_._....., ................ CHOICE AND PERCEIVED AUDIENCE I ' ATTITUDE AS DETERMINANTS OF . COGNITIVE DISSONANCE AND A SUBSEQU—ENT ATTITUDE CHANGE FOLLOWING COUNTERATTITUDINAL ADVOCACY} 1:5?{1 MICHIGAN STATE UNIVERSITY EDWARD M BODAKEN TheSIS f6r the Degree 6f Ph D f I976 """ I‘H 55‘s LIBRARY Michigan Sub- University This is to certify that the thesis entitled CHOICE AND PERCEIVED AUDIENCE ATTITUDE AS DETERMINANTS OF COGNITIVE DISSONANCE AND SUBSEQUENT ATTITUDE CHANGE FOLLOWING COUNTERATTITUDINAL ADVOCACY presented by Edward M. Bodaken has been accepted towards fulfillment l of the requirements for Mdegree in COIIImunication flax/(Z; if 1/21” (”if Major professor Date April 3, 1970 0-169 .- llflflfl SI K m IIIE. LIIIAII' INDIE. mnn. IIIIIIAI CHOICE AND PERCEIVED AUDIENCE ATTITUDE AS DETERMINANTS OF COGNITIVE DISSONANCE AND SUBSEQUENT ATTITUDE CHANGE FOLLOWING COUNTERATTITUDINAL ADVOCACY by Edward M. Bodaken This study sought to assess the effects of choice and audience commitnent on attitude change following encoding of counterattitudinal communications. Students enrolled in undergraduate communication courses at Michigan State University were administered a pretest attitude questionnaire containing several issues considered to be salient and ego—involving for university students. The issue producing the most skewed distribution called for compulsory on-campus residency for all college students; thus, it was chosen as the experinental issue. Subjects were randomly assigned to experimental conditions. Approxinately two weeks after the pretest, experimental subjects encoded counter— attitudinal essays under varying conditions of choice and audience attitude toward the issue. Post~encoding measures were then taken of each subject's attitude toward the issue and the perceived difficulty in encoding the essay. The effectiveness of each manipulation was assessed by asking subjects to respond to items tapping subjects' per— ceptions of choice and audience attitude toward the topic. Subjects in a control condition completed pretest and posttest instrunents. The choice variable was dichotomized into Free Choice and No Choice levels. Subjects in the Free Choice conditions were inforned of teat-foe i‘Jm‘HJ yd Edward M. Bodaken their freedom to comply at specified points in the experinental induc— tion. Subjects in the No Choice conditions were not given such freedom. Audience attitude toward the topic was dichotomized into Committed Audience and Uncommitted Audience levels. Subjects in the Uncommitted Audience conditions were told that their essays would be presented to an audience composed of high school students who held no opinion on the experinental issue. Subjects in the Committed Audience conditions were told that their essays would be presented to an audience composed of high school students who favored the proposition that all college students should be required to live on campus. The major dependent variable, attitude change, was measured by use of subjects' pretest to posttest attitude change ratings of the topic on four, seven—interval semantic differential—type scales. In addition, the counterattitudinal messages were analyzed on neasures of persuasive— ness, organization, attitudinal direction, and essay length. Three hypotheses were tested in this investigation: First, that attitude change would be significantly greater in Free Choice than in No Choice conditions; second, that subjects in the free Choice: Uncommitted Audience condition would demonstrate significantly greater attitude change than individuals in all other conditions; and third, that attitude change demonstrated by individuals in the No Choice: Committed Audience condition would be significantly less than change demonstrated by subjects in all other conditions. The results failed to confirm the research hypotheses. However, a significant audience effect was found, indicating that subjects who Cd'fd. hash-Q". I“! I-fhlr-I fiat" ME: FM m an“ a ' 73-“T"'--'" - “"u'u‘ -~::'~'".:-5":-.H." I---'r~ ..-:-. .‘. M." 3-12. fianaiw mm. A .— . .-. _ . - -..I .- _- .. dusty-I 3.5” I‘DZ'DOiI, Edward M. Bodaken encoded counterattitudinal communications directed toward an un- conndtted audience demonstrated amounts of attitude change significantly greater than those subjects who encoded such nessages for a cmnmitted audience. Analyses perforned on message persuasiveness, organization, and encoding difficulty failed to demonstrate differences among the experinental groups. Judgments by independent raters indicated that the essays were counterattitudinal fer the subjects. Finally, analysis of essay length revealed that subjects in the Uncommitted Audience conditions wrote significantly longer messages than did those of the Committed Audience conditions. Analyses of subjects' ratings indicated that the levels of choice were successfully manipulated. The nanipulation of audience attitude was successful in all experimental conditions except the No Choice: Conndtted cell. The results were discussed in light of dissonance formulations and a number of suggestions were offered for future research in counter— attitudinal advocacy. I ' . . . _' : -'-. "II-ill»; _ '_.-I : _ fig-I l} I”, ' . f ' I. =‘,_II1. r -':_ .- A; ‘ .6.“ _ . - I}: 'I . I I. r 1%.”; u, ‘.' . I .f ,l . mohair-taint.- . r.-. " -'.' pmm Lag/.53“ 'IT- await-m3 not!“ .M' . -.;-m.-. -- -':r.'=‘-‘??I.- - - + awn": .slv'l'rtrme gm - . ‘_ . m.— ... - “Fur 1 1-. '.. . - : . . . _ u” “I; .