THE EFFECTS OF CHOICE AND‘ COMMITMENT 0N ATTITUDE CHANGE AND PRODUCTIVITY GAIN Thesis for the Degree of M. A. MICHIGAN STATE UNIVERSITY GERALD WINN CARD 1969 ssssss IIIIIIIIIIIIIIZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 310438 9121 ' ABSTRACT THE EFFECTS OF CHOICE AND COMMITMENT ON ATTITUDE CHANGE AND PRODUCTIVITY GAIN By Gerald Winn Card According to dissonance theory, variations in choice and/or commitment lead to Variations in the arousal and magnitude of cognitive dissonance. While in most dissonance experiments subsequent pressures to reduce dissonance are used in predicting attitude change, this study predicted both attitude and behavioral change on the basis of hypoth— esized attempts at dissonance reduction. In addition, the choice and commitment variables were manipulated outside the laboratory in an industrial setting. Following pretest attitude and productivity assessments, A6 female key—punch operators were exposed to departmental letters within the context of interviews introducing two methods of objectively evaluating individual productivity within their department. Half of the employees were allowed to choose between the two methods, while half were assigned one or the other. In addition, half of each of the above groups were told that they would learn their productivity Gerald Winn Card scores in private, while half were told that their scores would be presented in public. The pretest productivity scores were presented a week later, followed immediately by an attitude posttest and a perceived choice assessment. All subjects were told at this time that productivity score presentations in the future would be based on actual productivity. Sub- sequent score presentations were made for productivity periods of two weeks and a calendar month. Dependent variables in this study were (1) attitude change toward both the rating methods and the individual's Job, and (2) productivity gain, for the two periods of productivity observation. It was predicted that individuals in the Choice- Private condition would perceive greater "freedom of choice" in the experimental situation than would individuals in the other three conditions, and would, in turn, experience greater attitude change. The productivity gain hypothesis predicted that individuals in the public commitment condi- tions would increase productivity more than would individuals in the private commitment conditions. While analysis of variance supported the perceived choice hypothesis, the attitude change hypothesis was not supported. The productivity gain hypothesis was supported for the first productivity period of two weeks, but not the second period of one month. Overall, or grand mean productivity for the second period of productivity observation, Gerald Winn Card however, was statistically significant when compared with the pretest productivity grand mean. The failure to achieve significance with the attitude change inducements was viewed as a joint function of a "group retaliation" effect (wherein several subjects, influenced by one disgruntled subject's comments, reaCted negatively against the experiment by checking the attitude posttest scales in a consistently negative manner), and a posttest subject deletion which resulted in identical pre- and posttest attitude scores for two experimental groups. These two factors seriously distorted the attitude change data, and made an unambiguous interpretation of the obtained results virtually impossible. While the productivity increases during the two week productivity period were statistically significant as predicted, it appeared that a productivity ceiling had been attained within the public conditions, so that additional increases during the calendar month productivity period within the public conditions were unlikely to occur. Informal interview data also suggested that over the term of the experiment, interaction between experimental subjects resulted in the diffusion of performance information along lines other than those artificial channels of the experi- mental design. This would have tended to build motivational pressures withing the private conditions, thus leading to the productivity increases within the private conditions that vitiated the results for the calendar month productivity period. THE EFFECTS OF CHOICE AND COMMITMENT ON ATTITUDE CHANGE AND PRODUCTIVITY GAIN By Gerald Winn Card A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Communication 1969 ACKNOWLEDGMENTS I would like to express my most sincere appreciation to Dr. Erwin P. Bettinghaus, for his advice and counsel over the past few years, and his guidance and assistance in the development and execution of this study. I would also like to thank Mr. Robert A. Schlappi, Supervisor of Data Processing Operations, Oldsmobile Divi- sion, General Motors Corporation, for his efforts that made both this study and my graduate education possible. 11 TABLE OF CONTENTS ACKNOWLEDGEMENTS LIST OF TABLES . . . . . . LIST OF APPENDICES . . . . . . . . . . Chapter I. INTRODUCTION . . . . . . . . . . . Previous Research . . . . . . . Rationale and Hypotheses . . . . II. DESIGN AND PROCEDURES Independent Variables . . . . . . . Dependent Variables . . . . . . . . The Sample . . . . . . . . Procedure of the Experiment . . . . . Statistical Analyses III. RESULTS . . . . . . . Randomness . . . . . . . . . . Perceived Choice . . . . . Induced Choice and Attitude Change . . Commitment and Productivity Gain . IV. DISCUSSION AND CONCLUSIONS . . . . Perceived Choice . Induced Choice and Attitude Change Publicness of Score Presentations (Commitment) and Productivity Gain Summary of Experimental Findings . . . . REFERENCES . . . . . . . . . . . . APPENDICES . . . . . . . . . . . . . . iii Page ii iv 22 28 30 31 32 33 AA AA 147 A8 149 EA 5“ 55 6O 66 69 72 Table 10. 11. 12. 13. LIST OF TABLES Comparison of Pretest Attitude Scores by Treatment Group (Phase I) . . . . . Summary of Analyses of Variance on Pretest Attitude Scores (Phase I) . . . Comparison of Pretest Productivity Scores by Treatment Group (Phase I) . . Summary of Analysis of Variance on Pretest Productivity Scores (Phase I) . . . . Comparison of Perceived Choice Scores by Treatment Group (Phase III) . . . . . Summary of Analysis of Variance on Amount of Perceived Choice (Hypothesis I) . Comparison of Attitude Change Scores Toward Two Concepts by Treatment Group (Phase III) Summary of Analyses of Variance on Attitude Change Scores (Hypothesis II) . . Comparison of Posttest Productivity Change Scores by Treatment Group (Phase IV) . . Summary of Analysis of Variance on Posttest Productivity Change Scores (Hypothesis III) Comparison of Posttest Productivity Change Scores by Treatment Group (Phase V) . . Summary of Analysis of Variance on Posttest Productivity Change Scores (Hypothesis III) Comparison of Departmental Grand Mean Produc- tivity by Experimental Period . . . . iv Page A5 A5 A6 A6 A7 A8 50 51 51 52 52 53 LIST OF APPENDICES Appendix A Pretest Attitude Instrument Bl-6 Experimental Messages C Posttest Attitude Instrument D Pretest Departmental Productivity Distributions . . . . . . Page 73 88 95 99 CHAPTER I INTRODUCTION Lower level industrial employees normally have little voice in the organizational decision-making process. How- ever, the employee does have attitudes and opinions about many changes in the organizational environment, especially as concern him personally; attitudes and opinions that may be crucial in the success or failure of organizational plan- ning. I If we assume that many organizational changes are neg- atively evaluated by lower level employees when introduced, it would seem that several rather interesting predictions concerning both acceptance of the change and related behavior could be made based on Festinger's (1957) theory of cognitive dissonance. Dissonance theory is concerned with the antece- dent conditions leading to the arousal of dissonance (a pSychological tension with motivational characteristics), and various means of dissonance reduction. The most investigated variables relating to the arousal of dissonance have been choice and commitment, while the means of dissonance reduc- tion most investigated has been attitude or opinion change. Other means of dissonance reduction, however, including behavioral change, might also be predicted on the basis of attempts at dissonance reduction. The dissonance formulation and the choice and commit- ment variables may have some very interesting "low cost" operational implications as concerns the introduction of or- ganizational change. For example, dissonance theory predicts that the amount of choice an individual perceives in a choice between alternatives should influence subsequent evaluations of the alternatives involved in the choice. The question thus becomes whether the change situation can be structured as a choice situation. If so, perhaps one could predict greater liking for, or acceptance 03 the change. On the basis of the dissonance research, one could also predict that the greater the commitment to information which is dis- crepant with one's self-concept, the greater the arousal of dissonance. To the extent that behavioral change can in fact be predicted on the basis of attempts at dissonance re- duction, it would seem that if the unpleasant or discrepant information were concerned with individual performance or productivity, changes in performance or productivity could also be predicted. The present study is an attempt to test choice and commitment hypotheses based on Festinger's theory of cogni- tive dissonance in the introduction of change in performance evaluation in a highly programmed industrial department. The experimental setting is as follows: two rating methods designed to objectively evaluate individual productivity are to be introduced into the key-punch section of a large cor- poration's data processing department. Half of the depart- mental employees will be given a choice between the two methods; i.e., will be allowed to choose the method by which they will be evaluated, while half will be assigned one or the other methods. In addition, half of each of these two groups will learn their subsequent productivity scores in private, half in public. The dependent variables in this study are (l) attitude change toward the rating methods and the individual's job, and (2) productivity gain. The independent variables are choice and commitment, operationalized as follows: (1) Choice—Private: Presentation of a departmental letter within an interview introducing the rating methods, with employees in this condition being allowed to choose between the two methods, and their sub- sequent performance scores to be presented in private. (2) Choice-Public: Same as (1) above, with the - exception that the score pre- sentations are to be made in public. (3) Denied Choice- Presentation of a departmental Private: letter within an interview in- troducing the two rating methods, but assigning one of them, with subsequent performance scores to be presented in private. (A) Denied Choice- Same as (3) above, with the Public: exception that subsequent scores are to be presented in public. It is hypothesized that attitude change toward both the rating methods and the individual's job will vary directly with the subjective perception of choice in the situation, with the greater increases occurring within the Choice—Pri- vate condition (condition 1, above), while productivity gain will vary directly with the publicness of subsequent score presentations (commitment), With the greater increases occur- ring within the public score presentation conditions (condi- tions 2 and A, above). Previous Research According to Festinger (1957), dissonance is the exis— tence of non—fitting relations between cognitive elements, which he describes as "any knowledge, opinion, or belief about the environment, about oneself, or about one's behavior (p. 3)." Specifically, "two elements are in a dissonant re- lation if, considering these two alone, the obverse of one element would follow from the other (p. 13)." Two basic hypotheses are offered: The existence of dissonance, being psychologically uncomfortable, will motivate the individual to try to reduce dissonance and achieve consonance. When dissonance is present, in addition to trying to reduce it, the person will actively avoid situa- tions and information which would likely increase the dissonance (p. 3). Much of the dissonance research may be viewed from the standpoint of the effects of choice on attitude change. The dissonance research generally supports the notion that the amount of choice an individual perceives in a decision between two or more alternatives should influence subsequent evaluations of the choice alternative. A variable closely related to choice in the dissonance research is that of commitment. As Brehm and Cohen (1962) have pointed out, most of the empirical investigations have been concerned with the "special case" of commitment. How- ever, as defined by Brehm and Cohen, commitment appears to be more or less synonomous with choice or decision: "a per- son is committed when he has decided to do or not to do a certain thing, when he has chosen one (or more) alternatives and thereby rejected one (or more) alternatives, when he actively engages in a given behavior or has engaged in a given behavior (Brehm and Cohen, 1962, p. 7)." Commitment has