literature is in order. The literature review will consist of four sections: (l) a brief description of reinforcement theory, (2) the application of reinforcement theory in industry, (3) the effects of feedback on job satisfaction and performance, and (4) intrinsic versus extrinsic feedback as reinforcer. Review of Literature Prior to the description of the research design, some of the relevant literature related to this study should be reviewed. This section starts with a brief description on reinforcement theory, fol- lowed by discussion on industrial application of reinforcement theory. Major emphasis, however, is placed on literature reviewing the effect of feedback on attitude and behavior, and the effect of different types of feedback on attitude and behavior. Reinforcement Theory and Its Industrial Application One of the basic premises of reinforcement theory is that all behavior is learned. Learning is the acquisition of knowledge, with performance being the translation of this knowledge into practice (Hamner, 1974a).Skinner and his followers assume that learning is a prerequisite for behavior and any set of conditions which favor learn- ing is called "reinforcement." Therefore, according to behavioralists, the highly differentiated behavior among people can only be explained in terms of the history of the reinforcement of that behavior (Morse, l966). Operant conditioning is the process by which behavior is modified by manipulation of the contingencies of the behavior. Operant condi- tioning rests on the assumption that an individual learns mainly by producing changes in his environment and that the consequences of that environmental change influence behavior. Skinner (l969) suggested that the interaction between an individ- ual and his environment should contain three elements: (I) the setting in which a response occurs (i.e., environmentL (2) the response itself (i.e., performance) and(3) the reinforcing consequences (i.e., rein- forcement). The interrelationships among these three elements are called "contingencies of reinforcement." Skinner maintained that the consequences determine the likelihood of behavior. Therefore, this theory suggests that to change behavior, the consequences of the behav- ior must be changed, i.e., the contingencies must be rearranged (Beh- ling, et_al,, l974). Several types of reinforcement and schedules by which reinforce- ment can be arranged have been suggested. For example, positive rein- forcement and avoidance learning are types of reinforcement designed to increase the probability of an operant response. Alternatively, extinc- tion and punishment illustrate types designed to decrease the probabil- ity of an operant response. The effectiveness of these types is dependent upon the manner in which they are scheduled. A reinforcement schedule is a formal speci- fication of the occurrence of a reinforcer in relation to the behav- ioral sequence to be conditioned (Adam and Scott, 1971) Aldis (l96l). suggested two basic types of schedules by which positive reinforcement can be offered: continuous reinforcement schedule (in which reinforce- ment follows every correct operant response) and partial reinforcement schedule (in which reinforcement does not occur after every correct operant response). This latter category may be classified into two types: variable partial reinforcement schedules and fixed partial reinforcement schedules. Under fixed interval reinforcement schedules, a reinforcer is administered only when the desired response occurs after the passage of a specified period of time since the previous reinforcement (i.e., fixed interval reinforcement schedule) or a reward is delivered only when a fixed number of desired responses take place (i.e., fixed ratio reinforcement schedule). 0n the other hand, under variable interval reinforcement schedules, reinforcement is administered at some variable interval time (i.e., variable interval reinforcement schedule) or a reward is delivered only after a number of desired responses with the number of desired responses changing from the occurrence of one rein- forcer to the next (i.e., variable ratio reinforcement schedule) (Hamner, l974a). The effect of each type of reinforcement and the various methods of arranging contingencies on worker performance as summarized by Behling, §t_al,, (1973) is shown below in Figure 1. In recent years, several theorists and practitioners have shown considerable interest in the applicability of reinforcement theory to industrial settings. Aldis (1961) examined the implications of various reinforcement schedules on wage payment systems, Christmas bonuses, etc. He suggested new methods of wage payments which rely on more immediate rewards will have impact on both productivity and worker satisfaction. Nord (1969) suggested the possibilities of applying positive reinforcement in areas in industrial organizations such as training and development, compensation, job design, supervision and leadership, etc. For example, positive reinforcement can be applied in the area of compensation and alternative rewards by making rewards FIGURE 1 OPERANT CONDITIONING SUMMARY Arrangement of Schedule of Effect on Behavior When Effect on Behavior When Reinforcement Reinforcement Applied to the Removed from the Contingencies Contingencies Individual Individual Continuous Fastest method Fastest method Reinforcement. to establish a to extinguish a new behavior. new behavior. Partial Slowest method Slowest method Reinforcement. to establish a to extinguish a new behavior. new behavior. Variable More consistent Slower extinction Partial response fre- rate. Reinforcement. quencies. Fixed Less consistent Faster extinction Partial response fre- rate. Reinforcement. quencies. Positive Increased fre- Return to pre- Reinforcement quency over pre- conditioning Avoidance conditioning level. Reinforcement level. Punishment Decreased fre- Return to pre- Extinction quency over pre- conditioning conditioning level. level. Behling,et_al,, "Present Theories and New Directions in Theories of Work Effort,” Journal Supplement and Abstract Service of the Ameri- can Psychological Corporation, 1973. contingent upon performance and/or by applying the desired variable ratio schedule of reinforcement. Further, Nord (1969, p. 399) argued that the operant approach, which focuses on the exchange of reinforcers within and between an organization and its environment "may prove to be an invaluable asset to both administrators and students of admini- stration and organization." Describing the role of the organization in the future in influenc- ing human behavior, Porter (1973) argued that both human behavior and work environments are more adaptable than commonly recognized. Thus, an organization can influence human behavior in the future by making the work environment more rewarding. In other words, Porter (1973, p. 114) suggested, "organizations will need to apply the concept of reward contingencies -- that is, to relate rewards to specific work behavior." He stated, the principles of shaping (operant conditioning) can be powerful tools to help the marginal worker transform his actions into acceptable employment behavior. Similarly, Scott and Cummings (1973) stated that organizational leaders must resort to environmental changes as a means of influencing the behavior of workers, and reinforcement principles are the most use- ful method in this regard. Skinner (1973) also offered his ideas on how operant conditioning can be appliedto industrial setting. Some of his suggestions include supervision by positive reinforcement, and a lottery system for por- tions of compensation and also for reduction of absenteeism. Several others have argued for a behavioral approach for a selected aspect in the industrial setting. Goldstein and Sorcher (1974) addressed themselves to supervisory training. Specifically, their attempts have been geared to the change of the work-related attitudes and behavior of supervisors through a procedure which includes the sequence of modeling, role-playing, and reinforcement "aiming directly at behavior change without relying on the diversionary tactics of atti- tude change." (Goldstein and Sorcher, 1974, p. 41). Another behavioral approach for training supervisors suggested by Luthans and Lyman (1973) is called organizational behavior modifica- tion. Like Goldstein and Sorcher (1974), this training program meas- ures whether an observable job behavior has been changed rather than attempting to measure and change internal states of one's attitudes and values. Since the measures take the form of observable job behavior - such as the units produced, orders filled, etc. - the continuous feedback of performance in behavioral terms is possible and can be used as a learning device and source of reinforcement. Further, Luthans and White (1971) argued that behavior modification can be applied in the areas of compensation, absenteeism, tardiness and motivation of organizational participants. Some other areas in which the behavioral approach has been suggested for potential benefit include organizational development (Beer, 1969), and counseling and guidance (Michael and Meyerson, 1962). To summarize, the arguments discussed above share at least two points in common. First, they all seem to agree that the procedure and schedules of reinforcement can be applied to change the supervi- sory and worker behavior in the industrial setting. Second, they suggest abandoning the preoccupation with the inner life of man (i.e., attitudes) and concentrating on man's observable behavior and trans- actions with the environment. The emphasis on observable behavior enables more effective feedback, which can become a source of rein- forcement. Effect of Feedback on Attitudes It has been generally conceived that feedback is one of the most critical factors in learning, motivating, and enhancing