‘I-lt'. CHOICE AND PERCEIVED AUDIENCE ATTITUDE AS DETERMINANTS OF COGNITIVE DISSONANCE AND SUBSEQUENT ATTITUDE CHANGE FOLLOWING COUNTTRAETITUDINAL ADVOCACY By Edward M 3 Bodaken A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Communication 1970 .. ,, .. group-I» ‘ n3- _ II- 1“! I. .t' in. R. W... I I It fl hr ”lulu-II.“ I. ..I.Inl.lln Ital-IE.- .l. .| l . . . .. I n I III .kr ll I..lll||l.|l||||..lll. I u . . . I I I . . . . II Illlnul I I III In . IIIII v Accepted by the faculty of the Department of Communication, College of Communication Arts, Michigan State University, in partial fulfillment of the requirements for the Doctor of Philosophy degree. flm/flg «3? Wt/flfl/r Director of Thes1s Guidance Committee: /:é//a_//f{ é %%fl’ , Chairman ,7 a ,7 /‘> 7 u ., ..:-__ ,Ifw __ ' - -. "FL-:3?! ._l".*':". {9.1! *i‘ii-u'fiiwl. ' -.‘1¥ MM (WM-hm IJ_ ‘- If; I... . .- . ._ . ,'..l': -- in Li- I -. ’5‘: H l‘fi'l’! ? '- :|- {LII-F 'd J -..-v.- as ._ ..— . 'I-‘."‘I-III-.I .—.-—. ._- _ DEDICATION To ACKNOWLEDGMENTS The author wishes to express his sincere appreciation to Dr. Gerald R. Miller fOr his assistance in all phases of this study. More than any other individual, he has made graduate study a rewarding experience and a fUture in the study of human communication behavior an exciting prospect. For conscientiously serving as members of the guidance committee, the author acknowledges Dr. Eugene Jacobson, Dr. David Ralph, and Dr. Lawrence Sarbaugh. The advice of Dr. William B. Lashbrook is also acknowledged. Dr. John R. Wenburg provided friendly motivation and intellectual stimulation as a graduate colleague, and his assistance in the pilot study was invaluable. H. Michael Burgoon‘s candid and competent evaluations as the study progressed are gratefully acknowledged. The services of Frank Millar as resident sounding board are also acknowledged. For their assistance as experimenters in both pilot and final studies, the author acknowledges the following colleagues: L. Blaine Goss, R. Samuel Mehrley, C. Edward Wotring, and Frank Zink. The author wishes to acknowledge Ruth Langenbacher for the rapid and professional service in the typing and final preparation of the manuscript. In addition, there are those who provided a more personal kind of assistance and for them the third person seems inappropriate. iii My parents, while not Imowing the exact nature of graduate study, consistently offered encouragement and moral support during the course of those studies. My wife De Ann and our children Matthew, Amy, and Christopher, made the work necessary for this degree worthwhile. Their loyalty and independence during these years of graduate study gave me the opportunity to work without distraction and the motivation to succeed without qualification . iv . 's.::':._w.* 5'1"."- 1'Im". ,ignr. Marita-Ti .-.-.-.-1..:,I b mm {.115 nnl d gang '1 I ' II .-".Ie.'._'-.-'_ -"' A113". {‘13: -- -.--. if? 1:): .maag-m 31-“. 9+, rh- -I-.-i TABLE OF CONTENTS ACKNOWLEDGMENTS . . . . . . . . . . . LIST OF TABLES . . . . LIST OF APPENDICES . . . . . . . Chapter I INTRODUCTION . . . Statement of the Problem . . . Review of Relevant Research . . . Choice . Audience Attitude Toward the Topic Rationale and Hypotheses II METHODS AND PROCEDURES Overview . . . . . . . Pretest . . . . . Independent Variables Inductions and Posttest Dependent Variables . Manipulation Checks . III RESULTS Pretest . . Experimental and Control Differences Manipulation Checks . Choice . Audience Attitude . . Test of the Hypotheses . . . Hypothesis 1 . . . Hypotheses 2 and 3 Supplementary Analyses . H mercy—1 |_l —"I.-.—- Chapter Page ‘ IV DISCUSSION......... .. . .3u 1 Implications ForFutureResearoh. . . . . . . . . '43 o APPENDICES...................H9 vi Table 10 LIST OF.TABLES Means, standard deviations, and analysis of variance of pretest attitude scores for experimental and control groups . . . . . . Attitude scores, anount of change, and analysis of variance of attitude change scores of experimental and control groups . . . Analysis of differences of attitude change scores between each experinental group and control . . Pbans, standamd deviations, and analysis of variance choice ratings . . Comparisons of nean choice ratings of experimental groups . . . . . . . . Percentage and frequency of subjects accurately per— ceiving the audience condition. Italics indicate intended inductions . Attitude scores, anount of change, and analysis of variance of attitude change scores of experimental groups . . . . . . . . . . . Comparisons of mean change scores of experimental groups . . . Means, standard deviations, and analysis of variance difficulty ratings Pbans, standard deviations, and analysis of variance essay length . . . . . . . . . . . vii Page 2a 25 26 27 28 LIST OF APPENDICES Appendix Page 1 A PRETEST ATTITUDE QUESTIONNAIRE . . . . . . . 50 B POSTTEST ATTITUDE QUESTIONNAIRE . . . . . . . 51% C CONTROL POSTTEST ATTITUDE QUESTIONNAIRE . . . . 