not, however, been used extensively in the dissonance research to describe or explain a behavioral process. Perhaps the most parsimonious approach to review- ing the relevant dissonance literature is to subsume commit- ment under the rubric of choice. For the purposes of this review, it will be assumed that once a choice or decision has been made, commitment occurs, and one may expect to see maniiestations of attempts at dissonance reduction. While sucri an approach is not without its negative aspects (see, fkn' example, Deutsch and Krauss, 1965, pp. 73-75), it would appear to be consistent with the manner in which commitment has been treated in the literature. For the most part, the dissonance research supports the notion that the greater the commitment to unpleasant or discrepant information or behavior, the greater the arousal of cognitive dissonance, and the greater the consequent attempts at dissonance reduction. Choice and commitment, then, may be viewed as very important variables in the dissonance research. As the hypo- theses of this study are directly concerned with the choice and commitment variables, the review of the research evidence will center around the effects of choice and the effects of commitment in various dissonance producing situations. Cohen, Terry, and Jones (1959) tested a choice in ex- posure hypothesis by asking male undergraduates in a high choice condition three times if they wished to hear the experimentor read a counter-attitudinal message to them con- cerning the advisability of young men marrying before the age of 23. In a low choice condition, subjects were merely told that the experimentor was going to read the contrary informa— tion to them. The results of this experiment generally indicated that among extreme subjects, high choice produced greater change toward the position advocated in the communi— cation than did low choice. Choice, however, had little effect on those who were initially moderate toward the issue. An experiment by Davis and Jones (1960) suggests that a choice or commitment must be perceived as relatively irrevocable before dissonance reduction and attitude change occurs. In this experiment, half of the male subjects were assigned the task of reading an unfavorable evaluation to an unseen stimulus person, while half were persuaded to do so (i.e., they could give either a positive or negative evalua- tion, but what the experimentor really needed was a negative evaluation). In addition, within each of these two groups, half of the subjects were told that immediately following the reading they would be allowed to meet the recipient and be able to explain why it had been given. The rest were told that there would be no opportunity to retract their state- ments. As predicted, the amount of negative change toward the message recipient was greater in the choice-irrevocable condition than in the other three conditions. However, as nearly one—quarter (12 of 52) of the subjects in this experi- ment were eliminated for various reasons, the obtained results should probably be accepted with some reservation. The following two studies investigated choice under the rubric of "justification for compliance." The dissonance prediction would be that the greater the justification used to secure compliance, the less the arousal of dissonance and consequent attitude change. Cohen, Brehm, and Fleming (1958) operationalized high justification (low choice) by giving 92 male undergraduate subjects several reasons why they should write a counter-attitudinal essay (e.g., it would help the experimentor, and the results of the survey would be of interest to both social scientists and the university authori— ties). In the low justification condition (high choice), subjects were given only the necessary instructions. While an analysis using all subjects did not result in significantly different attitude change in the two conditions, when those subjects with original extreme attitudes were deleted from analysis and the hypothesis retested, the difference level reached the .07 level. The authors report that the failure to achieve significance may have been due to the fact that all subjects wrote essays, so that there was a generally com- pelling force to comply in all conditions. Rabbie, Brehm, and Cohen (1959) manipulated the inde— pendent variable in much the same manner as did the previous study. In the high justification condition, the writing of a counter-attitudinal essay concerning the elimination of intercollegiate athletics was strongly justified by the ex— perimentor. In the low justification condition, the experi— mentor barely justified writing the essay. Subjects in the high justification condition reported significantly greater feelings of obligation (less choice) than did subjects in the low justification condition (more choice). When atti- tudes were measured immediately after the decision to take the discrepant position, as well as after the essay had been written, subjects in the low justification (high choice) condition showed significantly more change than did subjects in the high justification (low choice) condition. The issue under consideration in this study was apparently highly salient, as A6 percent of the subjects contacted refused to write the counter-attitudinal essay. Unfortunately, the authors report no data for these subjects. External justification may take many forms. The fol- lowing two studies successfully manipulated the amount of monetary reward in two forced compliance situations as a determinant of subsequent attitude change. Festinger and Carlsmith (1959) tested the hypothesis that "the greater the pressure used to elicit overt behavior, the weaker will be the tendency for the individual to change his opinion to bring it into correspondence with what he has said or done." Pressure was manipulated by varying the amount of reward offered to male subjects to take part in a very boring and tedious task of an hour's duration, and then tell the next experimental subject (a confederate of the experimentor) that the task was, in fact, enjoyable. Subjects in one con- dition were given twenty dollars as a reward (low choice), while subjects in another condition were given one dollar (high choice). As predicted, subjects in the high choice (one dollar) condition rated the experimental tasks signi- ficantly more enjoyable than did those subjects in both a control group and the low choice (twenty dollar) condition. Cohen (1959) obtained similar results in an experiment using rewards graduated from $.50, $1.00, $5.00, and $10.00, for writing an essay contrary to private views on a current" campus issue. The results of this experiment, as in the Festinger and Carlsmith study, showed a consistent inverse 10 relationship between the amount of reward and attitude change. In addition, a measure of perceived volition ob— tained by Cohen generally indicated that volition decreased as the inducing force increased; although there was a slight reversal in the high dissonance ($.50) condition. As convincing as these two studies appear, both have been recently re-examined due to possible alternative explana- tions for the results. Janis and Gilmore (1965) suggest that an "incentive" effect might have accounted for the results of the Festinger and Carlsmith study, while Rosenberg (1965) posits that Cohen's results were artifactual, and due to negative affect and evaluation apprehension. However, the "replications" of these latter experimentors are not exact replications, and so the theoretic controversy continues (see, for example, McGuire, 1966). The following two studies investigated the choice variable in a rather different context; that of selecting between relatively attractive articles as payment for parti— cipating in what might be termed "consumer research." Brehm (1956) had 225 female college students rate eight $15 to $30 manufactured articles on an eight—point scale as to desir- ability. Choice in this study was manipulated by allowing certain subjects to choose between a specified two of the articles with the impression that they could keep the arti- cle chosen as payment for participating in the research. In a no-choice, or "gift" condition, the experimentor chose an 11 Iarticle previously rated as highly desirable by the subject (either 5, 6, or 7 on the eight-point scale) and gave it to her as a gift. Attitude change was measured by comparing pre- to posttest desirability rating changes for both chosen and unchosen articles. In support of the revaluation of alter- natives hypothesis (one of three tested), the results indi- cated that chosen articles increased in desirability from the pre- to posttest period, while similar articles in the no- choice or "gift" condition did not. Brehm and Cohen (1959) tested the hypothesis that re- valuation of alternatives following choice will increase with an increase in the number and qualitative dissimilarity of the choice alternatives. Two hundred and three sixth-grade children rated 16 toys on the basis of liking both before and after being allowed to choose one toy from a selected group as a reward for participating in the research. Half chose from two toys, half from four. In addition, half of each of these groups chose from qualitatively similar toys, half from dissimilar toys. The authors report that significantly greater attitude change toward the chosen alternative occurred in the four alternative than in the two alternative condition, and in the dissimilar rather than the similar condition. In general, chosen toys became more liked, unchosen toys became less liked. While both of these studies suggest that postdecisional reorganization of choice alternatives can occur, a large 12 number of subjects were deleted from the overall analyses in each study. In the Brehm study, for example, A8 subjects reversed their choice from pre— to posttest, and an addi- tional 30 subjects' ratings were such that the pre-determined dissonance conditions could not be created. In the Brehm and Cohen study, 130 of the total 203 subjects were deleted from analysis for various reasons! Subject loss in this experiment becomes crucial when the authors state in a foot- note that for the entire sample, only the similar-dissimilar manipulation had any effect. Due to such inordinately large subject losses, the results of these experiments should pro— bably be viewed with some apprehension, especially those conclusions dealing with other than simple revaluation of alternatives following choice. In another "free choice" experiment, Deutsch, Krauss, and Rosenau (1962) tested the hypothesis that "a chooser will experience post-decisional dissonance only when he perceives his choice in a given situation to be inconsistent with the conception of himself which he tries to maintain (for him— self or for others) in that situation." Their subjects, 58 paid undergraduate volunteers, rated six bread spreads on a nine-point scale in terms of flavor and overall preference. Two spreads that had been rated close to midscale were then selected for each subject and he was asked to choose one of the two as a gift. Half of the high self-involvement group was told prior to choosing that a person's ability to judge l3 subtle differences in the quality of foods was closely related with leadership aptitude, executive potential, and artistic judgment. The other half of the high self-involve- ment group first made their choices and then received the A above inducement. A low self-involvement group, on the other hand, received no inducement. All subjects then re— rated all six spreads. The results of this study indicated that the amount of change toward the chosen spread, from pretest to posttest, was significantly greater for the high self—involvement group than for the low self-involvement group in terms of flavor ratings, but not for overall preference ratings. The two high self-involvement groups (inducement before-choice and inducement after-choice) did not differ significantly. As no subjects were eliminated from this study, the disson- ance theory prediction that revaluation of choice alterna— tives occurs following a choice or decision receives some fairly strong support. The fact that only the flavor ratings and not the over— all preference ratings increased from pre— to posttest would appear to be directly related to the involvement inducement in the above study. As the choice involved appears to have been rather trivial, it would seem that only the involvement inducement made the choice sufficiently important so that rating changes occurred. The importance of a decision or commitment would appear to be directly related to the magnitude of the dissonance 1A aroused in the situation. Festinger (1957) has advanced an hypothesis concerning this relationship: "the magnitude of the dissonance . . . which exists between two cognitive elements will be a direct function of the importance of the two elements (Festinger, 1957, p. 262)." Thus, variations in personal involvement should have implications for both the arousal and magnitude of cognitive dissonance, as well as for attempts at dissonance reduction. The remaining stu- dies to be reviewed are all concerned with rather direct manipulations of importance or involvement. And, while there may or may not have been a meaningful prior choice manipula- tion, it is the variation in commitment that is of major