performance level. The reinforcing effect of feedback has been well expressed in terms of stimulus-response by Thorndike (1911). He asserted that, of several responses made to the same situation, those which are accom- panied or closely followed by satisfaction will be more likely to recur; those which are accompanied or closely followed by discomfort will be less likely to occur (Thorndike, 1911, p. 244). This is the basis of reinforcement theory which views that all forms of feedback can be regarded as rewarding or punishing and that a rewarding result preserves the behavior which preceded it by a relatively simple mecha— nism. As defined by Dunnette and Kirchner (1965), an attitude is a "relatively stable or enduring syndrome of consistent responses made by an individual with respect to some psychological object -- any symbol, slogan, product, institution, person, group, or issue -- with which he may be confronted." (p. 215) However, only a limited number of researchers have addressed themselves to the effect of feedback on attitude. Watson (1969) found the potential effect of immediate and direct feedback on interpersonal changes in self-awareness, perception of others, acceptance, self-confidence, etc. Another study which tends 10 to support the findings of Watson was made by Leavitt and Mueller (1951). They found that zero feedback was accompanied by low confidence and hos- tility, while free feedback was followed by high confidence and friend- liness. In an experiment involving the T-group process, Kolb, Winter, and Berlow (1968) found that subjects in groups which received feedback reported higher self-perceived change than did non-feedback subjects. Several others studies related to the performance appraisal pro- cess in organizational settings also demonstrate the effect of feedback on attitude (Mahler, 1957; Meyer and Walker, 1961a; Kay and Meyer, 1962). For example Meyer and Walker (1961b) at General Electric found in a study based on managers and specialists that the participants' attitude toward the merit-pay salary plan was significantly correlated with their views as to how well feedback discussion was handled during the performance appraisal. Similarly, Ivancevich, Donnelly and Lyon (1970) found in their study on Management-By—Objective that the sub- jects of one company in which a feedback session was employed every three months showed more positive results in perceived need satisfac- tion than those of another company in which feedback was employed only once a year. Especially, significant improvement was indicated in four of the five need categories for middle managers. Furthermore, they contended that the exact number of feedback sessions for optimum level of need satisfaction must be determined on an individual firm basis. The preceding literature related to feedback effect on attitude has been primarily concerned with feedback provided by superiors or experimenters; that is, extrinsic feedback. This extrinsic feedback 11 is often confused with intrinsic feedback which refers to the degree to which employees receive information as they are working which reveals how well they are performing on the job. Hackman and Lawler (1971) found that there was little agreement between employees and researchers in their ratings of jobs as high or low on the feedback dimension, while consensus existed in ratings on the other three job dimensions - variety, autonomy, and task identity. Hackman and Lawler attributed this disagreement to raters' confusion in distinguishing between extrinsic feedback and intrinsic feedback. They suggested that this difficulty partially stems from the fact that some feedback is provided directly by the job as an employee works, while additional feedback may be provided by the supervision. Further, they reported that feedback of task (i.e., intrinsic feedback) was significantly correlated with all job satisfaction items under study, (such as self- esteem obtained from job, independent thought and action, security, pay, feeling of worthwhile accomplishment, promotion, etc.), especially for those employees with high-order need strength. While the authors reviewed to this point seem to suggest the poten- tially positive impact of feedback on attitude, several other authors have presented different results, Harrison (1969), using graduate stu- dents as subjects, examined the effect of direct formalized feedback on attitude in a group counseling situation. In his study, the experi- mental group received the direct formalized feedback of the other group members‘rating of their behavior, while the control group received no information. One of the key findings of this laboratory study indi- cated that the use of direct formalized feedback techniques in groups 12 did not produce attitude change. Another study which tends to concur with this finding was conducted in a large corporation by Smith and Knight (1959) on first level supervisors who participated in a manage- ment training program. Smith and Knight reported that the groups receiving personalized feedback and the groups receiving no feedback on their behavior showed no difference on ”self insight" which was measured by having each conferee indicate who assumed the leadership role well. Therefore, evidence pertaining to the impact of feedback on atti- tude is still tentative and the specific dynamics of feedback on atti- tude are not fully known at this time. Especially, none of these auth- ors specifically address themselves to the interactive nature of extrin- sic and intrinsic feedback on attitude. Further, attitude as the dependent variable employed by these various researchers varies so widely that any inferences drawn from these studies are far from being conclusive. Effect of Feedback on Behavior As discussed earlier, numerous theorists and practitioners sug- gested that the reinforcement-based approach applied in organizational settings can produce changes in behavior (Aldis, 1961; Porter, 1973; Scott and Cummings, 1973). Studies regarding behavioral change focus on the impact of monetary reinforcement and reinforcement schedule on productivity (Yukl, et_al,, 1972), absenteeism (Gamboa and Pedalino, 1974; Luthans and White, 1971; Adam, 1972). Prior to the discussion of feedback effect on behavior, some of these studies will be reviewed in this section. l3 Yukl, Wexley, and Seymore (1972) examined the effectiveness of pay incentives under various schedules of reinforcement. They reported that pay incentives were more effective in motivating increased produc- tion when used with a variable ratio schedule than when used with a continuous reinforcement schedule. However, Latham (1974), in con- junction with Yukl, found the opposite results among tree planters of Weyerhaeuser Company. When monetary reward was used as a reinforcer, he reported that the highest increase in performance was shown under continuous reinforcement condition, rather than under two other vari- able reinforcement schedule conditions. He attributed these results to the difference in sex, leadership, and situational variables of the study. Adam (1972) investigated in a laboratory setting the effectiveness of operant conditioning procedures in obtaining changes in performance quality over time. One of the key findings was that conditioning pro- cedures, when compared to an absence of conditioning, facilitated higher levels of performance quality. Gamboa and Pedalino (1974) conducted a field study at a manufac- turing-distributive company regarding the effect of a lottery incentive system on absenteeism. They reported a significant decrease in absen- teeism after a lottery system was introduced. Further they indicated that stretching the reinforcement schedule from weekly to biweekly could be achieved without resulting in a significant increase in the rate of absenteeism. As stated earlier, these studies employed primarily monetary reward as a type of reinforcer to induce behavior change. There are 14 a number of studies which examined the effect of performance feedback as a reinforcer. Ammons (1954) after reviewing studies on the effects of knowledge of results on performance, suggested that knowledge of results universally tended to improve the performance in laboratory settings. Both the studies of Smith and Knight (1959) and Harrison (1969), which were previously discussed, showed the effect of feedback on behavior. For example, the study of Harrison indicated that feedback produced behavioral changes but not attitude change in a group counsel- ing situation. Similarly, Smith and Knight reported that groups receiv— ing personalized feedback consistently improved group problem solving efficiency as compared to groups receiving no feedback. This finding is consistent with the findings of Pryer and Bass (1950) which showed groups receiving feedback solved their problems more accurately than control groups. A field experiment conducted by Hundal (1969) indicated that the performance of subjects in a repetitive industrial task increased in direct relation to the degree of awareness of their performance. Addi- tionally he reported statistically significant differences in perform- ance between the pre-experimental and the post-experimental periods for the subjects who were given accurate information about their perform- ance. Another field experiment on life insurance agents reported by Weitz, et_al, (1954) indicated that the experimental group (those with whom the home office maintained contact by means of production bulle- tins and personal letters) showed better performance than did the con- trol group (which had no home office contact). Miller (1965) at 15 General Electric reported that increasing the amount of feedback from foremen to workers improved the performance of the workers. He also provided evidence showing that the effects of feedback were related to performance quality. Partially