58 CHAPTER I INTRODUCTION Statement of the Problem Generally, investigators concerned with counterattitudinal advo— cacy have concentrated on the effects of varying amounts of incentives used to induce a subject to encode a belief-discrepant communication. The foundation for such research can be found in Festinger's (1957) original formulation of the relationship between the amount of incen— tives and the nagnitude of dissonance. the magnitude of the reward or punishnent, that is, the attractiveness and desirability of the offered reward or the unpleasantness and the undesirability of the threatened punishment is an important deter— minant of the magnitude of dissonance which exists once compliance is exhibited. Tbo great a reward or punishment will result in only little dissonance. (p. 91) Taken collectively, the results of the Heny studies investigating the relationship between justification and attitude change are far from conclusive. For exanple, the classic Festinger and Carlsmith (1959) investigation demonstrating an inverse relationship between the two variables was replicated by Cohen (1962), reversed by Janis and Gilnore (1965), and methodologically decimated by Rosenberg (1965). Central to Festinger‘s (1957) theory of cognitive dissonance is the variable of choice. It is only when the individual makes a choice that he experiences dissonance. Festinger writes: 2 after having made the decision he is no longer in conflict; he has made his choice; he has, so to speak, resolved the conflict. He is no longer being pushed in two or more directions simultane- ously. He is now committed to the chosen course of action. It is only here that dissonance exists, and the pressure to reduce this dissonance is not pushing the person in two directions simultaneously. (p. 39) From this formulation, it follows that the amount of dissonance and the subsequent attitude change are functions of the extent of the individual's commitnent to the chosen course of action. For if the individual feels no real commitment to or involvement in the choice, the decision will produce little dissonance. On the other hand, if the individual has freedom of choice, one would expect a positive rela— tionship between commitnent and dissonance. Brehm and Cohen (1962) suggest that an individual's perception of control over his own be— havior is a mediating variable in generating greater dissonance. They term this perception, or feeling, volition, and they define it as the "conscious control of one's own behavior." (p. 201) Volition also fixes responsibility for the consequences of an act. Stated another way, if an individual feels that he freely engaged in behavior with harmful or unpleasant consequences for himself or others, the magnitude of dissonance should be positively related to the degree of volition and the perceived potential harm of the behavior. Presumably, most dissonance experiments have offered subjects some choice as to whether or not they will perform belief—discrepant behaviors. Generally, however, the choice, or volition, operational procedures are not replicated from one study to another. for example, the method of cajoling subjects to perform in the Festinger and .- . .—‘- q.- v . I .r . 1‘ .1 1 n . . .. 1 . . .- .[ I .n. .A . "II. ._ . I. h. . ... 1. . ul- na .. III-ll in . v . I i u I 3 Carlsmith (1959) study (e.g. "It will only take a few minutes.") was quite different from.the method employed by Janis and Gilmore (1965). In the latter study, subjects were visited in their dormitories by the experinenters. It seems reasonable to assume that subjects in such a situation would perceive less threat or force than the Festinger and Carlsmith psychology laboratory setting. At any rate, the role of choice in counterattitudinal advocacy is obscured by the paucity of reproducible manipulations. Moreover, investigators have not systematically tapped the con- sequences dimension of the chosen alternative. Collins (1969) suggests that dissonance may exist if an individual feels personally responsible for potential harmful or aversive consequences. Aronson (1968) states that it is incorrect to say that the cognitions, "I believe the task was dull," and "I told someone the task was interesting," are dissonant. Rather, the following represent dissonant cognitions: "I am a decent truthfhl human being," and "I have misled a person; I have conned him into believing something which just isn't true; he thinks that I really believe it and I cannot set him straight because I probably won't see him again." (p. 29) If one adds the cognition, "I freely chose to en— gage in misleading behavior," dissonance should be greater. The preceding discussion suggests a refbrnwlation of the justifi— cation—attitude change question. The overriding concerns for an indi— vidual faced with a choice may be his perceived freedom of choice and the consequences of his act. Furthernore, the choice itself may serve as a reinforcing contingency for the act, so that justification may be subsumed by choice. Existing research on counterattitudinal advocacy in gen-.155. qua- 'm- 'oméaiip ' 'J - . . .~ .‘i-‘i I.) _. ‘ ‘ . | = -"“-'-’-:.‘-.--' ‘v-r 'xr..=-'-- er simmer: a... 3'3 . 1“ ..... ‘ ' may.“ '10.?! 917.902!!! Mum; 'r'umfh --. ' '--.:.' I I" 3?