importance. Zimbardo (1960) had female college students appear in friendship pairs for a study dealing with how friends diag- nose and evaluate social issues. All subjects were asked to read a case study on juvenile delinquency and then determine the locus of the blame for the crime described in the study. Half of the subjects were told that it would be impossible to learn anything from their reactions to the case (low in- volvement), while half were told that their reactions to the case would indicate their basic social values, personalities, and outlook on important life problems (high involvement). After having written their opinions of the case study, subjects were asked to judge delinquents from a group of eight photographs; a task which would measure their "judg- mental and perceptual" abilities. Each subject was then told 15 that her friend had rated all eight of the pictures correctly, but that she had rated only five of the eight correctly. Sub- jects were then told of their friend's ratings of the case study, under either high or low discrepancy conditions between friend and subject. The results of this study indicated that the high self- involvement group (high commitment), which reported stronger feelings of involvement than did the low self-involvement group (lower commitment), showed significantly more change toward the positions of their friends than did the low self- involvement group. Aronson and Mills (1959) induced dissonance in female subjects by requiring them to take an "embarrassment" test in order to join a series of group discussions on the psy- chology of sex for which they had volunteered. In a mild initiation condition (low involvement, hence low dissonance), subjects were required to read five rather mild sex related words to the experimentor; in a severe initiation condition (high involvement, hence high dissonance), subjects were required to read a list of twelve obscene words and two descriptions of sexual activity to the experimentor; while, in a control condition, subjects were merely told that they would become members of the discussion group. Following these "embarrassment" tests, all subjects listened to a very boring tape recorded discussion on animal sexual be- havior by the group they were supposed to join. After hear- ing the tape, subjects were asked to indicate both how good 16 they thought the discussion group was, and how much they liked the members of the group. The results of this experi- ment indicated that the more embarrassing or painful the test required to join the group, the more the discussion group and the participants were rated more favorably. Brehm (1960) had subjects fill out a questionnaire estimating the vitamin content and liking for a disliked vegetable, and then meet with the experimentor to eat a portion of the vegetable. Commitment in this study was varied by having the subject eat either a little or a lot of the vegetable. After eating, subjects read a report giving fic- ticious information either supporting or not supporting par- ticular attributes of the vegetable eaten. After reading, a questionnaire was again administered. High eating (high commitment) subjects were significantly different from low eating (low commitment) subjects in mean change of estima— tion of vitamin content of the vegetable. Liking for the vegetable, however, increased to approximately the same extent for both high and low commitment subjects. Cohen and Brehm (reported in Brehm and Cohen, 1962, pp. 206—210) report on yet another experiment varying the amount of personal involvement. Thirty undergraduate fra— ternity pledges were asked by their pledge—masters to report individually for a short project taking some fifteen to twenty minutes. The experimentor, a professor not connected with the fraternities, demanded, however, that they sign up 17 to c0py random numbers for from three to four hours. Three conditions of coercion were used: high, where subjects were threatened with strong efforts to keep them out of their fraternity houses; low, where they were threatened with extra hours of pledge duty; and control, where the subjects were simply asked to cooperate. After this manipulation, each subject was given a schedule sheet on which to indicate his "free hours." Perceptions of the coercive force, perceived choice in participation, evaluation of the number copying task, and amount of personal annoyance were measured in a questionnaire given immediately after the above commitment to participate was obtained. The high coercion group felt that they had been most strongly threatened, had the greatest amount of choice or volition, and were most strongly disposed toward the task. On the basis of other dissonance research, how- ever, one would predict that favorability toward the task should decrease as threat is increased. The authors con- clude that the increase in perceived choice accompanying increased threat or coercion was due to the perception of the threat involved as being illegitimate, thus demonstrat- ing, supposedly, that volition in itself is a key determi- nant of the magnitude of the dissonance aroused. One might ask, however, if the illegitimate coercion in this experiment was conceived before the experimental manipulations took place. If so, then perhaps the authors' .2... Id: II‘IDtp . — . 18 position is valid. If not, then the experiment demonstrates that certain ancillary manipulations do not always have their desired effect. However, as the group perceiving the greatest freedom of choice also was most favorably disposed toward the number writing task, this study does provide sup- port for the effects of choice and commitment on dissonance arousal and reduction. An experiment by Brehm (1959) investigated the effects of a fait accompli on attitude change. Brehm had subjects (eighth-grade students) rate 3A vegetables as to liking, and then three weeks later, offered them two movie tickets or two phonograph records of their choice if they would eat a portion of a disliked vegetable. Subjects were free to refuse if they chose. In the low consequence condition, the subject merely ate the vegetable and then filled out the questionnaire again. In the high consequence condition, when the subject had nearly finished eating the vegetable, he was told that his parents would receive a letter from the experi- mentor informing them of the subject's behavior; the implica— tion being that the subject would have to eat more of the disliked vegetable at home. Favorable attitude change toward the previously dis- liked vegetable was significantly greater for the high con- sequence group than for the low consequence group. Apparent— ly the parents' knowledge of the subject's discrepant behavior increased the need to justify having been persuaded 19 to eat the vegetable for those in the high consequence condition. The question of "who knows," or the publicness of behavior or commitment has not received a great deal of attention in the dissonance research. One could argue, perhaps, that all behavior in the experimental setting is somewhat public, as the experimentor is normally present and aware of what has taken place. While perhaps a moot point, one could additionally argue that observation by the experi- mentor may be a critical factor in understanding many of the "non-obvious" dissonance theory results. If we assume that others' knowledge of one's discre- pant behavior should enter as a cognition against that behavior, particularly when such behavior reflects adversely on one's self-image, it would follow from dissonance theory that the arousal of dissonance should be greater under public rather than private conditions. In this vein, Cohen, Brehm, and Latane (1959), replicating an earlier (1957) study by Festinger, varied the publicity attached to subjects' per— formance in a gambling experiment. Public condition subjects were told that both their choice of sides and their winnings in the low stake card game would be published in the school newspaper. Private condition subjects were told that their choice of sides and winnings would be kept confidential. As in the Festinger study, the dependent variable was the amount of time, following the twelfth hand, spent looking at a graph supposedly showing the true probabilities of winning. 20 The graph, a different one for each side, gave the false information that the chosen side was the losing side. It was assumed that small winners and losers would spend only a moderate amount of time viewing the graph, as it would confirm their behavior, while heavy losers were expected to ignore the graph, as it would show that they had been wrong in their choice of sides in the game. The data generally confirm this, with tendencies toward selective exposure and avoidance most pronounced under public rather than private conditions. Chapanis and Chapanis (196A) direct a number of cri- ticisms toward the Festinger experiment and the Cohen g£_a1. replication. One criticism has to do with what they consider to be Festinger's overly elaborate interpretations of his results. According to the Chapanis's, had Festinger's pattern of results been exactly reversed, his explanations would still apply! Their most important criticism, however, is that the two experiments were not so much experiments in postdecision processes as they were experiments in predecision processes. Subjects had been told that they could change sides and were, in fact, given an opportunity to do so when presented with the graphs. While Festinger and Cohen e£_a1. reported that many subjects did announce at this time their decision to change sides, the Chapanis's claim that what was not reported was the number of subjects who looked at the graphs ._.. ... .- ...>u III. V 21 in order to reach a decision whether or not it would be more profitable to change sides. The Chapanis's conclude: "tak- ing all of these factors into consideration, we are forced to conclude that . . . the successful replication of the experiment suggests-—not that the cognitive dissonance form— ulations are valid, but only that the results of experiments of this type are reproducible (Chapanis and Chapanis, 196A, pp. 10—11)." The criticisms advanced by the Chapanis's are well taken. However, as they point out, Festinger's explanations are both elaborate and persuasive. Perhaps on this basis, plus admiration for Festinger's rationale, the Festinger experiment and the Cohen, Brehm, and Latane replication are included at this review as offering at least general support for the dissonance theory prediction tested. The above two experiments, as well as the Brehm (1959) study could additionally be criticized for failing to demon- strate that choice had any effect at all! There was no variation in choice in the Brehm study (i.e., all subjects complied), and the Festinger and Cohen et_al. studies can be understood on other than a "choice of sides" basis. Given the importance variation in the Brehm study (eating more of the disliked vegetable at home) and the Cohen, Brehm, and Latane study (publicity), one could speculate that greater commitment alone would account for the obtained results; perhaps irrespective of there having been a prior choice manipulation. 22 While many of the individual dissonance experiments have been criticized for a number of reasons, taken as a whole, the evidence from these experiments generally support the notion that variations in choice and/or commitment lead to variations in the arousal and magnitude of cognitive dissonance. And, while most of the dissonance experiments have dealt with attitude change as the dependent variable, other means of dissonance reduction, including behavioral change, could certainly also be predicted on the basis of hypothesized attempts at dissonance reduction. Rationale and Hypotheses The experimental setting of this study is as follows: Objective productivity rating methods are to be introduced into the key-punch section of a large corporation's data processing department. One theoretic question of interestis whether choosing between the rating methods has an effect on favorable evaluation of the rating methods or the indi- vidual's job? On the basis of the research cited, the attitudinal hypothesis of this study is that it does. More specifically, it is hypothesized that: H Individuals making a decision between two productivity rating methods with subsequent private score presentations will perceive greater freedom of choice than either those making a similar decision with subsequent public score presentations, or those who are assigned a rating method; regardless of the type of score presentation (C Pr > C Pb, DC Pr, DC Pb). 