contradicting the findings of Smith and Knight (1959) and Harrison (1969), Cook (1968) found positive impact of feedback on performance as well as on attitude of the participants. She reported that the attitude and performance of the participants were directly related to the frequency of feedback. Several industrial organizations have attempted to apply the rein- forcement principles through "positive reinforcement programs." Some efforts of the most publicized companies include the work of E. J. Feeney at Emery Air Freight (Business Week, December 18, 1971, December 2, 1972), E. D. Grady of Michigan Bell Telephone Company (Business Week, December 2, 1972; International Management, October, 1973), Clifford Mrazek of Belden Corporation, J. LeCheminant of Sun Life of Canada, among many others (Hamner, 1974b). While these companies emphasize different aspects of reinforcement theory, they generally follow the work done by Mager (1962, 1970) and Rummler (1972) as guidelines. Both Mager and Rummler focus on operationally defined goals and feedback on worker's performance as critical elements for the success of the pro— gram. Following this argument, Feeney stated that most performance prob- lems in industry are execution problems; that is, low performance may result from lack of feedback, task interference or punishment but rarely from lack of motivation (Laird, 1971). Further, Feeney stated: 16 I see feedback and reinforcement as basic organizational activities; operat- ing at different levels with different behaviors for reinforcement at each level. (Laird, 1971, p. 55) Feeney's claims include the increased profits of about $2 million annually, the immediate improvement of performance sustained for four years, and beneficial side effects in terms of worker-management rela- tions. Part of the empirical evidence derived from the program imple- mented by Feeney in Emery Air Freight and Grady in Michigan Bell were reported by Brethower (1973) and by Grady and Hamner (in preparation), respectively. Brethower (1973) inquiring about the extent to which feedback can be used by managers in business organizations, investigated whether self-recording of performance information and subsequent posi- tive reinforcement increase the number of callbacks completed on sched- ule to customers by customer service employees in Emery Air Freight. Results showed performance improvement from a group mean percentage of 25% during base line period to a group mean percentage of 94% during the feedback period of 21 weeks. Further, this improvement of perform- ance over baseline was maintained over 15 months. Based on the results of this and other studies, Brethower concluded that (l) managers can use performance indicators as feedback, (2) feedback significantly improves performance, and (3) effects of feedback can be maintained over time. One of the limitations of her studies, however, lies in the absence of control or comparison groups, which may weaken the degree of internal validity of the findings. Therefore, research evidence tends to suggest that feedback can improve performance of various tasks, and feedback can be employed 17 effectively in industrial organizations for the improvement of perform- ance. Again, as has been the case for studies on attitude, the ques- tion regarding different effects on performance resulting from dif- ferent modes of feedback remains unanswered. For example, does intrin- sic feedback have a different degree of impact compared to extrinsic feedback on performance or vice versa? Studies dealing with extrinsic versus intrinsic reinforcement in general and feedback in specific are discussed in the following section. Effect of Extrinsic and Intrinsic Feedback on Performance The cognitive theorists, notably expectancy theorists, tend to emphasize intrinsic reward, while reinforcement theorists place more emphasis on extrinsic reinforcement (Brethower, 1972). Yet, both expectancy theory and reinforcement theory of task motivation make a common assumption that the effects of intrinsic and extrinsic rein- forcement are additive; i.e., a worker will be more motivated (or responsive) to complete a task when both kinds of rewards (or rein- forcement) are present. Bandura and Perloff (1967) have suggested that self-reinforcement (which seems implicit to the concept of intrin- sic motivation) directly parallels the effects of external reinforce- ment. O'Leary, Poulos and Devine (1972) maintained that the use of tangible reinforcers is valuable in building "intrinsic” reinforcers. Aronfreed (1968) suggested that emphasis on the intrinsic by some and emphasis on the extrinsic by other theorists is unwarranted since one without the other is relatively ineffective and inefficient. Deci (1972a, 1972b) has suggested that extrinsic rewards and intrinsic task motivation can be conflicting in nature. Engaging 18 college students as subjects in a puzzle game under three different conditions (rewarded with money, rewarded with positive feedback, and ' no reward), Deci used the amount of time they worked on the puzzles during the free-choice time as the dependent measure of motivation. Based on this measure, he interprets that, when money was contingently paid to subjects for performing intrinsically motivating tasks, their intrinsic motivation decreases, with the exception of positive verbal reinforcement. However, several authors (Hamner, l974b Yukl, 1974) have suggested that the cognitive evaluation theory of Deci is sub- jected to a number of methodological problems that raise doubts about the data upon which Deci's theory is based. Erskine (1974) conducted a laboratory experiment designed to test the effects of task character- istics and monetary payments on the persistence of behavior. His data suggested that money can add to the reinforcing characteristics of task as an additional source of reinforcement. Similarly, Hamner and Foster (in press) tested cognitive evalua- tion theory in a laboratory setting using college students as subjects. They found support for a reinforcement based prediction that task characteristics and money payments when appropriately arranged and scheduled increase task performance and/or self-reports of task attrac- tiveness and interest. Thus, they indicated support for a reinforce- ment and expectancy theory explanation that the effect of intrinsic and contingent extrinsic rewards on task motivation and performance are additive in nature. Several researchers (Bandura, 1969; Kanfer, 1971) have shown intrinsic feedback (i.e., self-generated evaluation) can serve as a 19 basis for self-rewards and self-criticisms. Further, intrinsic feed- back has reinforcing properties (Bandura and Perloff, 1967; Montgomery and Parton, 1970). Warm et_al,, (1972) investigated the relative effi- ciency of two forms of feedback (experimenter controlled feedback and subject controlled feedback) in enhancing vigilance performance. They found that both feedback groups had faster response time than control group receiving no feedback. However, results did not show a signifi- cant difference between feedback groups, although subjects with extrin- sic feedback performed better than did the intrinsic feedback subjects. Therefore, as Nord (1969) argued, it is reasonable to assume that intrinsic rewards may have the same consequences as extrinsic rewards if they are given on the same reinforcement schedule. Under this pre- mise, it seems possible to explain the "intrinsically motivated" behav- ior within a framework of reinforcement. Further, if these two types of reinforcement are in fact closely associated with each other, it can be possible that an interactive effect of intrinsic and extrinsic feed- back may empirically exist in a given situation. Yet, Baron and Ganz (1972) found in an experiment using elementary school children that intrinsic feedback, extrinsic feedback and the combined feedback of both failed to produce significant difference in performing a simple form discrimination task. To summarize, evidence arguing for the positive impact of feedback on performance seems to be stronger than evidence showing the positive impact of feedback on attitude. What seems unclear at this time is whe- ther there iS a differential and/or interactive effect of extrinsic and intrinsic feedback on performance. This issue of the effect of 20 extrinsic-intrinsic feedback on work performance and/or job satisfac- tion within a reinforcement based framework has not been examined in real industrial settings. Summary of Literature Review As the preceding review illustrated, the arguments surrounding reinforcement theory and its potential value in application to indus- trial organizations are, by and large, testimonial and generally lack supportive, empirical evidence. Most of the empirical studies related to this issue, having been conducted in laboratory situations, provide little comparability and offer conflicting evidence. While several researchers and practitioners seem to suggest the potential effect of feedback on attitude and/or performance, evidence is still tentative with the specific nature of feedback on attitude and/or performance not being fully known. Especially lacking is research focusing on the effect of different types of feedback on attitude and/or performance. Presently no study has attempted to investigate the impact of extrinsic and intrinsic feedback on attitude and/or performance and its interrelationship in the industrial organi- zation. Therefore, several questions regarding feedback in industrial setting may be summarized as follows: 1. Can feedback enhance the degree of job satisfaction of workers in an industrial setting? 2. Can feedback as a reinforcer be effectively built into an organization setting to increase performance? 21 3. Can extrinsic and intrinsic reinforcement produce differential effects on performance? 