— - - - '- ' “ minke) n has not focused on the relationship between freedom of Choice and possible aversive outcomes for a Source's audience. The present study is designed to test such a relationship. Review of Relevant Research Choice Nearly a decade ago, Cohen (1960) indicated that the perception of commitment to choice could serve as a necessary pre— condition for the creation of dissonance; Where choice is varied, expectations from dissonance theory are fulfilled only under high—choice conditions; under low—choice conditions, straightforward motiva— tional or resistance effects seem to account for the results. (p. 306) More recently, Collins (1968) has termed the manipulation of choice the "classical dissonance manipulation." (p. 822) The following discussion will focus on research which has yielded results relevant to the role of choice in producing dissonance. Cohen, Brehm, and Fleming (1958) hypothesized that the greater the justification the less the magnitude of dissonance and subsequent attitude change. Relevant to this discussion is their rationale for failure to confirm the study's hypothesis: A11 Se in both conditions complied. It may thus be inferred that the general force to comply was very compelling, and little attitude change is therefore to be expected. (p. 277) In fact, the instructions given subjects by Cohen et_a1. were very persuasive and emphasized such factors as the positive qualities of considering the other side of the issue, the benefits to research, and the fairness of the task. This type of justification manipulation tends to reduce the individual's perceived choice. But, more importantly, 5 it provides the subject with an intellectual exercise which minimizes commitment and thus lessens dissonance. Brehm and Cohen (1959) directly manipulated the variable of choice in an attempt to answer two questions: Can a chance event affect the magnitude of dissonance?; and Does the effect of such a chance event depend upon there having been a prior relevant choice. They hypothesized that under greater degrees of choice, increasing rela— tive deprivation results in increasing satisfaction. Subjects were told that an experiment was being conducted and that the task involved was not very interesting (copying random numbers). Schedule sheets were 1 then distributed so subjects could indicate their free hours. In the high choice condition, subjects were told that if necessary they could be excused; in the low choice condition, subjects were not told of the possibility of being excused. The schedule sheets were collected and subjects were given a questionnaire ostensibly to evaluate the general issue of participation in research. Relative deprivation conditions were induced through the questionnaire instructions. Subjects were told that while most persons would receive payment for the work, a few randomly chosen people would receive nothing. The instructions concluded by noting that the subject would or would not receive payment depending on whether a check—mark appeared on an appropriate line. All subjects were led to believe that they would receive no payment. In the low relative depriva— tion condition, subjects were told that those being paid would receive $1.00; in the high relative deprivation condition, the amount was $10.00. Manipulation checks indicated a significant difference between deprivation conditions but no significant difference between choice 6 conditions. The researchers then analyzed the data according to per— ceived choice. Under conditions of low perceived choice, the more relative deprivation, the more negative were attitudes toward the fait ; accompli; while under conditions of high perceived choice, the more ' relative deprivation the more positive were attitudes toward the fait V accompli. This choice manipulation would seem to be entirely consistent with Cohen's (1960) thesis concerning the extent to which self—selection becomes a problem when subjects are able to refuse to participate. He argues for presenting the illusion of choice but cautions against letting the subject leave the experiment. The difficulties experienced by Brehm and Cohen (1959) in presenting such an illusion forced them to analyze the data on the basis of subjects' perceptions rather than experimental assignment. Therefore, their findings are, by admission, somewhat equivocal. A different approach to the problem of choice was taken by Brock (1962). Incorporating Zajonc's (195”, 1960) cognitive structure measure— ments and Festinger's (1957) theory, Brock was interested in the effects of choice and confrontation on subsequent attitude change. Subjects were administered questionnaires on the issue of conversion to Catholicism. In addition to the questionnaire, each subject completed the following sentence on 16 slips of paper: "For me becoming a Catholic would mean:." After completing the sentences subjects were asked to group the slips of paper. Also, they were asked to specify relations (bonds) among the slips. One week later, subjects were contacted and asked to participate in a follow—up study on the issue. The task (an essay on "Why I would 'l .. " :. is” . up)?“ , ~ _. If??? "i" "L's-ta: swim m. &M:' ' a 1' " " _. . . . : ' n '4‘. L! i“: W ' I. "i I - ‘hZIUi‘WE‘T "ii" In -*::-.u'.fl'hm -flam and? 2....' ' i: - _: "' '1 at ' .151 ~ _ ..'_ I ' L I ' 'i ' ‘ -'.....;.- 9w». um .-:....-.