12 As concerns the subjective perception of choice, the litxarature is fairly explicit. If choice is denied, or no 23 choice is made, there should be little or no perception of choice. On the other hand, of those who do choose between rating methods, only those with subsequent private score presentations should perceive themselves as having much choice in the introduction of the methods. This would be due to the relative lack of external pressures within the private score presentation conditions. Owing to the external pressures involved in the public score presentation condi- tions, those making a decision with subsequent public score presentations should feel that the choice is less theirs to make, and due more to chance factors. The second attitudinal hypothesis to be tested is: H2: Individuals making a decision between two productivity rating methods with subsequent Drivate score presentations should change more favorably in attitude toward both the rating methods and their jobs than either those making a similar decision with public score presentations, or those who are assigned a rating method; regardless of the type of score presentation (C Pr > C Pb, DC Pr, The basis for this hypothesis is the dissonance theory prediction that both arousal of dissonance and subsequent attitude change are directly related to perceived choice. As the Choice-Private group is expected to perceive the greatest choice in the experimental situation, this group is also expected to experience the greatest attitude change. The productivity gain hypothesis to be tested is as follows: H3: Individuals who learn their performance scores under public conditiOns will increase 2A productivity more than those who learn their performance scores under private conditions (C Pb and DC Pb > C Pr and DC Pr). Objectively, the performance information to be presented is the same for all experimental groups. That is, having randomized subjects to treatments, one's best bet would be that the groups are equal (within random error expectations) on the variable under consideration. It is difficult to ima- gine that a prior choice of rating methods wou1d have much effect on subsequent productivity increases. The type of commitment made, however, imposes a subjectivity bias on the performance information. In dissonance theory terms, others' knowledge of one's possible poor performance should enter as a cognition against that level of performance, and one should be motivated to reduce this dissonance by increasing produc- tivity. In addition, learning the scores of others' should present a standard of comparison for evaluating one's own performance in the public score presentation conditions. Of course, these factors do not apply in the private score pre- sentation conditions. A key factor behind this hypothesis was the finding by this investigator during the pretest productivity assess- ments that nearly all employees exceeded the productivity levels that IBM Corporation and local management regarded as "good" performance by an employee. In fact, the producti- vity levels considered "good" were substantially below the departmental individual productivity means or averages! Thus, as employee performance had been reinforced at a very 25 low level in the past, the performance feedback based on actual departmental performance should be viewed as extremely discrepant by most employees. And, of course, learning one's performance score in the presence of others should accentuate the impact of this information. There is a possibility that the productivity gain hypo- thesis may appear to be inconsistent with the dissonance research cited. It was stated earlier that commitment has normally been subsumed under the rubric of choice in the literature, and there is thus little direct evidence for the efficacy of commitment on the arousal of cognitive dissonance. Yet, in the productivity gain hypothesis, no mention is made of the choice variable as being important. Is this consis- tent with the theory and research? The score presentation conditions in the present study can easily be viewed as examples of involuntary exposure situations, in which case, one could evoke Festinger's failure to mention choice as necessary for the arousal of dissonance in the accidental or involuntary exposure situa- tion. On the other hand, Brehm and Cohen maintain that some prior volition is necessary before the involuntary exposure situation is viewed as a dissonance producing situation. The position taken, which is consistent with the Brehm and Cohen View, is that the employees have made a relevant choice in having chosen to produce at a certain level in the past. Thus, past volition could have prevented the unpleasant consequences viewed as an integral part of the feedback conditions. 26 To test the above hypotheses, four ways of presenting departmental letters within individual interviews were developed: (1) Choice-Private: Presentation of a message offering a choice between 'two objective productivity rating methods, with the employee's subsequent scores under the chosen method to be presented in private. (2) Choice—Public: Same as (1) above, with the exception that the employee's subsequent scores under the chosen method would be pre- sented in public. (3) Denied Choice-: Presentation of a message Private mentioning both methods, but assigning one of them, with the presentation of the em- ployee's subsequent scores to be in private. (A) Denied Choice—z Same as (3) above, with the Public exception that the employee's subsequent scores would be presented in public. If the choice and commitment inducements of this study are successful, one would predict: The Choice-Private group should perceive greater "freedom of choice" than the other three experi— mental groups in the introduction of the rating methods, as indexed by the average sum of scores on a seven—point perceived choice scale (Hypothe- sis 1). The perceived choice differences in Hypothesis 1, above, should be reflected in attitudinal differ- ences; i.e., the pre- to posttest differences in the sum of scores on four seven-point semantic differential scales for the Choice—Private group should be statistically greater than for any of the other three groups (Hypothesis 2). 27 The public score presentation groups should increase productivity more than the private score presentation groups, as determined by a comparison of productivity difference scores at the end of each period of productivity observation (Hypothesis 3). CHAPTER II DESIGN AND PROCEDURES This study was designed to assess (1) change in atti- tude toward two job related concepts and (2) productivity gain, under four conditions of choice and commitment. A two-by-two factorial design was used in this investigation, with the hypotheses being: (1) attitude change would vary. directly with the amount of choice perceived by subjects in a decision between two productivity rating methods, while (2) productivity gain would vary directly with the publicness of subsequent score presentations (commitment) under the applicable rating method. A five phase before-after design was used. In Phase I (pretest), subjects' attitudes toward two experimental concepts were assessed, while independently, pretest produc- tivity figures were extracted from extant records and used to compute departmental and individual productivity norms. Phase II took place eight weeks later, and consisted <1f exposing each subject to both an interview and a depart- nmnital letter from local management introducing two methods Of‘ C Pb, DC Pr, DC Pb). Table 5 Comparison of Perceived Choice Scores by Treatment Group (Phase III) Treatment Group Measure C Pb C Pr DC Pb DC Pr Perceived Choice Mean 2.100 A.A00 2.300 2.A00 S.D. 1.792 1.075 2.58A 1.647 As predicted, the Choice-Private group perceived greater freedom of choice than did any of the other three treatment groups. Analysis of variance of the perceived choice scores (Table 6) showed that the main effect of pri- vateness of commitment was significant at the .05 level of confidence, while the interaction between choice and commit- ment was nearly so at the .08 level. With greater experi— mental precision, one might also expect the interaction A8 between choice and commitment to have reached significance. The confirmation of the perceived choice hypothesis would indicate the success of the experimental manipulations in creating the differential perception of choice posited as a necessary condition for the arousal of cognitive disson- ance in the attitude change portion of this study. Table 6 Summary of Analysis of Variance on Amount of Perceived Choice (Hypothesis 1) Source of Probability of Variation df SS MS F F Statistic Perceived Choice Rows (Choice) 1 8.100 8.100 2.355 .13 Columns (Commitment) 1 1A.A00 1A.AOO A.l87 .05 Interaction 1 12.100 12.100 3.518 .08 Error 36 123.800 3.A38 Total 39 158.A00 III. (a) Induced Choice and Attitude Change The attitude change hypothesis tested in this study was merely a restatement of perceived choice hypothesis (H1): H2: Individuals making a decision between two productivity rating methods with subsequent private score presentations will change more favorably in attitude toward both the rating methods and their job than will those making a decision with subsequent public score pre- sentations, or those who are assigned a rat- ing method; regardless of the type of score presentation (C Pr > C Pb, DC Pr, DC Pb). A9 In short, one would expect the perceived choice differ- ences presented in Table 5 to be reflected in differential mean attitude change scores in Table 7. An examination of the mean attitude change data in Table 7 shows that while the Choice-Private group experienced the greatest positive attitude change toward the rating method concept, and the least negative change toward the job concept, neither of these differences were statistically significant (Table 8). Table 7 Comparison of Attitude Change Scores Toward Two Concepts by Treatment Group (Phase III) Treatment Group measure 0 Pb c Pr DC Pb DC Pr Mean attitude change Mean —.700 -.500 -l.100 -2.A00 toward job concept S.D. 1.567 3.2A0 1.792 2.675 Mean attitude change Mean —l.800 1.000 -2.700 -1.800 toward rating method S.D. 6.729 6.815 6.86A 5.308 concept Table 8 Summary of Analyses of Variance on Attitude Change Scores (Hypothesis II) Source of Probability of Variation df SS MS F F Statistic Job Concept Rows (Choice) 1 13.225 13.225 2.268 .lA Columns ~ (Commitment) 1 3.025 3.025 .518 .A8 Interaction 1 5.625 5.625 .96A .33 Error 36 209.775 5.830 Total 39 231.775 50 Table 8. Continued Source of Probability of Variation df SS MS F F Statistic Rating Method Concept Rows (Choice) 1 3A.225 3A.225 .819 .37 Columns (Commitment) 1 3A.225 3A.225 .819 .37 Interaction 1 9.025 9.025 .219 .65 Error 36 1503.300 Al.758 Total 39 1580.775 III (b). Commitment and Productivity Gain The productivity change hypothesis tested for both the Phase IV and Phase V productivity periods predicted that: H3: Individuals who learn their performance scores under public conditions will increase produc- tivity more than those who learn their perform— ance scores under private conditions (C Pb and DC Pb > D Pr, DC Pr). It had been suggested that public commitment condition subjects would be motivated to increase productivity in order to reduce the dissonance engendered in the public score pre- sentations. Such motivational pressures should be minimal in the private commitment conditions, so that productivity increases over and above random error expectations would not be expected to occur within these conditions. Phase IV Results.--As predicted (Table 9), the public commitment condition subjects increased productivity to a greater extent during the two week Phase IV period than did the private commitment condition subjects. Analysis of the productivity change scores (Table 10) yielded a 51 significant F ratio of A.633 for the main effect of public- ness of commitment. This statistic was significant at the .0A level of confidence, thus supporting the public commit- ment hypothesis for the Phase IV productivity period. Table 9 Comparison of Posttest Productivity Change Scores by Treatment Group (Phase IV) Treatment Group Measure C Pb C Pr DC Pb DC Pr Productivity Change Mean .A32 .007 1.013 .208 Scores (Phase IV) S.D. .750 .A82 1.276 .711 Table 10 Summary of Analysis of Variance on Posttest Productivity Change Scores (Hypothesis III) Source of Probability of Variation df SS MS F F Statistic Phase IV Productivity Change Rows (Choice) 1 1.376 1.376 1.887 .18 Columns (Commitment) 1 3.397 3.397 A.633 .OA Interaction 1 .32A .32A .A33 .52 Error 32 23.A67 .733 Total 35 28.567 Phase V Results.——The data presented in Tables 11 and 12, based on productivity change for the Phase V productivity 52 period, fail to support the commitment hypothesis. As a comparison of the mean productivity change scores by experi- mental period shows (Tables 9 and 11), the public commitment condition subjects maintained the same relative level of productivity during Phase V as for Phase IV, but private commitment condition subjects increased productivity during the Phase V period. This mean increase in the private con- ditions negated the significant main effect of publicness of commitment for the Phase V period. A tenable explanation for this mean increase is offered in the following chapter. Table 11 Comparison of Posttest Productivity Change Scores by Treatment Group (Phase V) Treatment Group Measure C Pb C Pr DC Pb DC Pr Productivity Change Mean .Al7 .AA2 1.121 ,A18 Scores (Phase V) S.D. .826 .7A5 1.353 .550 Table 12 Summary of Analysis of Variance on Posttest Productivity Change Scores (Hypothesis III) Source of Probability of Variation df 88 MS F F Statistic Phase V Productivity Change Rows (Choice) 1 1.0AO 1.0A0 1.23A .27 Columns (Commitment) 1 1.003 1.003 1.226 .28 Interaction 1 1.188 1.188 1.A09 .2A Error 32 26.969 .8A8 Total 35 30-231 53 Overall Productivity Change.