4. Are extrinsic and intrinsic feedback interactive or offsetting in nature? Hypotheses The following hypotheses are primarily derived from the existing research evidence pertaining to the effect of feedback on attitude and behavior. They provide empirical tests for the relationships between different modes of feedback as the manipulated independent variable and the observed effects on the job satisfaction and the performance meas- ure as the dependent variables. Hypothesis 1-a. Workers receiving feedback will have a greater increase in job satisfaction than workers receiv- ing no feedback. Research findings on this issue show conflicting results. Smith and Knight (1959) and Hundal (1969) suggested the degree of attitude change resulting from feedback is not significant. 0n the other hand, several authors (Watson, 1969; Leavitt and Mueller, 1951; Kolb §t_gl,, 1968; Ivancevich g§_gl,, 1970) suggested positive impact of feedback (i.e., extrinsic feedback) on attitude. Similarly Hackman and Lawler (1971) showed that feedback of task, which is intrinsic in nature, is posi- tively related to job satisfaction. This hypothesis, therefore, is derived from the findings of the latter group of research studies. Hypothesis l-b. After the feedback intervention, the degree of job satisfaction for workers receiving extrinsic feedback will be higher than for those receiving intrinsic feedback. No specific studies are available at this time to argue that the effect, if any, of extrinsic feedback on attitude is significantly 22 greater than that of intrinsic feedback. However, it is generally assumed that individuals who receive frequent reports on performance (which is extrinsic feedback in nature) tend to have better attitude (Cook, 1968) and higher need satisfaction of participants (Ivancevich, g§_gl,, 1970), although neither studies made a deliberate attempt to provide intrinsic feedback to the subjects. Hypothesis 1-c. After feedback intervention, the degree of increase in job satisfaction for those receiv- ing the combined feedback (i.e., extrinsic/ intrinsic feedback) will be higher than that of all other groups. It has been suggested that intrinsic and extrinsic reinforcement are closely associated with each other (Bandura and Perloff, 1967; O'Leary, g§_gl,, 1972; Aronfreed, 1968). Extending from this argument, it is hypothesized that there is a combined effect of self and super- visory feedback on job satisfaction. In addition to these hypotheses related to all dimensions of job satisfaction, a following prediction is made regarding a specific dimension of job satisfaction. Hypothesis 1-d. After feedback intervention, the degree of job satisfaction on supervision will increase to a higher degree for those who receive the extrin- sic feedback than for those who receive the intrinsic feedback. This hypothesis on a specific dimension of job satisfaction, supervision, is derived from the finding of Cook (1968) and Ivancevich g§_§l:, (1970) cited earlier. While Cook and Ivancevich, gt_gl,, did not deal specifically with "supervision," it can be argued that periodic feedback on performance and positive verbal reinforcement given by supervisors will affect the worker's perception of his satisfaction with his superior. 23 In addition to the hypotheses pertaining to the effect of feed- back on job satisfaction, the following hypotheses are made regarding the effect of feedback on behavior. Hypothesis 2-a. There will be a significant degree of increase in job performance for those receiving feed- back, after the feedback intervention (i.e., pre-post difference for experimental groups). Hypothesis 2-b. Workers in groups receiving feedback will have a greater increase in degree of job performance than workers in groups receiving no feedback. It has been shown that groups receiving feedback increased perform- ance in problem solving tasks (Smith and Knight, 1959) and in repeti- tive grinding tasks (Hundal, 1969), and that individuals receiving feedback improved performance (Brethower, 1973; Weitz, g£_§l,, 1954), performance quality (Miller, 1965), cost performance (Laird, 1971). It is hypothesized, therefore, that these findings are generalizeable to tasks which are service type in nature. Hypothesis 2-c. There will be no significant difference in per- formance between those receiving extrinsic feed- back and those receiving intrinsic feedback, after feedback intervention. This hypothesis was derived from the findings of Warm gp_al,, , (1972) which indicated that external feedback and self feedback made no significant difference in enhancing vigilance performance. It is hypothesized that this finding is generalizeable to tasks which are service type in nature. Hypothesis 2-d. After feedback intervention, the degree of increase in performance for those receiving the combined feedback (i.e., extrinsic and intrinsic feedback) will be higher thanTTHat of all other groups. 24 While Baron and Ganz (1972) suggested no significant impact of the combined feedback effect on performance among elementary school students, Hamner and Foster (in press) found that the effect of intrin- sic and contingent extrinsic rewards on performance are additive. Additionally, Aronfreed (1968) argued that one without the other is relatively ineffective and inefficient. Hypothesis 3-a. Absenteeism among workers receiving feedback will decrease after feedback intervention. Hypothesis 3-b. After the feedback intervention the degree of decrease in absenteeism for those receiving the combined feedback will be higher than that of all other groups. A number of studies have shown that absenteeism is related to job satisfaction (Vroom, 1964; Talacchi, 1960; Waters and Roach, 1971). Thus, it can be argued that to the extent that feedback affects job satisfaction, absenteeism will be affected, although the feedback to be provided is related to job performance rather than to absenteeism. CHAPTER II METHODOLOGY Introduction In the preceding chapter, literature dealing with reinforcement theory, application of reinforcement theory, effect of feedback on attitude and performance, and hypotheses derived from the literature were presented. This chapter will begin with a brief description of the research design employed in this study followed by a discussion of (l) the independent variables, (2) the dependent variables, (3) the subjects and procedure, (4) the measures used, data collection, and scoring, and (5) analysis of data. Experimental Design The nonequivalent control group design (Campbell and Stanley, 1963) was employed in this study. This experimental design was modi- fied to include a time-series by adding three monthly posttest meas- ures related to performance. This quasi-experimental design, as opposed to a "true" experimental design, typically involves applied settings where it is not possible to control all the relevant varia- bles but only some of them (Isaac, 1974). This study employed this design because random assignment of sub- jects to the experimental groups and the control group was not feasible 25 26 in the research site. Therefore, this design is different from the "pretest-posttest control group design" in which experimental subjects are assigned randomly from a common population to the experimental and the control group. Yet, Campbell and Stanley (1963) have regarded this design as controlling effects which can weaken internal validity, such as history, maturation, testing, and instrumentation, since a control group insures against mistaking these effects for the treatment effects. Since this design lacks complete randomization of subjects, group differences undetected by pretest measures may possibly exist, thereby contaminating posttest data (i.e., selection-maturation interaction, selection testing interaction, etc.). As Stanley and Beeman (1958) and Thorndike (1942) suggest, analysis of covariance techniques can be used to increase internal validity when randomization is not possible. Although the selection of the control group in this study was influ- enced by its apparent higher performance, pretest data did not reflect control group superiority on the two objective task performance meas- ures. Therefore, problems of regression in this design should be mini- mized in this study. As for external validity, a possible problem with this design is interaction of pretesting and treatment; that is, administration of a pretest may influence subjects' awareness of treatment, reducing their representativeness of the unpretested population. Precautions were taken to avoid this effect, as described in the procedure section of this chapter. The experimental design for the study of the feedback effect on job satisfaction is shown in Figure 2. 27 FIGURE 2 EXPERIMENTAL DESIGN FOR THE STUDY OF THE EFFECT OF FEEDBACK ON JOB SATISFACTION Group Period Pretest Posttest* Extrinsic Feedback X11 X14 Intrinsic Feedback X21 X24 REESE}: $22.1ka X31 X34 No Feedback X41 X44 *The posttest measures correspond to the 90 days measures; see Figure 3. For performance variables, three consecutive posttest measures were obtained to detect the experimental effect. Further, where these repeated measures for posttest periods were obtained, the mean of the preceding year served as a pretest measure of these dependent variables for both control group and experimental groups. The experimental design for the study of feedback effect on performance variables is shown in Figure 3. Independent Variable In this study, the different modes of feedback given to three experimental groups and a control group constituted the four levels of the independent variable, feedback. The modes of feedback (i.e., treatments) included: (1) extrinsic feedback, (2) intrinsic feedback, (3) extrinsic and intrinsic feedback, and (4) no feedback. 