-su.ms..5 sum _ _ _ . . g I: . T'n' 7-;- I 7 like to become a Catholic.") was presented to subjects and the choice inductions were administered. Subjects in the high choice condition were told that the task was voluntary and that they were under no obli— gation to write the essays. Each subject was asked individually to participate and each said "Yes" at least once. Subjects in the low Choice condition were not given this option. High confrontation was established by asking the subjects, after they had written the essays, to rewrite them, ranking the sentences in terms of their originality and persuasiveness. Subjects in the low confrontation condition were told to rewrite their essays according to a syllables chart. Manipulation checks indicated that high choice subjects reported significantly less obligation and more option than low choice subjects. Subjects in the high confrontation condition reported significantly more awareness of meaning and more deliberation than low confrontation subjects. Also, the results indicated that more attitude change was obtained under high than under low choice conditions. Finally, given high choice, mcme favorable revaluation occurred under high than under low confrontation; given low choice, the effect was reversed. These findings support a positive relationship between choice and attitude change. Moreover, the interaction finding suggests that when an individual is forced to perform a belief—discrepant act under conditions of confrontation, he will resist a revaluation of cognitions in the counterattitudinal direction. Linder, Cooper, and Jones (1967) attempted to resolve the contra— dictory findings of the Cohen (1962) and Rosenberg (1965) studies, claiming that the latter study failed to offer the subject freedom not to comply. Rosenberg's investigation attempted to eliminate evaluation 8 apprehension on the part of the subject by separating the belief— discrepant task and the evaluation of the attitude object. When Rosenberg introduced this methodological change, he found an incentive effect: attitude change and justification were positively related. Linder et_al, (1967) argued that Rosenberg's findings could not be compared with Cohen's because different conditions existed. Furtherh more, they claimed that the crucial difference between the Cohen and Rosenberg studies was the freedom not to comply. Therefore, they con— ducted two experiments to determine the effects of choice in interaction with justification. In the first experiment, similar to Cohen's (1962), subjects were assigned to conditions of high or low incentive ($2.50 or fifty cents), and no—choice or free—choice. In the free—choice condition subjects were told that the decision to perform the task was their own and they could leave if they desired. In the no—choice conditions, no mention was made of such freedom. Analysis revealed a positive relationship be— tween attitude change and incentive in the no—choice conditions and a negative relationship in the free-decision conditions. Linder et_al. demonstrated the possibility of testing both dissonance and incentive predictions within the same paradigm. They also demonstrated that the effectiveness of testing these theories rests on the importance of choice in compliance. The second experiment attempted to illustrate that a more precise method existed for removing evaluation apprehension from an experiment. The variables remained the same, but the design was changed to a replica— tion of Rosenberg's (1965) method. Linder_gtlal. claimed that subjects whim m: bmfi d jhflu‘h 130%“ g -' nr.¢ -' .. _- a: 1 _y_- III-:m .“.~_‘”'_",|_L‘.qu1u’[ ~15 W' “MHZ. ' . --:'2 H" 1 art-5 ("5"??- . . .13 g: «sixth! :i: a 9 in the Rosenberg study found it difficult not to comply because of various commitments they made. In Rosenberg's study the subject strengthened his commitment to participate in "another little experi- ment." Iinder et_al, eliminated this possibility in Experiment II by indicating to the subjects before they moved to the other experiment that they did not have to participate. Subjects were asked to write essays in favor of i§_1999_parentis regulations. Analysis of the results revealed a significant inverse relationship within the free-decision conditions. Within prior commitment conditions a significant positive relationship was found. Taken together, these two experiments strongly indicate that decision freedom is an important determinant of dissonance under con— ditions of counterattitudinal advocacy. Moreover, it seems plausible to conclude that, all other things equal, the greater the freedom to encode belief-discrepant communications, the greater the dissonance and subsequent attitude change. Audience Attitude Toward the Tgpic The implication underlying dissonance arousal in such studies as Festinger and Carlsmith (1959) is that "Holding an opinion implies telling it to an interested party." (Abelson, 1968, p. 130) The notion here is that the arousal of dissonance and the choice of an appropriate mode of reduction are not processes that can be limited to the circumstances explicitly defined in the experi— mental paradigm. Whether the subject is encoding belief—discrepant essays or delivering a counterattitudinal speech before a video tape camera, he will probably be aware of some consequences of his behavior. If his communication is perceived to be anonymous, the aversive conse— 10 quences should be perceived as mdnimal. Carlsmith (1968) asserts: anytime a person makes some statement counter to his attitudes, and a listener (whose opinion is important to the speaker) is unaware of both the speaker's attitudes and his motivation for speaking against these attitudes, dissonance will be aroused. (p. 806) Thus, if a subject perceives his audience to be persuasible and unaware of his private opinions, dissonance should exist. The present study will not be concerned directly with the private—public aspects of a counterattitudinal act. Conflicting evidence from several studies suggest that this variable may confound more than anything else. Collins (1969) concludes his review of research in the area of public— private encoding as follows: The . . . results lead me to abandon the notion that the public—private dimension is the