-—Whi1e the predicted Phase V productivity change was not statistically significant, a comparison of the overall, or grand mean productivity scores (Table 13) shows that overall mean productivity for the Phase V period was statistically significant at the .01 level (by t test) when compared with the overall or grand pretest productivity mean. Table 13 Comparison of Departmental Grand Mean Productivity by Experimental Period Experimental Period N Mean . SD Obtained t Pr. of 3 Phase I 36 .0AA 1.0A7 —-- ' —-- Phase IV 36 .A60 1.16A 1.56A .06 Phase V 36 .6AA 1.011 2.A19 .01 CHAPTER IV DISCUSSION AND CONCLUSIONS This study was designed to determine the effects of (1) freedom of choice in a decision between two productivity rating methods on attitude change toward both the rating methods and the individual's job and (2), publicness of subsequent score presentations (commitment) on productivity gain. Four choice and commitment conditions were estab- lished in which choice or commitment was varied. The speci- fic hypotheses tested will be discussed in turn in the following pages. Perceived Choice The perceived choice hypothesis tested was as follows: H1: Individuals making a decision between two productivity rating methods with subse- quent private score presentations will perceive greater freedom of choice than either those making a decision with sub- sequent public score presentations, or those who are assigned a rating method; regardless of the type of subsequent score presentation (C Pr > C Pb, DC Pr, DC Pb). The data presented in Chapter III indicate support for this hypothesis. The Choice-Private group's perceived choice rmean was A.A0, or about two scale points (on a 0-6 scale) gzneater than any of the other three groups, with the main 5A 55 effect of privateness of commitment statistically signifi— cant at the .05 level of confidence. There were no real differences between the other three groups; all fell at the lower end of the same scale, with the group means ranging from 2.10 to 2.A0. The confirmation of the perceived choice hypothesis would indicate the success of the experimental manipulations in creating the differential perception of choice posited as a necessary condition for the existence of cognitive dissonance in the attitude change portion of this study. Induced Choice and Attitude Change The attitude change hypothesis tested in this study was merely a restatement of the above perceived choice hypothesis: H2: Individuals making a decision between two productivity rating methods with subsequent private score presentations will change more favorably in attitude toward both the rating methods and their jobs than will those making a decision with subsequent public score pre- sentations, or those who are assigned a rat— ing method; regardless of the type of score presentation (C‘Pr > C Pb, DC Pr, DC Pb). This hypothesis was based directly on the dissonance trueory generalization that attitude change varies directly Imith the subjective perception of choice. It had been (expected that subjects in the Choice-Private condition, gnarceiving the greatest freedom of choice, would also eaqaerience the greatest dissonance, and would change atti- tnides in order to reduce it. While the Choice-Private grmNJp did indeed perceive the greatest freedom of choice, 56 the predicted attitude change did not occur. While there are undoubtedly many theoretically based explanations for the failure of the differential perception of choice to be reflected in attitudinal differences, the following commen- tary will be based primarily on an event that took place during the Phase III portion of this study. It is this ex- perimentor's position that other explanations are of secondary or even tertiary importance in comparison. The event of interest is termed a "group retaliation" effect, and had its genesis in the following manner. During the course of the Phase III manipulations (presentation of initial productivity scores, followed by the attitude post- test), a member of the Denied Choice-Private group (Employee A) was told that her initial score under the applicable method was -.l60 (z score). She became visibly upset and rather hostile, but completed the attitude scale checking, and then returned to her work area and went on a break with a number of other employees. The Phase III score presenta- tions and attitude posttests continued for some ten minutes, and were then halted due to the lack of available subjects. A few minutes later, a number of employees returned from their break and the Phase III manipulations continued. During these later treatments, no fewer than five sub- jects made direct reference to the experimentor that Em- ployee A had been very threatened, and had made some very disparaging remarks about the study and the organization. As these five subjects were among the last to be tested, and 57 their posttest instruments were readily available, the experimentor was able to verify their attitudinal responses. Four of the five subjects (Employee A's group leader, one of the five, showed only minor change) checked the four scales under the "Rating Method" concept consistently nega- tive as follows: Employee B: Checked each scale as zero. Net change in attitude toward the concept: -20. Employee C: Checked all but one scale as zero. The other was checked as 1. Net change in attitude: -19. Employee D: Checked all scales as neutral, or 3. Net change in attitude: —12. Employee E: Checked all scales at neutral, or 3. Net change in attitude: -8. While these employees may have been retaliating against the experimentor for upsetting an informal work group mem- ber, it is also possible that they were protecting Employee A. That is, by their all checking the scales in the same manner (consistently negative, as also had Employee A), it would be impossible to single out any one employee for criticism. It is also quite possible that subjects other than these four were decrementally influenced by Employee A. While no other subjects gave such an indication or checked the scales in such a fashion, it would seem reasonable that potential attitude change was severely curtailed by Employee A's behavior. The "group retaliation" effect does, however, serve a worthwhile purpose. Commitment to some type of negative, 58 unpleasant, or discrepant behavior is posited as one of the necessary conditions for the arousal of cognitive dissonance in the literature. It is suggested that Employee A's reac— tion to the performance information illustrates the negative aspects of the score presentations. It should be remembered that Employee A's initial score was not objectively "had" (-.160, z score), but rather "average," and that this score was presented in private rather than in the presence of others. Even though the "group retaliation" effect seriously distorted the attitude change data, the Choice-Private group, for whom the greatest attitude change had been predicted, did in fact change the most. As concerns the rating method con— cept, this group had a mean change of 1.000. All other groups changed negatively in mean score as follows: the Choice-Public and Denied Choice-Private groups each changed by -l.800, and the Denied Choice-Public group changed by -2.700. Of course, as mentioned in the footnote to Table l in Chapter III, due to the post-experimental subject deletion procedure used to equalize cell n's at ten, identical results were obtained for the Choice—Public and Denied Choice-Private groups for both the pre— and posttest periods. While no pun is intended, as change would have it, none of the four subjects mentioned as participants in the "group retaliation" effect were deleted from the study. 59 All attitude change toward the job concept was in a negative direction. This concept had been considered to be of secondary importance, and was included in the study only to see if the predicted attitude change toward the rating method concept would generalize and also affect the evalua- tion of one's job. This mean negative change by treatment group was as follows: Choice—Private: —.500, Choice-Public: -.700, Denied Choice—Public: -1.100, and Denied Choice- Private: -2.A00. These changes were not statistically significant. At the time this study was proposed, an additional attitudinal assessment was included in the experimental design. This assessment was to have taken place after an additional score presentation, which too, was deleted. These portions of the study were deleted due to a conflict in time. Given the statistically significant productivity increases to be discussed in the following section, one might speculate whether the effects of a public commitment might also have resulted in greater "liking" for the rating methods. At any rate, to minimize the chances of random error, these additional assessments had been included in the original design. In summary, neither the theory nor the research on which this portion of the study was based have been questioned. The obtained attitudinal results, confounded as they were by the "group retaliation" effect, and additionally muddled due to the extremely rare results of the subject deletion 60 procedure, are nearly impossible to assess via any theoretic considerations. As the predicted attitude change was in the right direction for the group for whom intended, one must conclude that faulty experimental control, manifesting itself in unintended interaction between experimental subjects, rendered meaningless the attitude change data of this study. Publicness of Score Presentations (Commit- mentTIand Productivity_Gain The hypothesis tested in both phases of the producti- vity gain portion of this study was: H3: Individuals who learn their performance scores under public conditions will increase productivity more than those who learn their performance scores under private conditions (C Pb and DC Pb > D Pr and DC Pr). The rationale underlying this hypothesis was as follows: the performance information to be presented to the experimental subjects was objectively the same for all groups. That is, having randomized subjects to groups, one's "best bet" is that the individual groups are the same, within random error expectations. Subjectively, however, the information presented should be differentially perceived depending on whether subsequent score presentations are public or private. Individuals in the public commitment conditions, where one's possible poor performance would be learned in the presence of others, should be motivated to increase productivity and thus avoid such discrepant infor- mation in the future. This would follow from the dissonance theory prediction that the arousal of dissonance increases 61 as the potential unpleasantness of the situation increases. In other words, others' knowledge of one's poor performance should enter as a cognition against that level of perform- ance, and one should be motivated to increase productivity and thus reduce or eliminate the dissonance engendered in the public score presentations. In addition, within the public score presentation conditions, but not the private conditions, knowledge of others' performance should provide a standard of productivity comparison for the individual. Within all four conditions, one would expect producti- vity scores that were extracted from extant records when productivity was not a particularly salient issue, to be perceived as generally low when productivity was a salient issue. Of course, the nature of the score presentation should also modify this perception. In short, due to the nature of the commitment made, motivational pressures directed toward increasing producti- vity should be greater in the public than in the private conditions. The pages following will be directed to a dis- cussion of this rationale as applied to the results obtained by experimental period. Phase IV Period.-—As predicted, mean productivity gain within the public conditions for the two week Phase IV period was significantly greater than that within the pri- vate conditions, with the main effect of publicness of commitment significant at the .0A level of confidence. It would appear that the theoretic rationale is supported by 62 these results. That is, public commitment subjects, appar- ently motivated by the dissonance engendered in the public score presentations, increased productivity so as to avoid such dissonant information in the future, while private condition subjects, experiencing dissonance of a much lesser degree, did not increase productivity above random error expectations. Phase V Period.