28 FIGURE 3 EXPERIMENTAL DESIGN FOR THE STUDY OF THE EFFECT OF FEEDBACK ON PERFORMANCE Group Pretest Posttest 30-Days 60-Days 90-Days Extr1n51c Feedback X11 X12 X13 X14 Intr1n51c Feedback X21 X22 X23 X24 Extrin51c and X31 X32 X33 X34 Intrinsic Feedback No Feedback X41 X42 X43 X44 Extrinsic feedback was Operationally defined as having work groups receive information from their foremen as to how many workers in the work group had met the previously determined weekly objective. That is, using the rating forms (see Appendix A), foremen rated their worker's performance in behavioral terms on a daily basis. They pro- vided feedback weekly on worker's performance against the weekly objec- tive. The form used for the feedback is shown in Figure 4. Further, each foreman praised workers on the performance categories which dis— played good performance. This group served as an experimental group with externally reinforced treatment without receiving additional judgment. Intrinsic feedback was operationally defined as having workers rate themselves as to how well they performed on a daily basis, using the same rating forms (see Appendix A). At the beginning of each week, the supervisor set the goal for each work group without providing any 29 FIGURE 4 GROUP FEEDBACK SHEET Date Department Week SAFETY SERVICE COST Objective Actual Objective Actual Objective Actual Excellent Excellent____ ____Excellent ___. Good Good ___. Good ____ Fair ____ ____. Fair _____ ____ Fair .____ ____ Poor Poor ___. Poor ATTENDANCE WORKER COOPERATION SUPERVISOR COOPERATION Objective Actual Objective Actual Objective Actual Excellent Excellent Excellent Good Good Good Fair Fair Fair Poor Poor Poor 30 other feedback on their performance. This procedure allowed the workers to keep a record of his or her own work. (i.e., self-feedback), permitting the worker to gain intrinsic feedback from the task itself (Hamner, l974b). Thus, this group served as an experimental group with intrinsically reinforced treatment without external judgment. Extrinsic and intrinsic feedback was operationally defined as having workers obtain both extrinsic feedback from their foremen and intrinsic feedback from themselves as described above. Thus, at the beginning of each week, each foreman pointed out the discrepancy between self-rating and supervisory rating on performance, and then praised the worker groups on those performance categories which demon- strated good performance. This procedure was repeated on a weekly basis using the same feedback form (see Figure 4). No feedback was defined as having workers under the current prac- tice without introducing any change. Some feedback may have been already provided to workers through regular supervisory practices. The current practice will be described in detail in the procedure sec- tion of this chapter. Dependent Variables Two categories of dependent variables were measured in this study: job satisfaction and performance. Job satisfaction was measured by the Job Description Index (Smith, 33 21,, 1969) and the five single item scale of job satisfaction questionnaire constructed for this study. The Job Description Index (JDI) includes the five dimensions of job satisfaction: work on present assignment, present pay, oppor- tunity for promotion, supervision on present assignment, and people 31 on present assignment. The remaining five single item scale of job sa- tisfaction covers such dimensions as opportunity for independent thought, sense of accomplishment, job security, recognition of good performance and physical conditions in the work place (see Appendix 8). Another dependent variable category deals with performance of the work group. One subjective performance measure on "service" was observed. Service, defined as a performance measure, was obtained by the superior's subjective rating on the quality of building service, maintenance, motor vehicle, and supply services., The actual score ranges from 1 to 100 and was obtained by inspection on specifically described inspection categories on a periodic basis. The 1974 service objective for the work groups under this study was 83 points. (More detailed description of "service" and other performance variables are given in Appendix C). In addition, three objective measures of performance were investi- gated in this study: "cost," "safety," and ”absenteeism." The "cost objective" was computed based on two basic categories: (1) actual dollar expenditures per approved budget assigned to departments, (2) actual dollar expenditures per approved hourly budget assigned to spe— cific labor classes. The actual cost performance for this study was obtained by computing the ratio of forecasted over actual cost. The measure of "safety" was obtained by adding the actual points for such categories as job disability accidents, professional care and restric- tions, duty accidents, motor vehicle accidents, and accident prevention plan. For example, the point for job disability accidents for a given period is a function of the number of lost time injury cases and the 32 total days of absence due to accidents adjusted by the total number of employees for the work group. The 1974 net safety objective for the work groups under this study was 87 points. These three dependent variables - service, cost ratio, and safety - are the major performance measures which are directly related to the tasks involved in the job situation. Finally, the absenteeism objec- tive is obtained by allowing .7 percent for the incidental absenteeism and 4.0 percent for the disability absenteeism for the total working _ hours for the work group. Therefore, the actual measures of these variables reflects the rate of absenteeism of the workers. ‘ "Worker cooperation" and "supervisory cooperation” are two meas- ures created in this study. Therefore, since only posttest data were available, these measures are not included in the findings. Subjects and Procedure The research was conducted in four separate plants within a large midwestern telephone company. The plants are located in a suburban area of a midwestern city. There are approximately 220 workers employed in these plants working on service type jobs. Approximately 25% of employees in these plants are white-collar workers involved in manager- ial and clerical types of work. The remaining employees, who were direct participants in this study, were 37 workers in the extrinsic feedback group, 26 workers in the intrinsic feedback group, 26 workers in the extrinsic and intrinsic feedback group, and 24 workers in the no feedback group (control group). These workers hold one of these five job titles: building equipment mechanic, motor mechanic, building 33 serviceman, cleaner, or stockman. The description of each of these five jobs is summarized in Figure 5. These participants are blue-collar unionized employees with a high school education on the average. About 60 percent of these employees are in the age category between 40 and 60. Approximately 40 percent of these employees are women. The number of foremen who supervise the work group in each plant varies only slightly; 6 for the three experimental groups and 7 for the control group. The number of workers in each work group, however, ranges from 2 to 13. Prior to this experiment, workers were rated by their supervisors in six categories, such as productivity, quality of workmanship, dependability, safety, job knowledge, and customer-employee relations. In each work group, each foreman rated his subordinate on a scale rang- ing from 1 to 5 either by direct observation or by job inspection. The frequency of rating by each foreman varied for individual workers. If workers were rated "below average” (i.e., lower than an average of 3 in a composite score of six performance categories), they were more fre- quently observed and rated. Conversely, if workers were rated "above average," the frequency of observation on these workers for the rating purpose was subsequently decreased. Therefore the frequency of feed— back given to workers varied by their performance. Currently, three measures of performance are available at the plant level: cost, service, and safety. These measures served as the indicators of the relative efficiency in each plant. 0f four plants, three served as experimental groups, and the fourth served as the control group. Before any feedback intervention 34 FIGURE 5 JOB DESCRIPTIONS OF THE SUBJECTS BUILDING MECHANIC - Inspects, maintains, makes repairs, and installs electrical and mechanical equipment including plumbing, ventilating, heating and refrigeration equipment on Company and customer's prem- ises. Makes repairs and alterations to buildings on customer or Company premises including carpentry, plumbing, masonry, plastering, painting, and electrical work. May install, maintain and repair building and equipment hardware. May repair fixtures, furniture, and furnishings including cabinet work and finishing. May do other related assigned work. BUILDING SERVICEMAN - Cleans buildings and central office equipment; services and inspects fire equipment; receives and disposes of sup- plies; operates the heating plant; maintains building grounds and . performs other house service work. GARAGE MECHANIC - Maintains motor vehicles and motor driven tools, 'including inspections, adjustments, repairs and overhauls. Performs other work such as disbursing gasoline and cleaning, lubricating and painting motor vehicles. CLEANER — Does general cleaning work, such as wet and dry mopping; washing and polishing furniture; washing walls; scrubbing stairs; vacuuming rugs; dusting; and picking up waste papers. STOCKMAN - Receives, stores and disburses telephone apparatus and sup- plies. Performs other duties incident to maintaining stockrooms and pole yards such as reconditioning some equipment, disbursing gasoline and taking inventories. 