crucial one for the production of a dissonance, negative relationship between financial inducement and attitude change. (p. 218) While this conclusion refers to financial incentives, there is no reason to think that the absence of such incentives and the manipula— tion of choice and audience attitude will serve to clarify the public- private dimension. Inextricably bound to a consideration of research relating to consequences of belief—discrepant behavior is the relevance of the self- concept. Bramel (1968) suggests that the self is affected as a result of concern for what others may think. He writes: The theoretical relevance of the self in this view of dissonance theory now becomes clearer. . . dissonance is a feeling of personal unworthiness (a type of anxiety) traceable to rejection of oneself by other people either in the present or in the past. Any information which implies that one is incompetent or imnrral arouses dis— sonance. The reason dissonance is greatest when the person feels personally responsible for his behavior is /__- 11 that rejection by other people is usually greatest when they believe the person voluntarily acted in an appro- priate way. (p. 365) Therefore, of interest in self—persuasion research are results suggesting the importance of self and its implications for the percep— tion of possible harmihl effects that may occur as a result of the counterattitudinal behavior. Carlsmith, Collins, and Helmreich (1966) manipulated role playing, experimenter bias, and monetary incentives to test dissonance and in— centive theory predictions in the same paradigm. Subjects were asked to strike out certain digits in several booklets of random numbers. After an hour, subjects engaged in either essay writing or role—playing conditions. In the former, the subject was instructed to write essays detailing the interesting aspects of the experiment. He was told that no other subjects would read the essays, that the experimenter would use them as source material. In the role-playing condition, the subject was asked to describe the task as interesting to a waiting student (confederate). In both conditions, subjects were told that since the experimenter was "in a bind" and needed help, he could pay them (fifty cents, $1.50, $5.00). After completing the task, the subject was led to another room under the guise of participating in a different research project and posttest measures were obtained. Carlsmith e: 31. hypothesized a dissonance effect in the role— playing condition: the greatest attitude change should occur in the fifty cent condition. An incentive effect was predicted for the essay writing condition: the greatest attitude change should occur in the $5 condition. Results confirmed both hypotheses. Relevant to this discussion was the _. - I - ' -.;..-.r_.-._- 5-3.1 -:s.--'- .wy§.=.7~i'..."r.'..:ri : ' ..-.u.- Us? 1.; W It. - --L"'"' _.-. J5 .-_. 3-. - ' "a". [41.71111 "JCTME. 12 the finding that increased pressure for compliance yielded less attitude change in the role—playing condition. These researchers suggest that the social consequences of the role—playing were much more dissonant- producing because the subject believed the audience (the confederate) perceived him as sincere when he described the dull task as exciting, interesting, and enjoyable. Major methodological differences preclude a great deal of reliance on the findings. For example, the subjects in the role—playing condition were given two minutes to describe the task as compared to the maximum 17 minutes allowed in the essay condition. According to the authors, subjects in the role—playing condition performed under more "hectic" or "crisis" circumstances than essay subjects. At any rate, there is some basis for concluding that implications of social consequences deriving from the belief—discrepant act are determinants of dissonance. More relevant to the notion of harmful consequences as a determi— nant of dissonance is the Helmreich and Collins (1968) study in which financial inducement and commitment were manipulated. Subjects were given a card with several arguments supporting the proposition that family size should be limited by the government. Subjects were assigned to one of three commitment conditions: no—takeback video, takeback video, anonymous audio. In the no—takeback condition, subjects were instructed to deliver a speech on videotape. Furthermore, they were not instructed to tell anyone their reasons for recording the speech nor their private opinions. Subjects in the takeback condition were told the audience who would see the speech would be debriefed and were asked to make a second tape telling the audience why they made the tape and revealing their l3 true private opinions. Subjects in the anonymous audio condition were asked to performlthe counterattitudinal task, but they were told their names were not needed. Financial inducement was either fifty cents or $2.50. The findings supported Helmreich and Collins' first hypothesis: a dissonance effect was obtained in the no-takeback video condition. Their second prediction was not confirmed; there was no difference be- tween the two video conditions. HOwever, with inducement conditions collapsed the no—takeback video condition was significantly higher (more favorable post—task rating) than the takeback video and the anonymous audio conditions. Thus, this study supports the assertion that subjects who are strongly (publicly)‘committed to a belief-discrepant