--During this second period of produc- tivity observation, a mean increase within the private commitment conditions vitiated the between treatment differ- ences present at the end of the Phase IV period. During this productivity period, the public conditions maintained their Phase IV level of productivity, while the private con- ditions increased productiVity above their Phase IV level, but to a lesser degree than that attained by the public conditions. This increase was large enough, however, so that the Phase V differences were not statistically signi- ficant. Of course, if the private conditions has maintained the same relative level of productivity for the Phase V period as for the Phase IV period, the differences would have remained statistically significant. There are a number of alternative ways of viewing the Phase V results. One would be to forward the idea of an "Hawthorne" effect as instrumental in destroying the between group differences present at the end of the Phase IV period. However, if one considers the results for the Phase V period 63 as being due to such an effect, then one would have to over- look the presumed dissonance effects for the Phase IV period. Fortunately, there is a common sense explanation that will account for the Phase V mean increases in the private conditions (my thanks to Dr. Eugene Jacobson for his assis- tance in developing this hypothesis). Unlike most psycholo— gical experiments where subjects are college students in contact with each other for only a short period of time, if at all, the experimental subjects in this study were indus- trial employees interacting freely in the normal functioning of their department over the term of this experiment. Over time, subjects undoubtedly communicated information about the study, including performance information, which would, of course, tend to diffuse along lines other than those artificial channels created by the experimental design. While it had been hypothesized that individuals learning their scores in private would be minimally motivated to increase productivity due to the fact that there was little external pressure on them to do so, and public condition subjects would be maximally motivated to increase productivity (due both to others' learning of one's performance, and one's learning of others' performance), communication along lines other than those specified in the experimental design would have tended to increase the motivational pressures within the private conditions. Over the term of the experiment, the private condition subjects probably found that their performance scores were not totally private, due both to the 6A information they themselves and transmitted, and the infor- mation furnished by the "grape-vice"; as distorted as it might have been. While certainly not conclusive, an informal check with several subjects in the private conditions revealed that this was the case. These subjects knew something of the performance scores of both other private condition subjects and public condition subjects. This information was trans- mitted over friendship lines, which, of course, were not the same as the information channels created by this experiment. Given the existence of such comparative information, one might expect productivity to increase. If one accepts this explanation for the productivity increases within the private conditions as plausible, the next question one would ask is why subjects within the pub- lic conditions did not further increase productivity during the Phase V period. While it could be that these subjects were satisfied with their Phase IV level of productivity, in that it reduced dissonance sufficiently so that addi— tional increases were not necessary (for psychological well- being at least), a re—examination of the obtained results by experimental period suggests at least two additional reasons for the non-significant Phase V differences: Treatment Group Experimental Period C Pb DC Pb C Pr DC Pr Phase I (pretest) mean .511 -.208 -.051 -.O7A Phase IV gain .A32 1.013 .007 .208 ghase IV mean .9A3 .805 -.OAA .13A hase v gain --.01% .108 .A35 .210 Phase V mean .92 .913 .391 .3AA 65 First, the factorial analyses of variance were based on productivity change scores, from pretest (Phase I) to both posttest periods (Phases IV and V). While the randomi— zation analysis of variance of pretest productivity scores did not yield a significant F ratio, the Choice—Public group was much higher in mean pretest productivity than the other experimental groups, and thus had less "room" to in— crease productivity. This factor would account for the apparent differences between the Choice-Public and Denied Choice-Public groups in productivity gain over the course of the experiment. Second, if one looks at actual mean productivity for each experimental group at the end of Phases IV and V, it would appear that a productivity ceiling had been attained within the public conditions. Clearly, at the end of both periods of productivity observation, the public groups cluster together in terms of mean productivity, as do the private groups. The fact that the public groups are nearly identical in mean productivity at the conclusion of the study would lend credence to this View. While the Phase V analysis of variance did not yield a significant F ratio, overall or grand mean productivity for this period was statistically significant when compared with the pretest productivity grand mean: t = 2.Al9, df = 70, probability of t = .01. In brief summary, the failure to achieve, or rather maintain statistical significance during the Phase V 66 productivity period appears to be due to (l) inadequate experimental control, manifesting itself in unintended interaction between experimental subjects, and (2) the existence of a rather high pretest measure on the produc- tivity variable by one of the experimental groups. Summary of Experimental Findings Perceived Choice Hypothesis (H1).--This hypothesis was supported. As predicted, the Choice-Private group perceived greater freedom of choice in the experimental situation than did any of the other experimental groups. This differential perception of choice was taken as an indication of the exis- tence of cognitive dissonance; an hypothesized necessary condition for subsequent attitude change. Attitude Change Hypothesis (H2).--This hypothesis was not supported. While the attitude change was in the right direction for the group for whom predicted, these differ- ences were not statistically significant. The failure of the the perceived choice differences to be reflected in differential attitude change was considered a joint function of a "group retaliation" effect, wherein a number of em- ployees, influenced by an employee who was extremely threatened by the performance information, checked the post- test rating scales in a consistently negative fashion, and a posttest subject deletion that resulted in identical mean scores for both the pre— and posttest periods for two ex- perimental groups. 67 Productivity Gain Hypothesis (H3).--This hypothesis was supported for the Phase IV period, but not for the Phase V period. Phase IV. During this productivity period, dissonance was presumed present in the public conditions due to the nature of the public commitment made by these subjects, and was deemed of sufficient magnitude so that productivity was increased so as to reduce it. Dissonance of a lesser degree was presumed present in the private conditions, so that only random fluctuations in productivity occurred. Phase V. During this second period of productivity observation, public condition subjects maintained their Phase IV level of productivity, while private condition subjects increased productivity above their Phase IV level, but to a lesser extent that that attained by the public con- dition subjects. This mean increase within the private con- ditions was great enough, however, so that the Phase V differences were not statistically significant. It was suggested that communication behavior beyond the scope of the experimental design increased motivational pressures within the private conditions, leading to con- comitant productivity increases that vitiated the differences present at the conclusion of Phase IV. In addition, it was suggested that a productivity ceiling had been reached by the public groups, so that further gains were very unlikely. A key factor in the failure to maintain significance for 68 this period was that one of the public groups was much higher than any of the other three groups in mean pretest productivity, and thus had less "room" to change. REFERENCES 69 REFERENCES Aronson, E. and J. Mills, "The effect of severity of initiation on liking for a group," J. Ab. Soc. Psych., 1959. :2. 177-181. Brehm, J., "Post decision changes in the desirability of alternatives," J. Ab. Soc. Psych., 1956, 55, 38A-389. , "Increasing cognitive dissonance by a fait accompli," J. Ab. Soc. Psych., 1959, 58, 379—382. , "Attitudinal consequences of commitment to discrepant behavior," J. Ab. Soc. Psych., 1960, 59, 379-383. , and A. Cohen, "Re-evaluation of choice alterna- tives as a function of their number and qualitative similarity," J. Ab. Soc. Psych., 1959,.58. 373-378. , and A. Cohen. Explorations in Cognitive Dissonance. New York: John Wiley and Sons, 1962. Chapanis, N. and A. Chapanis, "Cognitive dissonance: five years later," Psych. Bull., 196A, 5;, 1-22. Cohen, A., "Attitudinal consequences of induced discrepan— cies between cognitions and behavior," Public Opinion Quarterly, 1960, 23, 297-318. , J. Brehm, and B. Latane, "Choice of strategy and voluntary exposure to information under public and private conditions," J. Pers., 1959, 21, 63-73. , J. Brehm, and W. Fleming, "Attitude change and justification for compliance," J. Ab. Soc. Psych., 1958, pp, 276—278. , H. Terry, and C. Jones, "Attitudinal effects of choice in exposure to counterpropaganda," J. Ab. Soc. Psych., 1959, 58, 388-391. Davis, K. and E. Jones, "Changes in interpersonal perception as a means of reducing cognitive dissonance," J. Ab. Soc. Psych., 1960, 5;, A02-A10. 7O 71 Deutsch, M. and R. Krauss, Theories in Social Psychology. New York: Basic Books, Inc., 1965. Deutsch, M., R. Krauss, and N. Rosenau, "Dissonance or defensiveness?" J. Pers., 1962, 58, 16-28. Festinger, L., A Theory of Cognitive Dissonance. Stanfordf Stanford U. Press, 1957. , and J. Carlsmith, "Cognitive consequence of forced compliance," J. Ab. Soc. Psych., 1959, 58, 203-210. Janis, I. and J. Gilmore, "The influence of incentive con- ditions on the success of role playing in modifying attitudes," J. Pers. Soc. Psyph., 1965, 1, 17-27. McGuire, W., "Attitudes and opinions," Annu. Rev. Psychol., 1966, 11’ A75—51A. Rabbie, J., J. Brehm, and A. Cohen, "Verbalization and reactions to cognitive dissonance," J. Pers., 1959, 31, A07-u17. Zajonc, R., "Balance, congruity, and dissonance," Public Opinion Quarterly, 1960, 2A, 280-296. Zimbardo, P., "Involvement and communication discrepancy as determinants of attitude change," J. Ab. Soc. Psych., 1960, pg, 86—9u. APPENDICES 72 APPENDIX A PRETEST ATTITUDE INSTRUMENT 73 OLDSMOBILE DIVISION, GENERAL MOTORS CORPORATION E.D.P. Key Punch Operator Evaluation Please read the following pages in sequence and follow all Instructions exactly. There is no correct answer to any question; we are Interested only in how ygg feel about certain ideas. Please make sure you fill out the last page completely, as that data will be used for purposes of analysis. No one will ever know how you responded to the questions. Thank you. a . \ kw-.. . utl. ' I..- On each page of this booklet you will find a different concept to be judged and beneath it a set of scales. You are to rate the concept on each of these scales in order. Here is how you are to use the scales: If you feel that the concept at the tOp of the page is very closely (extremely) related to one end of the scale, you should place your check mark as follows: Good X : : : : : : Bad extremely quite' slightly neither slightly quite extremely or Good : : : : : : X Bad extremely quite slightly neither slightly quite extremely If you feel that the concept at the top of the page is quite closely related to one or the other end of the scale (but not extremely), you should place your echeck mark as follows: Fair : X : : : : : Unfair extremely quite slightly neither slightly quite extremely ‘or Fair : : : : : X : Unfair extremely quite ' slightly - neither ' slightly ' quite extremely If the concept seems only slightly related to one side as Opposed to the other side (but not really neutral), then you should check as follows: Cruel : : X : : : : Kind extremely quite slightly neither slightly quite extremely or Cruel : : : : X : : Kind extremely quite slightly neither slightly quite extremely If you consider the concept to be neutral on the scale, both sides of the scale equally associated with the concept, or if the scale is completely irrelevant, unrelated to the concept, then you should place your check mark in the middle space. Safe : : : X : : : Dangerous extremely quite slightly neither slightly quite extremely n. ... . i I ' . . II 'fi t ' ' I ' . .“~I l 5' ' '1 ‘. ‘ ' ’ ...--._......._-.. - --- -. - .-....... —-.-- ‘ o (2"! 1 ‘ ... , 1'. _ I I I. . m......-.---_.. . - . . .. .. - . --. m.--_ .. - . , ‘ l ' . v u -7- --_. .--.. . . . ...4--.....-. s. r « {— l ‘ I - .. ..- . ..-— ...._.. ..e.. --.-< . .. 1....-- n t i I . . I ' 1 r: ' ‘I . ‘ ' " ‘ . ~ , . , I l ‘ - . o - -_...I ..V .- , . . ‘ . ‘ v . ..p... ..--.-—~.~-w— fl-.-“ "b- w.- r-.— .av - -— - -.--.< _.- 1"- —‘>"“ . I I 3' , \ o -. ~< I . 3 g ' . .Ja ——q - a a . I ' I .'J 1 ‘ ‘ w}, 31. '.',)1.', i , . 1 ,.._.‘-, r 1 K' . ,,,. . .II -I ~‘ . . —.--. . 1.: Key Punch Operating _3 an Occupation. leasant . . . Unpleasant extremely quite slightly neither slightly quite extremely nstable . . . . Stable extremely quite slightly neither slightly quite extremely Nice . : . . Awful extremely quite slightly neither slightly quite extremely orthless . . . . Valuable extremely quite slightly neither slightly quite extremely Good . . Bad extremely quite slightly neither slightly quite extremely nskilled . . Skilled extremely quite slightly neither slightly quite extremely .‘u 9-. _- .- _-.. . . ---- . ., | .. ..... . ”“11 I / - .-.... ..-.. I . | '. I ...... -‘ I I l ... -¢.... I J . if me ___Tu-,_-- _ 1-, U f ‘ I ' l \ mun-aw _...._.