35 took place, a pre-intervention attitude survey was administered to all blue collar workers in the four groups to measure the degree of job satisfaction. This step was introduced in order to detect the experimental effect of feedback intervention once it had taken place. To avoid potential contamination (i.e., interaction of pretest and treatment), the employees were told only that the attitude survey was part of an independent survey which the Department of Management at Michigan State University was conducting. Prior to the experimental treatment, the jobs in each experimental group were described in concrete behavioral terms. (See Appendix A). For example, the task of building equipment mechanic has been broken down into the number of cards reviewed, the number of items incorrectly loaded, the ratio of purchases to inward movement, etc., which consti- tute the "service" category of performance. A composite group score appeared in frequency terms on the weekly feedback form, which is shown in Figure 4. As described earlier, the feedback intervention for the experimen- tal groups occurred for three consecutive months. For the extrinsic feedback groups, each supervisor rated each worker using a rating form (see Appendix A) at the end of each work day. This information was compiled on a weekly basis showing the number of workers who met the weekly objective on the performance categories. At the beginning of each week, the supervisor provided the feedback by announcing the num- ber of workers who had met the weekly objective and then praised workers on those categories which showed good performance. For the intrinsic feedback groups, each employee was given a rat- ing form at the end of each work day (see Appendix A). All employees 36 were required to rate themselves and put the form in a prepared box. At the beginning of each week, the supervisor set the goal for each wo rk group without giving any other feedback on their performance. For the extrinsic and intrinsic feedback groups each supervisor po ‘inted out the difference between the self rating and supervisory Pa “ting at the beginning of each week, and then praised workers on these performance categories which showed good performance. No posi- ti Ve or negative reinforcements were given to those who failed to meet the weekly performance objective. Again, this procedure was repeated on a weekly basis. At the end of the three months of experimental intervention, a second attitude survey identical to the one administered prior to the ex periment was taken to obtain the posttreatment measures on job satis- fa ction. During the experimental period of three months, two employee re signations, three retirements, and eight intra-company transfers took P 1 ace. IL!CILS§_Measures, Data Collection, and Scoring It has been stated that job satisfaction as a dependent variable Ma. 8 obtained through the questionnaire which includes Job Description I rlclex (JDI) and the five single item scale of job satisfaction con- 3 tructed for this study. Each of the five dimensions of job satisfac- ti on in the JDI actually includes 12 response categories, 6 of which were randomly selected in questionnaire construction. In scoring, each 0 1: the item responses for each job satisfaction dimension in the JDI We" 3- assigned either a 1, 2, or 4; the higher the score, the higher the d Q Sree of job satisfaction reflected by that item. The score of the 37 respondent for each dimension was obtained by computing the mean; that is , the sum of item scores divided by the number of items responded. Th'i s procedure was necessary since some respondents failed to complete a1 1 6 items under each dimension. In the other 5 single item scale, each item rated by the individ- ua 1 workers received a score ranging from 1 to 5, a higher score indi- ca ting a higher degree of job satisfaction. In addition, a composite 30 b satisfaction measure was obtained by adding the standardized Z Scores of both the JDI and the remaining 5 items. This procedure was "a cessary because different scoring ranges existed in the two job satis- fa. ction measures employed in this study. The individual job satisfaction measures for the pretest period We re then matched with the posttest measures after the 90 day experi- me htal period, with the exception of those subjects who had been trans- fe rred, had retired or resigned. Table 1 presents the test-retest reliabilities of all ten dimen- S‘i cans of job satisfaction scores and the internal reliabilities of the JD 1 scores. The internal reliabilities were computed by using Cron- ba Ch's Coefficient Alpha (Cronbach, 1951). While an individual worker is the unit of analysis for investigat- ing job satisfaction, a work group is used as the unit of analysis for De ”formance variables. Thus, the performance variables (i.e., cost, 8a 1“Tety, service, and absenteeism) which have served as the indicators OF the relative efficiency at the plant level were broken down for each wo bk group of each plant on a monthly basis for three consecutive peri- °qs ‘ 38 TABLE I TEST-RETEST RELIABILITY AND INTERNAL CONSISTENCY 0F JOB SATISFACTION SCORES .- Jo b Satisfaction Dimensions Test-retest Rel. Internal Rel .* WC.) rk on Present Assignment .26 .76 Pr" esent Pay .41 .81 Op portunity for Promotion .34 .83 Su pervision of Present Assignment .47 . .61 People on Present Assignment .53 .77 op portunity for Independent .43 Thought 3 e nse of Accomplishment .39 J o b Security .26 Re cognition of Good Performance .39 P’ hysical Conditions of Work Place .29 *r 3|: nternal reliability measures were computed based on the pretreatment d ata. A&<‘:filysis of Data Basically two different statistical techniques will be used in the a. r‘Pc‘tlysis of data; analysis of variance with repeated measures and CO Variance analysis. Analysis of variance with repeated measures is an e>oo on“ >2 umumzwum muowcma pcwspmwgppmoa owes» any com cams w;h« NmN.P mmo.~ omo.~ emcmmz umpmzwe< mmp. amp. F¢_. w_. om ems. ems. mqo.F op._ z Pmm. Ns_. saw. mo. am Pwo.P 0mm. umm.~ NR. 2 8mm. NNN. emp. N_. am NNF._ mm_.F asp._ __._ z app. _m_. _oF. N_. am wN_.P NmP.F mep._ NF._ 2 .mxmmumm .mxmmumm .mxmmuqm peasbaacbaca mcowcwa xuaneaaa oz xomnummm uwmcwcch new ommcwcpxm xomncmmu owmcwgucH xoonuwmm owmcwgpxm xumneaaa we owe: =mucmsgomgma “moo: we mcmmz umumzwu< an» egg mcowpow>mo vcmucmpm .mcmmz szpu< N m4mou asp an umpmzwcm mcovcma acmeummcppmoa moss» on» Low mamas mgpee .mucmsgomcma xumemm on» mepmn may .mcoum ms» gmgmw; asks we. Nk.m N~.m NF.m om mm.om mm.~m um.mm mm.mm um.em z mm.m om.“ om.o .NN.N om m_.om op.¢m mm.om mm.uw mm.mn z o_.m wo.¢ mo.¢ mm.m om mm.om ON.Pm mm..m oo.om ou.wa z NF._F m¢.o m~.m eo.N om o~.Pm mm._m mF.Nm om._m mm.- 2 samcaaz .mxmmumm .mxmmumm .mxmmumm “caspaacpaca umumsnu< - mcowcma xomn—uwmn— oz xoonummu owmcwcch new owmcmgpxm xomnummu owmcwcch xomnnmmu uwmcwcpxm xumnummm mo wvoz .s=»pacam= co mama: empmsnn< as“ use mcorumw>ma ucwucmpm .mcamz Fmspu< m m4mmo ULMUcM#m vcm mcmmz w m4m00 000 00 00000000 0000000 0005000000000 00000 000 000 00000 0E.t0 .00000000000 000>000 000 000000 000 .00000 000 000000 000.0 0m.wm om.mm mm.—w m©.mw 000000: 00000000 00.0 00.0 00.0 00.0 00 00.00 00.00 00.00 00.00 0 00.0 00.0 00.0 00.0 00 00.00 00.00 00.00 00.00 0 00.0 00.0 00.0. 00.0 00 00 00 00.00 00.00 00.00 0 00.0 00.0 00.0 00.00 00 00.00 00.00 00.00 00.00 0 .mxmmuqm .mxmmumm .mxmmumm 000000000000 0000000 00000000 oz 00000000 000000000 000 0000000xu 00000000 000000000 00000000 0000000xm 00000000 00 00oz 0=0u0>000= 00 00000 00000000 000 000 0000000>00 00000000 .0000: 00:00< NF m0m<0 57 scores for all four groups were below the 1974 service objective. While the no feedback group showed better performance in service than the other three groups, there was no statistical difference between the groups (F = 1.83, p 5 .17). At the end of the experimental periods, service performance of the intrinsic feedback group as per- ceived by superiors was still below the service objective for 1974, while the no feedback group showed a considerable improvement in the same category. Table 13 presents the summary of the analysis of variance with repeated measures on service. It reveals that there were signifi- cant main effects of both mode of feedback and time period. Addi- tional analysis on the effect of time period indicates that the dif- ference in “service" between the pretreatment measures and the three posttreatment measures combined was statistically significant (F = 45.78, p 5 .001). This finding supports the stated hypothesis that there would be a significant increase in job performance for those receiving feedback. For the differential effect of feedback mode on service the planned comparisons summary of covariance analysis are presented in Table 14. When the posttreatment measures (i.e., the three post- treatment measures combined) were adjusted by the procedure of co- variance analysis, it was found that the control group was signifi- cantly higher in service performance than the three experimental groups combined as seen in comparison 1 in Table 14. This finding is contrary to the prediction. TABLE 13 Analysis of Variance with Repeated Measures Table on Service for Mode of Feedback and Period 58 Within Groups (Error) 20 15.548 Source of Variance df MS F p- Between Mode of Feedback (A) 3 276.66 5.32 .007 Error 21 51.98 .0012. Period (8) 3 527.47 29.27 .001 A X B 9 20.60 1.14 ns Error 63 18.02 TABLE 14 Planned Comparisons Summary of Covariance Analysis of Service for the Posttreatment Periods Source of Variance df MS F p - Between Groups Comparisons: 1 1 73.483 4.726 .041 2 1 8.657 .556 ns 3 1 100.985 6.495 .019 59 Further, there was no difference in service performance between the combined feedback group and the two other feedback groups (ex— trinsic and intrinsic feedback) combined. Yet, there was a signifi- cant difference in service between the extrinsic feedback group and the intrinsic feedback group, indicating inferior service ratings by superiors on their subordinates in the intrinsic feedback