act experi— ence dissonance. Moreover, the manipulation used to induce the levels of commitment was handled in suCh a manner that subjects believed the tapes would actually be used to change the attitudes of others. Subjects in the no—takeback condition experienced dissonance because they were faced with cognitions relating not only to the performance of belief— discrepant behavior, but also the knowledge that their arguments would be used to persuade others. In addition, they may have perceived that they would not be able to tell the audience their real positions on the issue. Added together, these cognitions could create a magnitude of dissonance that would be reduced by subsequent attitude change. Steiner and Darroch (1969) had subjects deliver short counter— attitudinal speeches opposing draft deferments for college students. Public and private role assignment and agreeing and disagreeing audiences were the independent variables. Relationships between attitude change and subjective appraisal of performance were examdned. No significant 11+ differences were reported, but one finding that approached significance is of interest. As noted earlier, subjects who believe that their audience is unaware of their motives or real opinions should experience greater attitude change than those who are allowed to take back their acts. Steiner and Darroch's findings reveal just the opposite trend (p < .10): subjects who were privately assigned (audience unaware of position) demon- strated less attitude change than those publicly assigned (audience aware of position). It should be noted that encoding time was not controlled: the speeches varied from one-half minute to seven minutes. Nevertheless, this trend is at odds with the Carlsmith g al. (1966) conclusion that subjects who deceive an audience concerning their true beliefs should ex— perience more dissonance than subjects who do not engage in such deception. A recent study by Nel, Helmreich, and Aronson (1969) offers dra— matic support for an aversive consequences position. These researchers varied financial incentives (fifty cents, $5), and audience attitude to— ward the topic (opposed, in favor, no opinion). Subjects were given a card with several arguments advocating the counterattitudinal issue (legalization of marijuna) and instructed to deliver the speech on video tape. Nel e1 ai. hypothesized: (a) when addressing an audience of un— committed individuals, subjects undergo more attitude change in a low inducement condition than in a high inducement condition; (b) the strongest negative effect (i.e., dissonance) will be in the uncommitted audience condition; and (c) the condition most likely to have a positive relationship between attitude change and financial inducement is the condition where the audience is already committed to the position advo— cated. While subsequent analysis failed to yield significance, some a .’-" II :.--5. 1L; ' road >145? 13.: bah-1.15 ml: min M “E w , ' ':-.-.*: 512*. g- . .- a‘.-.1."~.-:."! .“rbcrrxdi bi!" - .. L. . "~ 11ft. . ' ‘ .f‘ _-I _ , ~ _- - __ ... ... _... 0,5. amm**+ ,1: .. " - '. l ‘ u - - €- - 0 - 3- Du? 933 a ’ ' .. ‘1': . 3...,- _ 15 priori comparisons indicated that a significant difference existed be— tween the fifty cents and $5.00 uncommitted audience conditions. The interactions were nonsignificant. The most relevant finding fer this discussion concerns the inverse relationship in the uncommitted condition. Nel et_a1, conclude: Our reasoning was that the anticipated evil results of the subject's counterattitudinal advocacy should be most damaging to the self—concept when the behavior is elicited under conditions of minimal justification and should lead to dissonance reduction through internalization of the advocated communication. (p. 12”) Some possible procedural difficulties may have militated against 3 more consistent support of the hypotheses. First, subjects were given little indication of choice until after they had agreed to perform the task. That is, using a two-experiment guise, subjects complied with "Experimenter 2" and moved to another room. This commitment was then rewarded by the fifty cent or $5.00 incentive. Secondly, the use of a prepared speech outline could have influenced the results, for subjects in the committed conditions might not think it worthwhile to incorporate any new arguments into the outline. By contrast, those in the uncommitted condition could well be persuaded more by the prepared arguments than by the encoding act itself. In other words, a direct persuasive effect could occur when subjects are given several important arguments. Nevertheless, Nel e:_al. have provided support, albeit equivocal, of the consequences position. Moreover, their operationalization of audience attitude is a worthwhile methodological addition to the self— persuasion research. “'51:! Wufiflme I it"! V1.1: graffiti) We“! rm “I“. . . :.. ___._ I, _- ._, . .. w.“ . 1‘ I .- .- . - '-. I '9 . '. $3.52.