—...-— - _-._.._....._.... ._._. .. . ‘1 ‘:l ‘7 ' x: - . . l .r .. - _... . I-.... _.. - . “l l y w. . l ‘ I . .—~. ....—.---..~._....v....r I . . I . I' . A n -. . .0... ...l.. a -e. a I - . .... ....—_ u . .. . .‘ 1:, A .I . _........-- —-_...- - r- 1 .._.' 4 .. ...........,. . i I , '7" :n}i.._ .’ ‘ ~ I Fl“; I -..-~ 1 . : r ,. .- ‘ I - ' .‘ "I ._‘ --—.-....-—-...- --.._. __..._-.._....._. _ A I l I 'l . I (‘W' ‘ ‘1 f ‘l ' ‘ I‘ . ~ . d . .ll.’ ‘ 1-"l,.) _ . . 7 9 ‘ ‘ o ‘ I - ~ .4 I ' — «4‘. . - .——hrq‘-’.‘~- - -—-‘.~o..-§ -—'-"-.—-. . .-. . -. -. . . , . r v . ---I ~ M f . v , 8 -~ I" ' ‘ a . -, . I . , , 1,.“ ..-... _ .. .4- .i._ .-».. .. . .-- _ I , l ,. , ’ ‘ .5 : ‘ r i . r1 - 4 .K M-.." .. - _.... . - . ~ _. .._- --.....--... - , -.. x . . . _ ~ , ‘ . ‘, -}i .v..-.-.- ...._....._-.- ._-,V__..... i .. -..- . . I ". I g I ‘ '- I t . ' J I p - . 0 ‘. ‘ - I . _ , . 7. . I M-,... - ._.- Your Job 55 Oldsmobile Division, General Motors Corporation aluable . . . . extremely quite slightly neither slightly quite extremely nskilled . . . . . extremely quite slightly neither slightly quite extremely leasant . . . . extremely quite slightly neither slightly quite extremely Nice . . . extremely quite slightly neither slightly quite extremely wstable . . . . extremely quite slightly neither slightly quite extremely Good . . . extremely quite slightly neither slightly quite extremely Worthless Skilled Unpleasant Awful Stable Bad Inganai I.. ~{ I l (I \j ' ..'-I'.'\A . f ‘ I : Al, ‘ I ~ A n 7' .I 1? .~ l . Al O O *—--~4--—.-rr--..O.'M-.~-' -—-.-' . §.m—.-‘ I .1- ‘O- .O-‘ I . ‘ . _ ~ . v . . W |_ ‘ _ ‘ ‘ I.'l;.-.lI:,-!J_~=- JIIUU 7! Mi: ;l ‘J'rll '1'. xvi, l:' l; I I ..._ . \nsrmuzo :3} Hi' w uh)! l' l‘;ll.l‘J.-l t-l :2»:- , Q -_—.-_-.-- 1. ‘- ._- .- . 0.. ~ “1“ . . u —-M.- . - -.“ 00—“ u..---~.--—--v ”10-- ¢-D.- -. .. . .z ' ~ . “ . .5 ' ‘I"‘t:!sl:~, ,- 3; lairlltllu Him-~77 gl'II-n! poo -O....~.~ _-..-.~-—- ‘ .. .._- 7.-.“. _ 0--.... ...7.... .._. --- ._..- l .‘ . ' , , . , a ., r: , I ';'.-‘-"~.J I. ’ . 3 J 1‘." "' --. -h- .-—— -.._.... . . . . .--.c--- I' ’ " - -_ g . ‘ . , I ,2, ‘F all) 1).»l!‘ . __... m .-.-- -__.._ . “m-.." A . \ 0- ‘ . A , , 1 ,2'3} :1 IA glzupllz \ u——.. - .r~———_- ——-.-— ,I-‘,-n ‘I‘i‘l‘ 2 \ ’. . Ii : I': ll -q-.. -. -..--_..._.7--...-_-._-. -._- --‘ x .- co -... o t . "l ‘ .' , u ' t A.) ‘ ‘ - V . _--.. . .n...‘ ....._. _- -..I . _.,..-u u I o . A \ J , . . \ *) n u... .. __,_.-._-.- 7..-- -. ”.— .—_- —__ . .. ,, - o I ' - O ' u . . 1 ‘ - I v o-...-u....-. _*»“_..r-- ”_— “- h-.\- _...-_ -- . J*-—-lnu.—..A--vvn I -.‘--»fl oabcr< ——-~- - .I . . _ ‘ .. -'I . l \- , ‘ ' ' ‘ - j- ‘ ' i \ ' ' leasant Cruel Bad aluable Awful Skilled Unfair The Other Key Punch Qperators lg Oldsmobile's Key Punch Section Doing The Same Work You Are Doing, extremely quite slightly neither slightly quite .extremely extremely quite 1 slightly 3 neither 1 slightly quite .extremely extremely quite . slightly neither . slightly . quite .extremely extremely quite . slightly neither . slightly . quite .extremely extremely quite slightly neither . slightly . quite .extremely extremely quite slightly neither slightly quite .extremely extremely quite . slightly neither . slightly . quite .extremely Unpleasant Kind Good Worthless Nice Unskilled Fair hi1? ‘ a -. , -n ___.-_.-,.-_,,,, . .- . ' l .. ---- WV. .. .. .7 ....... - .. - C a . a ,._.......—......_ ......_.... .... 4.. ~. . . . < r u n I n‘ ' 1 .I v ' - - l . ‘ ' ‘ . ' ‘. .I .- ‘ ' . . . ‘ . ~ I I . V... .. m- —» . .- _. u..-» . ...._.-.-.... “m. . ‘ l I "‘I . . , . , ‘ I I r - I HP." 1 t l -1! . l . . .--. _---~..-... _...- . ---~-... ... ....... . . l . L“ -. . . a l ,. -I I _ I li- .. ..-.._..............._‘ - .. .fl . --..- f ‘ ' ' I ’ a" ' ‘ I t :2 'l' , ,l , i o . I l l . . -., -......_-....__. .. .. . . . ,- . .. . ..... -4 a.--” I . _ -. -_-.. . . - _.._...-~._.. .1 , I l' a, . - z . . s. .. ..~—~._....._.-‘.._._ _. - ‘ i l ‘ l 1 i‘ V I I M: . \ --,.....-. - . . . i - , . ....-..._... H . ,1. l r . ‘ r I A .. A- s-...-...._ I . . .. 6 I! ............-,-.. .’ I . . H- ._ ,l . I I I .. .. .-- - .. . .- . . .. .. 7...”--. u .. -..., _...,~._. —.«—— _— ‘u t l . . I v‘ - . I l \ 1.. l . l ll.“ . ._ . ..M .H.. . .. .~.... A .... w-- . , . - . - ..- - a“... a. . .. c -._~—- ....--_...- . . . ‘ l' l n '. . ' l I“ l I l - . . l - '.’ . .,-- . ... ---....,. . ....__..‘...- -.. .. . _.--..... A-.—-.--H . , . ., L o ' l l . ' . 1| . I v I v . | A ‘. 0""..- ._ .... “— ___.— - .‘ i ‘ .ak‘ . .9 n A\‘ . \ | \ ‘ I ‘ -, . U ' V ‘ I .t ‘ . . -w---“ A _ - _. . . . I . ., _ \ ' l I ' I. I ,, . ”mi—q"-.. -. -...-—...- I _-._... -....- ---. . . r ‘f p '7 ‘ , A -._..._.-. .V.. .a..__._. .-- k. The Method lg Which Your Work is Evaluated Ialuable . . . extremely quite slightly neither slightly quite extremely Bad . . extremely quite slightly neither slightly quite extremely Awful . . . . . extremely quite slightly neither slightly quite extremely Safe . . . . extremely quite slightly neither slightly quite extremely Unfair . . . . extremely quite slightly neither slightly quite extremely Decisive . extremely quite slightly neither slightly quite extremely Skilled : : . . . extremely quite slightly neither slightly quite extremely Worthless Good Nice Dangerous Fair lndecisive Unskilled ‘ n . . 1 \ r l s . . ‘ - U v t I l I . « v . . - . 7-7 .. _. , - .. - . - . ._ ‘A I ...—...“ .. - , .. . ‘ , , _ . -.. ,- . .11 -. , . , . . . I . v I V I , c . . . u .- -....— - . .... - _ ...- -.s -. -.. ...r t ,... t -. . . -. .- .._ ..k... .....- r. ..u..._,....-..._- I ' I ‘ ‘ - I o l‘ ‘ A f ‘ ! f I - I ' ' . A I , _ ." .‘ _ . . A a ..I . I . .l - - ‘ I " _ ,\ 3‘ - o u . . o '2' -’ ' - - u . . .- --——.._~_..—-. -...A ---. .. . - . __ . - ...... . A ...... ‘ . . _. a- ...“-.. A A _ . . . . ,- .-..--fl... .... ., . .... .-__....—.— ..-—_—_--.. 4-.- A-.'—..' ' i l . j u t C .' r I . ~ v . _ , . 1 . » f to (7 t . . . ll l ‘ r I . a l . ~ 5 . . . - ~ I ; ' a ._ ’ ll .1 . . - - - - . u . .- ¢4 ' . . . I I ——....~—.—.... -.---.“ --.—.-.-.- . g." . ...-- .._.._——-—- ... A.. -.--.v—------.. ... r“... e - 7- - .. .... - — - f . - I . . . . . ‘ I . . ., . . ’ . 4 ,‘ ‘.. . . - ~ .7 i . ., 't ‘ u . .. . , ‘ . l v . | r u, , . ‘ ‘ ' .‘ . .‘ I - _ J I . I - . . s . . u . . . . .. .-.»--- ..w -. ...- . .-.. ._ .V.” .-.--fl -7 . . . _u“ .. -t A - . ,s .s — -— »_r_ ‘ a - . . a " . l . V l _ l . .o . ' ,- _ - . . ,7, ,I....‘_ ,l‘ , ,.A . V ' v z . .4 ,. I 4- _ ‘ , "- ‘ . - a . -' l I I 0 . . . “..-“tw... ... “...--..__.....—— . ..—-—._..........—- . . - _. ---”--_. . -..- - .....- .-- . ,-.. --_--.-- -..-..- -._ _ -- -..--- ' u v | . , u v ' . - | . § . ‘ L , .. ‘ u t . , i , . o . . ‘ . ‘ , I . ‘y l I", - . c . f I ._ l 1 ' . ' u _ 7‘ q x . . - .1 7‘» I . -—_-- -.—. .— ...”- _- _ -_._.....- ..AA .. --. . ._ ..H ...... _. - -7“ ...—..- -. .. . . - .---. - --. - . .. - . .......__ _- --.. .mmu- —_...... . . . . ‘ . . . - . . I . . - I ‘ ' . ' . \ , i . . A ' . . , .. l x .. A . l n g u . - . y I u I I “fl-mwa___-- ‘--._. .. -fl- . t .....- - . ,__.. . . 7 -.-- ..- . .-- ---V— . .. 4 . . . -. . .—,--..—. ~-—-- ..-——_~ -.—- - V , 7.....-— .,._.. . . , . . v u . ‘ , ' i. A ‘ . ..I . . t . .v . - a r . . , . l - l Dangerous Decisive Skilled Awful Kind Pleasant Unfair Valuable Good Your Job Performance Introducing a New Method 9_f_ Evaluatlagae extremely quite slightly neither . slightly quite extremely— extremely quite . slightly neither . slightly - quite .extremely extremely quite . slightly neither . slightly quite .extremely extremely quite slightly neither slightly quite .extremely extremely quite slightly neither . slightly quite .extremely extremely quite slightly neither . slightly . quite .extremely extremely quite slightly neither . slightly quite .extremely extremely quite . slightly neither slightly quite .extremely extremely quite slightly neither slightly quite .extremely Safe lndecisive Unskilled Nice Cruel Unpleasant Fair Worthless Bad I Q I ' A . —-» ---. _- r. ”v... .....-q... ”...... ' . , . ‘) . . .. .. \. I I. 9 - _,.l . . J l - , I . . ..» _ . - .-- , .-.. . ...~ —...-- .. . -.— A . A. . . I ' . ‘ i 0 1’1 . u t .V ..H_ -. --.. -A . , . . . ' . . ‘ I n .... t-..u-..-_4.. . .r.. .---........_- i. . _ 3 . - ‘ ’ .- 1 . . -‘ .‘ ‘L I I . . . - —. .V.-. - ,- ... -- .... -.-—...-”— l u ‘ I l l t .r l ‘ l . _' . ' :L) i 1 .. a . n V ' i ., l u ‘ l v c .- rv - .. I I - . » - I . l ‘ — I I o ‘ l \ I i . I A l . 0 v I V I ‘ n i ‘ . ‘ 4 . . . . . . o I . —-........._. -..--_--.... .- .-. -.--..q..- .- ......4- ..~.-..- ---.-.._.-. -7 .r- I.,_.._.....—~...II -. ..._ -......V.~..--.. ...- .-7 . . ...—~..——~.—.-...- .. . -—. , I I . . . . I . . I a - . ‘1 . > - l I - - | I l . . [SI , I D - . I I . K . _ .. . . . I l I \ . . . n a . ...--...I 7.. .--—.-.I_ ...H. “....-. -x-.-.“(7- ,...... .-..“ -_. .77-..._.. .. .. M...- .-I .. ...-.....-»— n‘ .. ..--..H. ..-. n, ... .........,r... . .- . . . . I I , I I, v . I P ' . l , I I . A . . . - . . .--- - ...._ . ...—n -~_ -.-- . V a -- . - U... -- . -. -——- .--~...... .... --..-. .7 . ..~-~...--- .- . . 7 .- . “...—.--- ..- . . . - . l I J ' t . . I ; I ., ‘ ‘ I r I l ' I l A . ' I A I If I I i "| ‘ ’ ‘ I' I ‘ ' N - .v. Q II .~.I. ...-I... . "...--...--J ....” --_- .. V. . .---....__..._. ...“...- _- _ ..-- .-— . .1- . .. -.. .. . —-.--—_-—'-‘- . U . ' . ' ' . u -..._I..,, .-—.—..——... -...— .---._., . I ...-.... t m- w. ._ - II .._,._,,_._ .—. ~——...—-—.‘ 4‘ .2. s U ‘ 'v. | l ' U l I (A 1' I I l .’ ' I i - r' I l‘ I V . l U :7. “4 I ‘. '. _~ ‘l .V\ b . . . . . ...n........- “m“..- . ...--..—-——.-....... . -. . .. .-. .... _._..._~—---I. . - ~h..-»... .... -.. ——-v- ...“..- _-....'—..-.-—...-—.- .-. --.. -t- .. m. ......"- ' - ,_ .— ...—....- ...-.. ----J. - ‘ . - ' l ' l 1 ' ' ‘ I r . . I u - I. ; ' ‘ -. ' x ' t'I ‘(‘J""‘"~" I.” .A- vll"".l I «Hm - ‘I’..'- \ V . . . . I . n . u .---..I.~... .... ‘~,._..-.--. ...—..-I--—__. --.. ._—-.. ....-._....-.-__.. .-~.—- ..I- -..-«,..a -I- ...~ .. ,___ .4... ..,. .- - ...—7-... , "....-- .7”- Oldsmobile Supervision Outside gf_Your Degartment. Good . . : . extremely quite slightly neither slightly quite extremely Stable : . extremely quite slightly neither slightly quite extremely Unfair . . : . extremely quite slightly neither slightly quite extremely Safe . . : . extremely quite slightly neither slightly quite extremely lndecisive . . : . extremely quite slightly neither slightly quite extremely Awful . : . extremely quite slightly neither slightly quite extremely Cruel . . : . extremely quite slightly neither slightly quite extremely Valuable . . : . extremely quite slightly neither slightly quite extremely Pleasant . : . extremely quite slightly neither slightly quite extremely Unskilled : . extremely quite slightly neither slightly quite extremely Bad Unstable Fair Dangerous Decisive Nice Kind Worthless Unpleasant Skilled . . ‘ ". .J I It ......_-. --~._.—.— ._ --. -. ... . n' . t , .; ‘.r ' . I: I I V ... . ..—....._~-~—... . ~ l9 ' . i ' ‘ I _ . . -..... ......I.-.~...---.-. .. E . . ~ ~' . I l .l l I - “-“_..-_-_-.._.. _ -.. ' l I I 1‘ .-. r ‘1 I . I . \- ‘ Jill ‘I , _) o n l ,' ' , 4 . . . I I ...—.— .-.n— ...-.“w - . u . . — V .l . I 3’ . z ' . ' o a. . , ...-I. ...I -—..._... -.....I .. . - ; \_ I .'I I I . | i I l .1 ~) ...—7. -—-.......__.I mv-i... '- . , . ' l ,I I ‘ o . ..w— -.-A,--_I_I.- , I I ...... - -. .. .- I n I l ' .' ‘ \ —.-,.-——.—~-v . .- I l l .I I, - ~. I a I I n - . -....I -...» a“ ‘_ 'I. . 1 " .‘ “.' l \‘ I~ 'I ' _ ....--.-- .. .... -..- H- .._. , ,___ C , . .. -..—.-- ..A,-..-,.____._._,.- .- .-.. _- . . I ; . ' ‘ . . —~»—-‘n-OI_~. . .— ...-”Q &—~ I— — ‘-u i '. . . I - ' '- . . I ' W. . . . -*. ... -..”--- . ..... .. . r . I 5 . . I . I . -- . ,..._.A . -... .-..h- . -..... - i I I .‘- ‘J l ' l \ . . . ....‘w . ~~~¢ . . . I. l I‘ . ' l I . I .. .--. .. ‘ —.‘ .- ....-ka...r- -.-- ; z '1 3'! l g: l 7‘ . .. ..--_.-..-._~_. - ...-.. ._ . ...... I I . I ' ‘ l .v E. \. i ' I1. . I h K c . - . _-_- .- ...... ...-.-....—. - - , .' .1 I , I . ' I I '.' ' I I I . .- . .......... H...” I- r I ‘ I I .l ... ‘ o -—'-~»—o—<---. M‘v— - . »-l «0-..- .-<—»—--- n . I u . . I I ‘ . ', . f! ., r- . I l I. 3 . n. I. \ I ... ll 'l ‘I l ,_ _. .— ‘ I I. —.I.. .I I l., o I I l l I I f I I ‘ I . . - I ...»-.A—“. - .- II .c . u - -.- - . . l I 1 u I I I I I - ... -... - I --.- ... - -... I. ‘f_ I i ' -' . ....A-_.._..-. . ' u ‘ i . i '3 . . .2 I '_ -...._-..__ ,... . I i -h. a-“ .-- ' I ~ I , l --- . - I, —-. I.; I— _ I l I. ,I I Oldsmobile Data Processing Key_Punch Section Pleasant . . extremely quite slightly neither slightly quite extremely Unskilled . extremely quite slightly neither slightly quite extremely Stable . . extremely quite slightly neither slightly quite extremely Good . . extremely quite slightly neither slightly quite extremely Aorthless . . . extremely quite slightly neither slightly quite extremely Awful . . . extremely quite slightly neither slightly quite extremely Cruel . . . . . . extremely quite slightly neither slightly quite extremely Unpleasant Skilled Unstable Bad Valuable Nice Kind -‘- ~ Vii-1'. / .. :AI -"il ', \u‘ ll ‘1 ... ’ l .' v 7' a ‘ ‘. — m-.. -.—. . . I a « .I‘I""" “ .0 . ..w—c-n u—‘i.’ I I t . '~ I.l . I O 4 ~ - taut-w...h. I..-.—.--~...1 . r...“ ..— ...-..- Io.“ -.......I .. .- ..1»_-‘s..- . . Inna—.... ., ....-II- _ . . . I . . ......I. ....-- .....- --.—.... .. -_ . ...