group. These findings are not in the predicted direction. To summarize the above findings and the related discussions of the four performance measures, a summary table showing the overall results on the performance data is presented in Table 15. In gener- al, two objective performance measures (cost performance and safety) and one subjective performance measure (service) indicated signifi- cant improvement after feedback was introduced. Also, those worker groups receiving feedback as compared to the control group showed significantly better performance in ”cost” and “safety“. However, there were no significant differences in performance among the groups receiving different types of feedback. Further, "absenteeism” seemed to be unaffected by the feedback on worker's performance, although the findings can be regarded as a relative improvement. Job Satisfaction In analyzing the data on each job satisfaction dimension, the same statistical techniques were used as had been for the performance measures; that is, the analysis of variance with repeated measures for the pre-post difference and covariance analysis for the group differences. As discussed previously, job satisfaction was measured at two different time periods: prior to the experimental treatment 60 00000 00000000 0000000x0 000 00000> 00000 00000000 000000000 000 00 00000000 00000000 000000000 000 000 00000 00000000 0000000x0 000 0:000> 00000 00000000 00000500 000 00 00000000 000000 0000050000x0 00000 000 0:000> 00000 0000000 000 ”0 m0o. v 0 m 00 N 000. v 0 000>00m :00. v a _. 00 m 00 0 00 00000000000 00 0 00 m 00 0 000. v 0 , 000000 who. v a _. 00 m 00 0 00. v 0 00005000000 0000 moo.v 0 0 00000000< 0000000>oov 0000000: 0000000mv 000000000000 00000 0000000000 0000-000 0000000: 00000000000 0000000< 0000000>oo 000 0000000: 00000000 0003 0000000> 0o 0000000< 00 00000002 00000000000 00 00000000 00 0000500 m0 m0m<0 6l and at the end of the treatment period (90 days after the treatment started). It was predicted that there would be a significant increase in job satisfaction after feedback intervention occurred, as compared to the pretreatment measures. It was further predicted that there would be a differential effect of feedback on job satisfaction, hypothesizing that the combined feedback group would show higher increase in job satisfaction than other groups. Additionally, it was hypothesized that the extrinsic feedback group would show higher job satisfaction than that of the intrinsic feedback group. Table 16 presents the summary of findings on each of the ten job satisfaction dimensions investigated in this study. As this table shows, in six out of the ten dimensions of job satisfaction, no statistically significant difference existed between the pretreat- ment and the post-treatment score of job satisfaction. Further, when the pretreatment differences were adjusted, no significant differences in job satisfaction between different groups were shown in any dimensions except "job security,“ which was decreased after feedback intervention. 0f the remaining four dimensions of job sat- isfaction which showed significant pre-post difference, "present pay” and "job security" were decreased, while "people on present assign- ment" and "supervision of present assignment" were increased after feedback intervention. In addition, Table l6 revealed that feedback on worker's perform- ance produced no impact on job satisfaction dimensions that were di- rectly related to work content (i.e., intrinsic job satisfaction) such as "work on present assignment,” "opportunity for independent thought”, 62 .00000000 000000 00000000000x0 00000 000 00000> 00000 0000000 000* m: m: a. S. .v. n— m: m: m: m: w: m: m: «0000000< 00000000000 0000000000 00000 mo. mo. 00. mo. VI VI VI V. m: m: m: m: m: m: QQQO. .100000002 0000000mq 0000000000 0000-000 0000000000 0000000 00 00000000000 000000000< 0000000 00 000000 00000000 000 000 0000000 000000000 000 00000000000 00000 0003 00 0000000000 00000000 00000000000 0000 000 00000000000 0000000000000< 00 00000 0000000 00000000000 000 00000000000 000000000< 0000000 00 0003 0000000000 000000000000 000 0000000< 0000000000 000 0000000: 00000000 0003 0000000> 00 m0m napmsmeg w3~m> Pmpoe LO msomcefipmumwz mcmpemso mov>cmm mwoo» w mmWFmamw LDLOLO NNNN mesa w .xwa acme: WUCWPm cmeoacma a mauaem .msoecea (\J eczocw w cowcmuxw mfiwa w mcoxuoo .cczd m “Ema .mcooa .m_~e% e w m— NP. 0 o mEoocgmmz w w mcoo_d mccwpmm a mmwcomwpmu corpomamcm \ \.. .\ \ c \ .- m. \J. SE“. \\ \ aged 02%” ompxos\me_n a \ acw_ ecm\oca\me_n _ . .w.z a .m.m “new sea a .a.: a .m.m aged >qsm L as mass m :mEmcod mop: mupm mrsvmgom cowpomameH mePezo FIGURE A2 RATING SUMMARY SHEET BUILDING MAINTENANCE INSPECTION gucoz mesh mzovpomamcH mo .02 xmncH her—mzo aspe> Boomshea m:_m> ~mpoe mF NF m m «wcwpmmz m m a m N Fmoficpo arm s m e m N 83%: or m o a chzpochm cowcmucH or m m a mceucvma cowcmpxm m a m N mucsocw my NF m o cowpmpwpcw> w .ncoo cwq OF w *0 a mecca me N_ m e meeocema coacaoxm or m o e m mCFuc ma co? .mp:_ m cwowm r1 mmwsommpmu cow“ mmmca a .m5 mw nu mm H" n, n, mw \wu “w 44 .nL hop m mm mm” WW mm mm” .MW mm” Aw WW MW mw MW ms case 0 ,w W W w u... 1 v m \m. mm. 8: ENSURE: . a a w . r u ....mzammoesmzo > \ \\ \cecm mcpfi pm_\oe\mure _ H \\ .m. s w .m. m mead >osm Q \ \ \ 432:: N \ \ \ \ \ :95.»th mop; 33m m_:nm;om cowpomgmcH xpwrmso GURE A3 r... F MOTOR VEHICLE INSPECTION RATING SUMMARY SHEET cwmgp m>wga4 van m+wo .mchbw (Y) (‘0 T mm_amu a Ngmppmmm cowpwucou mc%c::m wcwmcm mgmxuwpm Loom «mvgoumm mD34 LO Q' d” m .m.z w .m.m pgzm >wm saga mCWF pm~\ E\;m> N .m.: w .m.m gag; >Qsm \ x3\:m> m- cagm ugcw Luz mwsnmsom cowpquWCH mpwpmzo spcoz mFgF mcowpumamcH we 02 mchoa FmpoF ngcFu .mpcmo .Fmsm mummm w mgmz LooFm goFEmgcq (\J cmmflu .meamm mangzo .>mga LOLOLO m .w .m.m “mam >Fo PmmchcmW < . .m.s gag; ocFF me\oe\;m> N n m 22;; >nsm mwm n. a I w 2. pm.\..}w.\% : \ \r I x \ merm >quznm wzHH ache: Fn_p:ouv mq umpru \ .WC,” w t x3\£m> m- csgm pacm Luz onvmgom coFFuwamcH mpFFozo FIGURE A4 SUPPLY SERVICE AND QUALITY INSPECTION RATING SUIMARY SHEET :Fcoz mFgF conumamcF Fo Lanazz xmwcF muFFmso msFm> vmpmsmmm msFm> FmFOF p:m5m>o:_ngw2cF Top mmmmgugnm mo oFme NF m xmucF muF>me xFQgsm NF oF pcsou xuagsuom cgsme OOOO :oFFFucou gmcFmpomLmo :Egpmm ncm: co mpmm FmF Eoogmgowm Fo COFFFUCOU :ou umumoF waumggoo -cF mEmFF Fo Lwnemm m NNNCD OOOO mvmoF F.cou co .m.o mampF mo Lm3532 oF LO 0 .xw noumoF mFFungou -cF mampF mo gmmgsz 0F c mnmoF mocmguxm co oF OOCOCD £0 0 nmsz>mi mugmu .XQE .cF: mmFLommqu coFFquWCF Ly “W. m? ..a mF NW mu 9 J. 4 MW 9 J. m inn IU _n AM a s \ 1: FA my F \ F F x .0 AW cELF mcFF UCN\LFU\ELFW F .m.z a .m.m Fasm >Fo :ELF mcFF me\oE\ELFm F .w.z a .m.m cELL >uzm x2\ELFm F sag; mmFFaasm szumzom soFFummwcF NMFszo APPENDIX B ATTITUDE SURVEY APPENDIX B ATTITUDE SURVEY This is part of a survey which the Department of Management at Michigan State University is currently conducting. Primarily, the purpose of this survey is to investigate the attitude of industrial workers in the Detroit area to their present job. This questionnaire has been approved by the management of Michigan Bell. Information obtained from you will be completely confidential. Please do not put your name on the questionnaire. Once you have completed your questionnaire, place it in the self-addressed envelope and place it in the prepared box. Thank you for your cooperation. l. Age 2. Sex 3. Education: Elementary School ( ) Middle School ( ) High School ( ) Community College ( ) Technical Training ( ) 4 Year College ( ) Others ( ) 4. Number of years with Michigan Bell _m __yr(s). 5. Job Title 6. District: Trinity-Townsend ( ) Royal Oak-Pontiac ( ) Supply System ( ) Southfield—Birmingham ( ) Place a check in the space that reflects your degree of satisfaction on the following items most closely. 7. The opportunity for independent thought in doing your job. Very TT' Satisfied NeitherTSatisfied Dissatisfied Very Satisfied Nor Dissatisfied Dissatisfied 8. The feeling of worthwhile accomplishment at work. Very ' Satisfied Neither Satisfied Dissatisfied Very Satisfied Nor Dissatisfied Dissatisfied 9. The degree of job security of your current position. Very Satisfied Neither Satisfied Dissatisfied Very Satisfied Nor Dissatisfied Dissatisfied l0. The awareness that others have of your performance when you perform well. Very Satisfied Neither Satisfied Dissatisfied Very Satisfied Nor Dissatisfied Dissatisfied ll. The physical conditions of your working location. - Very Satisfied Neither Satisfied Dissatisfied Very Satisfied Nor Dissatisfied Dissatisfied Think of your present work. What is it like most of the time? In the blank beside each word given below write: __1___ for “Yes" if it describes your work .._N__. for "No" if it does NOT describe it __;:__ if you cannot decide T2. WORK ON PRESENT ASSIGNMENT Boring Useful ______Pleasant Tiresome A__Challenging - Routine l3. PRESENT PAY _. Income adequate for Less than I deserve normal expenses Highly paid _fl_Barer live on income ______Under paid “_ Bad T4. OPPORTUNITIES FOR PROMOTION Good opportunity for advancement Promotion on ability Dead-end assignment 15. SUPERVISION ON PRESENT ASSIGNMENT Asks my advice Praises good work " Doesn't supervise enough 16. PEOPLE ON YOUR PRESENT ASSIGNMENT Boring Ambitious Responsible Unfair promotion policy Infrequent promotion Opportunity somewhat limited Tells me where I stand Impolite Around when needed Talks too much Lazy Unpleasant APPENDIX C DESCRIPTIONS ON PERFORMANCE OBJECTIVES APPENDIX C DESCRIPTIONS ON PERFORMANCE OBJECTIVES COST The cost objectives have been established in two basic categories: 1. Actual dollar expenditures, per the approved budget, are computed and assigned to specific departmental codes. This computation is made daily and summarized on a weekly, monthly, quarterly and yearly basis for the following departments: 1. Building Service 2. Building Maintenance 3. Motor Vehicle 4. Supply Service 2. Actual hour expenditures, per the approved hour budget are computed and assigned to specific labor classes. This computation is also made daily and summarized weekly, monthly, quarterly and annually for the follow- ing worker groups: Cleaner Building Serviceman Building Mechanic Motor Vehicle Mechanic Stockman 01th—4 SAFETY The safety objective is to minimize the accidental injuries to employees by identifying a lack of safety knowledge and employee aware- ness of environmental hazards. To measure performance trends, several formulas have been estab- lished in four major categories: 1. Job Disability accidents: 24 points a) Points for No. of Cases = 16 - No. of Lost Time Injury Cases X 320 Total No. of Employees b) Points for Days of Absence = 8 - Total Days of Absence X l6 Total No. of Employees T974 objective: 22 points out of maximum 24 points Professional Care and Restrictions Duty Accidents: T0 points Points for No. of Cases = TO - “No. of Injuries X 40 Total No. of Employees T974 objectives: 7.5 points out of maximum TO points. Motor Vehicle Accidents: 24 points a) Points for No. of Vehicles = l2 - No. of Accidents X 3 Total No. of Vehicles b) Points for Miles Driven = l2 - No. of Accidents X 25,000 Total No. of Miles Driven l974 objective: l7.0 points out of the maximum 24 points. Accident Prevention Plan: 42 points The 42 points allotted for this category is computed by utilizing the AT&T minimum objective of 96% (per the T00% scale) in the following formula: Points = T.4 (Avg. Score per the AT&T plan -67) T974 objective: 40.5 points The T974 Safety Objective is 87 points out of a possible perfect score of TOO points. ATTENDANCE The attendance objective is to improve, through the process of identification, action, and follow-ups, our overall attendance results per worker, per Foreman group: Formula: Incidental # of employees X total work days X 0.7 = objective Disability # of employees X total work days X 4.0 = objective Total Incidental + Disability = objective SERVICE' The service objective is to maintain the quality of building serv- ice (housekeeping), building maintenance, motor vehicle and supply services at a specified objective level. Actual manager and/or worker inspections are utilized to gain the necessary data on a periodic basis. Service Quantitative Category Objective Building Services 83 Building Maintenance 83 Motor Vehicle 83 Supply Services 83 Thus, by computing the average score, in the categories for which each manager or foreman is responsible, the service rating for any given period can be determined. Building Service Foremen, Building Maintenance Foremen, Motor Vehicle Foremen and Supply Foremen utilize the appropriate inspection form (See Appendix A) to compute their service ratings as follows: Each manager observes worker's performance or inspects jobs which have been completed, and: l. Enters the total point value in the appropriate column of categories or classifications of items inspected. 2. Summarizes the maximum value points for the items inspected. 3. Divides the total point value of the inspection by the maximum point value of the items inspected. This equates to the adjusted point value index. 4. Enters the adjusted point value index in the appropriate box on the inspec~ tion sheets. This now becomes the quality rating. 5. The total point value inspected divided by the summarized maximum point value of the items inspected equals the adjusted quality results rating index. When summarizing inspections for the week or month, on the rating summary sheet, the managers followed the same procedures and formula as outlined above. APPENDIX D FINDINGS ON SIX JOB SATISFACTION DIMENSIONS APPENDIX D FINDINGS ON SIX JOB SATISFACTION DIMENSIONS WORK ON PRESENT ASSIGNMENT TABLE DI MEANS AND STANDARD DEVIATIONS SUMMARY TABLE OF JOB SATISFACTION WITH ”WORK ON PRESENT ASSIGNMENT” (N=TI3)* Treatment Periods Pretreatment 90—days M 2.90 2.54 Extrinsic Feedback S.D. l.22 1.35 E0d3b3f. M 2.81 2.76 ee “CK Intrinsic Feedback S.D. l.26 l.25 Extrinsic and M 2'89 2'72 Intrinsic Feedback S.D. 1.24 1.12 M 2.97 3.02 No Feedback S.D. 1.23 1.06 *The higher the value, the better the job satisfaction; the maximum of 4 points. TABLE DZ ANALYSIS OF VARIANCE WITH REPEATED MEASURES TABLE FOR ”WORK ON PRESENT ASSIGNMENT" Source of Variance _gf__ _M§_ _Ji_ _J3§_ Between Mode of Feedback (A) 3 .7728 .4031 ns Error 109 1.9174 mm Period (8) 1 1.2935 1.1475 ns A X B 3 .4679 .4151 ns Error 109 1.1271 TABLE 03 PLANNED COMPARISONS SUMMARY OF COVARIANCE ANALYSIS ON ”WORK ON PRESENT ASSIGNMENT" FOR THE POSTTREATMENT PERIOD Source of Variance _£H:_ _jfii_ _j:_ _Jg;_ Between Groups Comparisons: 1 1 2.152 1.524 ns 2 1 .1174 .0832 ns 3 1 .9025 .6394 ns Within Groups 20 1.4115 (Error) ”OPPORTUNITY FOR INDEPENDENT THOUGHT" TABLE D4 MEANS AND STANDARD DEVIATIONS SUMMARY TABLE OF JOB SATISFACTION WITH ”OPPORTUNITY FOR INDEPENDENT THOUGHT” (N=113)* Treatment Periods Pretreatment _j¥§g§QgL_ M 3.91 3.83 Extrinsic Feedback S.D. .98 .92 Mode of Feedback M 4.15 4.00 Intrinsic Feedback S.D. .88 .84 Extrinsic and M 3'69 3'88 IntrinSic Feedback 3.0. 1.15 .90 M 3.75 3.96 No Feedback S.D. .89 1.08 *The higher the value, the better the job satisfaction; the maximum of 5 points. TABLE D5 ANALYSIS OF VARIANCE WITH REPEATED MEASURES TABLE FOR ”OPPORTUNITY FOR INDEPENDENT THOUGHT" Source of Variance _Ji£_ _Jfii~ _J:_ .IES. Between Mode of Feedback (A) 3 .8029 .6048 ns Error 109 1.3275 mm Period (8) 1 .0398 .0748 ns A X B 3 .4636 .8704 ns Error 109 .5325 TABLE 06 PLANNED COMPARISONS SUMMARY OF COVARIANCE ANALYSIS ON ”OPPORTUNITY FOR INDEPENDENT THOUGHT" FOR THE POSTTREATMENT PERIOD Source of Variance df MS Between Groups Comparisons: 1 1 .3149 2 1 .2265 3 l .0665 Within Groups 20 .7310 (Error) .4307 .3098 .0910 )< 11$ ns 1'15 "SENSE OF ACCOMPLISHMENT" TABLE D7 MEANS AND STANDARD DEVIATIONS SUMMARY TABLE OF JOB SATISFACTION WITH "SENSE OF ACCOMPLISHMENT" (N=Tl3)* Treatment Periods Pretreatment 90-day§__ M 3.75 3.91 Extrinsic Feedback S.D. 1.21 .95 Mode of Feedback M 3.73 3.83 Intrinsic Feedback S.D. .91 .95 Extrinsic and M 3'6] 3’80 IntrinSic Feedback S.D. 1.13 .80 M 3.87 3.91 No Feedback S.D. .94 .82 *The higher the value, the better the job satisfaction; the maximum of 5 points. ANALYSIS OF VARIANCE WITH REPEATED MEASURES TABLE DB TABLE FOR "SENSE OF ACCOMPLISHMENT" Source of Variance _df__ _M§__ _j:_ _J$;_ Between. Mode of Feedback (A) 3 .3037 .2233 ns Error 109 1.3599 mm Period (8) 1 1.1324 1.8792 ns A X B 3 .0543 .0902 ns Error 109 .6026 TABLE 09 PLANNED COMPARISONS SUMMARY OF COVARIANCE ANALYSIS ON "SENSE OF ACCOMPLISHMENT" FOR THE POSTTREATMENT PERIOD Source of Variance df Between Groups Comparisons: 1 2 3 Within Groups (Error) 20 MS .0038 .0545 .0102 .6835 .0056 .0798 .0149 .122. ns ns "RECOGNITION OF GOOD PERFORMANCE" TABLE 010 MEANS AND STANDARD DEVIATIONS SUMMARY TABLE OF JOB SATISFACTION WITH "RECOGNITION OF GOOD PERFORMANCE" (N=TI3)* Treatment Periods Pretreatment 90-days M 4.00 4.00 Extrinsic Feedback S.D. 1.02 .94 Mode of Feedback M 3.73 3.57 Intrinsic Feedback S.D. .91 1.13 Extrinsic and M 3’76 3'80 IntrinSic Feedback S.D. .90 .89 . M 3.54 3.75 No Feedback S.D. 1.14 1.11 *The higher the value, the better the job satisfaction; the maximum of 5 points. TABLE 011 ANALYSIS OF VARIANCE WITH REPEATED MEASURES TABLE FOR ”RECOGNITION OF GOOD PERFORMANCE" Source of Variance _gf__ _M§_ _J:_ _Jg;_ Between Mode of Feedback (A) 3 1.7380 1.2352 ns Error 109 1.4071 w Period (B) l .0177 .0279 ns A X B 3 .2766 .4361 ns Error 109 .6342 TABLE 012 PLANNED COMPARISONS SUMMARY OF COVARIANCE ANALYSIS ON ”RECOGNITION OF GOOD PERFORMANCE" FOR THE POSTTREATMENT PERIOD Source of Variance df MS F gs Between Groups Comparisons: 1 1 .0460 .0513 ns 2 l .0147 .0164 ns 3 1 1.5447 1.7252 ns Within Groups 20 .8953 (Error) "PHYSICAL CONDITIONS IN WORK PLACE" TABLE D13 MEANS AND STANDARD DEVIATIONS SUMMARY TABLE OF JOB SATISFACTION WITH ”PHYSICAL CONDITIONS IN WORK PLACE" (N=113)* Treatment Periods Pretreatment 90-days M 3.91 3.83 Extrinsic Feedback S.D. .75 .92 Mode of Feedback M 3.80 3.50 Intrinsic Feedback S.D. .80 1.17 Extrinsic and M 3‘69 3'69 IntrinSic Feedback S.D. 1.08 .88 M 3.50 3.66 No Feedback - . S.D. .97 1.30 *The higher the value, the better the job satisfaction; the maximum of 5 points. TABLE 014 ANALYSIS OF VARIANCE WITH REPEATED MEASURES TABLE FOR "PHYSICAL CONDITIONS IN WORK PLACE" Source of Variance df _M§__ _F__ _p;k. seesaw. Mode of Feedback (A) 3 1.0063 .8014 ns Error 109 1.2557 B13319. Period (B) 1 .2167 .3117 ns A X B 3 .4895 .7040 ns Error 109 .6953 TABLE 015 PLANNED COMPARISONS SUMMARY OF COVARIANCE ANALYSIS ON ”PHYSICAL CONDITION" FOR THE POSTTREATMENT PERIOD Source of Variance df MS F as Between Groups Comparisons: 1 1 .1207 .1146 ns 2 l .0581 .0552 ns 3 1 1.3656 1.2956 ns Within Groups 20 1.0539 (Error) ”OPPORTUNITY FOR PROMOTION" TABLE D16 MEANS AND STANDARD DEVIATIONS SUMMARY TABLE OF JOB SATISFACTION WITH ”OPPORTUNITY FOR PROMOTION" (N=113)* Treatment Periods Pretreatment 90-day§~_ M 2.16 2.15 Extrinsic Feedback S.D. 1.25 1.20 Mode of Feedback M 2.46 2.53 Intrinsic Feedback ' S.D. 1.21 1.28 Extrinsic and M 2‘09 2'25 IntrinSic Feedback S.D. 1.22 1.19 M 1.99 2.24 No Feedback ' S.D. 1.12 1.14 *The higher the value, the better the job satisfaction; the maximum of 4 points. TABLE D17 ANALYSIS OF VARIANCE WITH REPEATED MEASURES TABLE FOR ”OPPORTUNITY FOR PROMOTION" Source of Variance _d:_ MS __5_ u_g§; Between Mode of Feedback (A) 3 1.6300 .8357 ns Error 109 1.9503 W Period (B) 1 .6370 .6536 ns A X B 3 .1765 .1811 Ans Error 109 .9745 TABLE 018 PLANNED COMPARISONS SUMMARY OF COVARIANCE ANALYSIS ON "OPPORTUNITY FOR PROMOTION” FOR THE POSTTREATMENT PERIOD Source of Variance Between Groups Comparisons: 1 2 3 Within Groups (Error) .511. ‘4 MS .0095 .0005 1.1305 ( 1.3070 .0073 .0004 .9108 BS. MS 1'15 113 IIIIIIIIIIIIIIII 111111111111111111111111151111