-.. .......---I....... -_ ...... . I .. I . _.,--__-_.,_.__.__. ..-.-., , I . . . . ........--I. ..--- .. - ' l . “...--.y-.- -- ..-... - . a... ,. I‘ . ...-.-_....-..... .. . . ..-.-... Oldsmobile Data Processing Department pleasant . extremely quite slightly neither slightly quite extremely Good . extremely quite slightly neither slightly quite extremely Valuable . . . extremely quite slightly neither slightly quite extremely nskilled . . extremely quite slightly neither slightly quite extremely Unfair . extremely quite slightly neither slightly quite extremely Decisive . . - extremely quite slightly neither slightly quite extremely Cruel . . . . extremely quite slightly neither slightly quite extremely Nice . . extremely quite slightly neither slightly quite extremely Unstable . extremely quite slightly neither slightly quite extremely Pleasant Bad Worthless Skilled Fair lndecisive Kind Awful Stable --- .... ... .... -. t ... ..-... .. . .. -..-..., . .. .. ...... . _. A . . y . . . " | _» ~ . [r l . ,‘ x ‘ I . 1. ' . i . . - _,. H - .. - __,,. - . ._‘ .-.._...-.. - - . ..l --..u-‘ . . , ' I . 5 ' 4 . l I . ; v ‘ 1 l . . ‘ 1 - “L . u . J v. i - u . . -.. m . ,. ...-fi- .- ...... . -. ..s ....—_‘_._ , ,s-.-....—_.‘ _ _. . u I ; I c ' I ‘ a l ' ' .. .. ._ . ”..--.. .... .- - “.4--- .... H... -.... .7 ....- ... _,._.____,_- 4. I f l ‘ — -, i -. 4 , l l _ , l . , . . . ' - .....- ... ....-. . _ ., _ . ...... ...”... -..—s -.. ...... - . ..-. , . 4 . u a ‘ A , l x , 1 . - 1 ‘ . 4 . l .‘ e C ‘ - “....-- .-.A..V ... .—.--l.~. .‘. - - .- _. a“- _H- . l . . . - . i ‘- ' . l l l .’ ' l A l . . . I v I , -t .-- -.. —iah~-— . q. _.l . . .s -- ,_A_.-._v-‘ ~ —. - . . i . - I» l- . ! ,. « ‘ . \ ‘ A I i _ .- J .- . . o . . . .. e--- --. «....-- . . ,. - .— mv..u- .. _. ....,._.....s..._ .. . - . - - -- _ -..-“ . I ‘ . u . -.‘ . 1 - 9 ‘. ~. .) i‘ r ‘ 0 ~ ' , . ...... a. ...s_ ..._ . _ _....-...._._ . ..- -- _ - ._‘-__-_- N.- . -.- - - - - ..,.._-,.s...- n | ' I l ‘ , u . V » l . ‘ x | i . - . '; I . , - ‘ . i J ' . I ~ u I . fi.s .. - - u a -.....»w... s ..|. M.,"... :'l ' ' ' . _ . . , . l k—‘-- . -. r l - a- ., -..- ...-... .../q- _.—.-- ..., ‘ ~ ‘. . t . _. -. -- - - u .— ‘ - Pleasant Oldsmobile Division - General Motors Corgoration slightly slightly extremely quite neither quite extremely Dangerous . extremely quite slightly neither slightly quite extremely Unskilled . extremely quite slightly neither slightly quite extremely Stable . . . . extremely quite slightly neither slightly quite extremely Worthless . . extremely quite slightly neither slightly quite extremely Kind . . . extremely quite slightly neither slightly quite extremely Fair . . extremely quite slightly neither slightly quite extremely Awful . . . extremely quite slightly neither slightly quite extremely Bad . extremely quite slightly neither slightly quite extremely ecisive . . extremely quite slightly neither slightly quite extremely Unpleasant Safe Skilled Unstable Valuable Cruel Unfair Nice Good lndecisive v ' . . ,1 , I . v . - - . -l- .- -... ,p. ....-- ..---- -.‘-‘ ‘ x . . é _ l . . , J J ’\ .l ” ' l I ' A -—-—ow-—--. .-. Il.-- . .7 >d‘lr --——-— --—‘ ..uQ - —..-- 4- Is ».-, r ‘ . . , . . . I ' l ‘ l - l I ’ l . l , . l . J 7, A A , . \ A .‘ ,I . ' ..v I ‘ . ..- ...... ,_’_., s .. 1 .h.__-.-___..._.'» . ... ...—.... ...U -7 , l . I = . . ‘ l '. - ' ' ’ . 'Ii \ ,' . i - 1 l .j 4 . . ‘ , . { , l | , ,1 - ‘ ii , ...—...-..— “ta—-...w‘...‘ ...A...._ -. -..- _._.- .s . ...—-..-.s. H a r . r ; . r , — . _ i . . . -, - l - s I ,~ - ' l - 'l. ' u I ‘ l z‘i' , I“ a . . .. . I , .. ...—.-._.‘ . -m ..._ . .. -‘,..-_._..H.- . - u .f-7d-...-.t ..---- I . . , ~ .... . . l . . ml ‘ . v I ‘-‘» ' n v I '. . I ,. , -1 ' y-ui 3 - ..-—-.- -... .... .--. . .. .....-“fi us. - .. .. ...“ _ .— - --. A-.- , , . r . l . I 7,, , . . . . . , xx , ‘ I . | l ..-... —...._. .. ...-..., ,-..... . . _4 s- .... - .... .-7 -.‘u .- __.- .-.-.—4 ‘ _ . . A . < r' :3; " . ‘ I ' ' . ‘ ‘ ' v , . ‘ , . -. .. ~_-~. ...... ... . .- _- s -.. .. .. W » ...h... H. ._ a ' . . l w :r~ l l l , . . ‘ . ‘ . . ' . ‘ u , . ‘ ‘ l1 . . . ‘ 1 . , - i c u . ..----“ _-..._--._. ...... .... . .._.,.__-. _ .....-u. . .E . . ... V D , ° , u . , . .~ -. . - . ‘ . l . I 4 r , l . l A V ’ '. ~ g , _ _l l I . . . l,‘-._"._ ' . -.M....- --.»............ ...... .. .-.. -...,..-.,. - . .---.....~.___ s. ' r " ‘ -;- - . I I . < . _ H , ~ _’ , l . ~ I : “’7‘ l‘ l . ‘ y I j ’r: . u - -—._.-. ..-...”H .7 - .—.-..-- -.. u .. . .7 u - -..- ‘-......_-._-.. . , . .V l A . ' , ' i ' . .2 l l ‘ ~...‘~ - . I . I.. p“ : .-- ._ _..... -...sk l f v l .....r .— -.4’ . - , o . . . . - »~ ~ .. a. .——-“-. 7---. .. ‘ l ...-.. . . ..-—...— v . V . - _ . . -- .r-v .... . , . A .. . .-_H.-— -——-—- l L -. 4--.. . ..- , s a . - v ‘-<~- .... e a—. » a . , ‘ ... ..-.--«-l-—. bad bad Name: Seniority date: (Month and Year): Shift: Days Afternoons Which type of work do you do the most? Key Punching Verifying _____ About equal As a general evaluation, how would you rate your job performance? (If less than one year's seniority, compare yourself to others with comparable seniority). good extremely quite slightly neither slightly quite extremely How do you think your supervisor would rate your job performance? good extremely quite slightly neither slightly quite extremely You will be contacted in a few weeks, either singly or in small groups, to learn the results of this evaluation. Thank you for your assistance today. .,-~-"' -s..._...~ u‘" ' W --‘\.___‘ ”m- 7’ l (I I " fl-'*. ~ . - ' ,. ‘3"! N" ‘ ...... ‘_ .. _1, . a.“ ’ - ~"- ‘— .. l I” \ [\- - . '- . - --. . u - ..__. f- . ..- -.--n "' .. l l l ' ‘V 4 _ ,» --— -"“- --’- .. g... -....—..._... v . l l . . my, 3 ‘ L- ‘ ' - '__.._...-.—‘ -.. .... ~— APPENDIX B 1-6 EXPERIMENTAL MESSAGES 88 JLDBMDBILE 89 In order to better evaluate the Data Processing Key Punch Activity, 3 new method of employee rating will be introduced during the month of April. For each employee, there are two ratings available: an ”Efficiency Score”, ranging from -h to +h (with -h being very poor, 0 average, and +4 very good), and an ”Effective Pro- ductivity Score”, expressed in percentages ranging from 0 to l00 (with 0 being very poor, 50 average, and l00 very good). You will be rated with the ”Efficiency Score“, and will be told of your initial score, in private, early next week. If you have any questions, please feel free to ask them now. General Supervisor Data Processing Operations OLOBMOBILE C?“-,-'-i8!ON t GEP£ERAL MOTORS CORPORATION 1* LANSING.MICHIGA!\! 48991 DLDBMOBILE 90 In order to better evaluate the Data Processing Key Punch Activity, a new method of employee rating will be introduced during the month of April. For each employee, there are two ratings available: an ”Efficiency Score”, ranging from -h to +h (with -4 being very poor, 0 average, and +h very good), and an “Effective Pro- ductivity Score”, eXpressed in percentages ranging from O to lOO (with 0 being very poor, 50 average, and l00 very good). You will be rated with the “Effective Productivity Score”, and will be told of your initial score, in private, early next week. If you have any questions, please feel free to ask them now. General Supervisor Data Processing Operations ..-—...-— AIQI‘IH fl‘lf‘ulnAM Ant-Q1 JLIJBMOBILE 91 In order to better evaluate the Data Processing Key Punch Activity, a new method of employee rating will be introduced during the month of April. For each employee, there are two ratings available: an ”Efficiency Score”, ranging from -h to +h (with -A being very poor, 0 average, and +h very good), and an "Effective Pro- ductivity Score“, expressed in percentages ranging from 0 to lOO (with 0 being very poor, 50 average, and lOO very good). In the following space, please check which of the two scores you would prefer to be rated with: A: Efficiency Score B: Effective Productivity Score In the interests of time, you will be told of your initial score in a small group from three to six employees, early next week. If you have any questions, please feel free to ask them now. General Supervisor Data Processing Operations flaiviCiE-ZHLE DIVISION t GENERAL MOTORS CORPORATION i LANOINO. MICHIGAN 4.!!931 DLDBMOBILE 92 In order to better evaluate the Data Processing Key Punch Activity, a new method of employee rating will be introduced during the month of April. For each employee, there are two ratings available: an ”Efficiency Score”, ranging from -# to +h (with -h being very poor, 0 being average, and +h very good), and an “Effective Pro- ductivity Score”, expressed in percentages ranging from D to ICC (with 0 being very poor, 50 average, and 100 very good). In the following Space, please check which of the two scores you would prefer to be rated with: A: Efficiency Score ’ . 8: Effective Productivity Score . You will be told of your initial score, in private, early next week. If you have any questions, please feel free to ask them now 0 General Supervisor Data Processing Operations DLOBMOBILE 93 in order to better evaluate the Data Processing Key Punch Activity, a new method of employee rating will be introduced during the month of April. For each employee, there are two ratings available: an ”Efficiency Score”, ranging from -h to +h (with -h being very poor, 0 average, and +h very good), and an ”Effective Pro- ductivity Score”, expressed in percentages ranging from O to 100 (with 0 being very poor, 50 average, and 100 very good). You will be rated with the ”Effective Productivity Score", and in the interests of time, will be told of your initial score in a small group of from three to six other employees, early next week. , If you have any questions, please feel free to ask them now. General Supervisor Data Processing Operations BLOBMOBILE 9A In order to better evaluate the Data Processing Key Punch Activity, a new method of employee rating will be introduced during the month of April. For each employee, there are two ratings available: an I'Efficiency Score“, ranging from -# to +h (with -4 being very poor, 0 average, and +4 very good), and an ”Efficetive Pro- ductivity Score”, expressed in percentages ranging from 0 to lOO (with 0 being very poor, 50 average, and lOO very good). You will be rated with the “Efficiency Score", and in the interests of time, will be told of your initial score in a small group of from three to six other employees, early next week. If you have any questions, please feel free to ask them now. General Supervisor Data Processing Operations APPENDIX C POSTTEST ATTITUDE INSTRUMENT 95 Your Job A£_Oldsmobile Division General Motors Corporation Valuable . . : . : . Worthless extremely quite slightly neither slightly quite extremely Unpleasant . : . : . Pleasant extremely quite slightly neither slightly quite extremely Awful . . : . : . Nice extremely quite slightly neither slightly quite extremely Good . : : . : . Bad extremely quite slightly neither slightly quite extremely n . I i .._i....._:. I . . :‘lmjul-c ” «:77: M '_ , 77:17" “7'5"” ! Jami-E , “..--.. ..-“... ..._,. .. — ... , i l . . ...—...- ,MJ Good Pleasant Worthless Nice The New Method 21 Evaluating Your Job Performance extremely. quite : slightly ~neither slightly: quite .extremely extremely. quite :slightly : neither slightly: quite .extremely extremely. quite 2 slightly: neither : slightly: quite .extremely extremely. quite :slightly : neither slightly: quite .extremely Bad Unpleasant Valuable Awful ‘ ... “L‘— ' I u .""“-'n.,-.. -. . .' ‘I T I i A r I . -~ ...—'7'. _- . u“ . o l . . l I. . ’ o o -o .-.. . -... --.... ' v .. - ~‘m_ . ._ ..-,_ -- - . I I . l ... 'e n ~.¢. “ -- .... Good Good Little Name: Seniority Date: Shift: Days Key Puncher Afternoons Verifier As a general evaluation, how would you rate your job performance? : : : : : : Bad extremely quite slightly neither slightly quite extremely How do you think your supervisor would rate your job performance? : : : : : : Bad extremely quite slightly neither slightly quite extremely Check on the following scale how much choice you feel you had in what you did here today. Much extremely quite slightly neither slightly quite extremely Please verify that you have made 13 check marks in this booklet, and then place the booklet in the envelope matching the code that appears on the first page of the booklet. Thank you. .. _ w..- -i l 4 ..fi.....-... 7 . I I ‘ M I ' . xl‘; v ".' 1 LI. ., ‘ —--—‘ ....—.—. 7*-—~”--—”4--~- O - >n ->-<-l—-Q---4 . . 'l , .. - l . ‘. ‘ 1 I I«. H ‘ »’ . . . ‘z ' ~ _ p ~_‘. ‘ .‘ ’ .: - i- ii.) a . l - ' : '.' " l li)‘ .' . . -..._ - .. - -. ....-- . _ ... l y. ,( I 4 A 1 . ( i - ‘ . . y! . l A ._._. . . .....- 1‘! .. x . APPENDIX D PRETEST DEPARTMENTAL PRODUCTIVITY DISTRIBUTIONS 99 100 PRETEST PRODUCTIVITY DISTRIBUTIONS*_ Job Classification Mean Standard Deviation N Verifiers 15189 2792 60** Key Punchers 10776 1890 63*** Beginners: Month 1 5281 1U52 46 Month 2 6790 1625 “5 Month 3 7172 1723 41 Month U 7902 1807 H2 Month 5 8067 1949 “0 Month 6 8957 1797 39 Month 7 8409 2053 38 Month 8 9105 2262 37 Month 9 8810 1573 36 Month 10 9023 16h0 39 Month 11 9909 1730 33 Month 12 95U9 2073 26 *Observational periods: verifiers and key punchers, November and December, 1965, January, 1966. Beginners: 1961-1965, by consecutive months, first year on Job. **N of 60 based on 20 subjects, 3 observations each. ***N of 63 based on 21 subjects, 3 observations each. "I7'111'1111111'1‘11115$