“; w m- 1 ..A_ ”31145” 12.. " “9...... n . _- fiL") _ 5-.II I I. - ”.34 _' _ . r. . ' ud- . I 16 j Rationale and Hypotheses Based on the research findings which demonstrate a positive relationship between choice and dissonance in a counterattitudinal en— coding task, the following hypothesis will be tested: Hl: Persons encoding counterattitudinal communications under conditions of free choice will demonstrate greater attitude change in the direction of the position taken in the communication than persons encoding counterattitudinal communications under conditions of no choice. Moreover, the dissonance-producing effect that exists when a counterattitudinal communication is directed toward an uncommitted audience suggests the second research hypothesis: H2: Attitude change in the direction of the position taken in the counterattitudinal communication will be greatest when the communication is encoded for an uncommitted audience under conditions of free choice. Finally, since conditions of no choice and committed audience should yield the least dissonance, the following relationship is pre- dicted: H3: Attitude change in the direction of the position taken in the counterattitudinal communication will be least when the communication is encoded for a committed audience under conditions of no choice. Support for Hypothesis 1 derives fromlthe research of Brehm and Cohen (1959), and from the theoretical notions of dissonance theory. Obvious difficulties arise in directly manipulating choice, which may account fOr the bulk of literature that indicates a more indirect testing of the variable. The major problem is one of self—selection; by directly manipulating the variable, the possibility exists that those who choose to stay are able to perform the behavior without experiencing dissonance. l7 Ashmore and Collins (1968) suggest that a subject may not change his attitude simply because the belief—discrepant act may fall within his range of acceptable behavior. However, given that the subject believes the decision to be his own, he should experience dissonance. For not only is he encoding a communication that is at odds with his private opinion, he has freely chosen to do so. Upon agreeing to perform this belief—discrepant act, he will choose a mode of dissonance reduction that will allow him to justify the behavior. A probable method of dis— sonance reduction, assumdng the plausibility of the induction, is a change in his attitude toward the issue he is advocating. l Hypothesis 2 is based on the findings cited earlier which suggest that dissonance is heightened when probable harmful consequences are perceived by the subject. A subject encoding a belief—discrepant message for an uncommitted audience should experience more dissonance than a subject performing the same act fOr a committed audience. Festinger's (1957) original formulation of dissonance theory states that the magnitude of dissonance is a function of the importance of the opinions or behavior. It seems reasonable that encoding a message——at— titudinally—discrepant or consistent——will be more significant to the encoder if the probability is high that the message may affect others. Given both a belief-discrepant act and a persuasible audience, the act should take on even more significance. For not only does the subject perceive the act as belief—discrepant, he also perceives that he may influence others to accept the counterattitudinal position. Add to this the fact that he cannot take back the act or inform the audience of his private opinions and one can predict a substantial magnitude of dissonance. -;...'1'.'- l..."‘ =.':'-"-“*t-:‘I3 nut-«1w: are mm mm (and! -- ~ , .r-rraa-aath Inny‘m .193 Rhoda ad .1110 an! ad (I? "-.. _- _ * :.-.-,-..i.mm:n 5 pm I . 1;, -._' V'f‘. 35f! 5” I p- l8 Hypothesis 3 predicts that under conditions of no choice and audience commitment, subjects engaged in counterattitudinal advocacy should experience little, if any dissonance and subsequent attitude change. For subjects will perceive the act to be without volition and the responsibility of another. Dissonance, if any, may be resolved by derogating the experimenter, accepting the task as an intellectual activity, simply tolerating inconsistency, or some combination of these modes. Moreover, these subjects may simply assimilate the behavior to their own opinions and recognize that such activity is part of some be- havior peculiar to experiments. The subject should not experience dis- sonance as a result of potential harmful effects for two reasons: first, aversive consequences that may occur cannot be the subject's responsi— bility since he assumes no responsibility for the behavior leading to such consequences; second, the audience should not be perceived as per— suasible, since it is already on record as supporting the belief—dis— crepant issue. Therefore, because the subject does not see the act as his own and because the probability of harmful consequences to his audience is low, he should not experience dissonance or subsequent attitude change. While predictions are not made concerning attitude change in the Choice: Committed and No Choice: Uncommitted cells, it follows from Hypotheses 2 and 3 that subjects in those conditions will experience moderate dissonance and will demonstrate amounts of change ranging be- tween the Choice: Uncommitted and No Choice: Committed conditions. 1 .- dim“ in: 1mm 'I I lit taut-.215 m- u , ' fi' ‘i " _ , .L ‘5 r‘s- _ _ ' _ p i _ ”Ir-.5 manila-v “Ian-415: r; a? .+--.- - '.. mats-sq LI- ' amt-1118 “I CHAPTER II METHODS AND PROCEDURES Overview On the basis of pretest results, an issue was chosen which was counterattitudinal for Ss. Ss were randomiy assigned to experimental and control groups. In the experimental conditions, Ss encoded belief— discrepant essays under varying conditions of choice and audience attitude. Ss then completed post—encoding questionnaires. In addition to containing attitude items dealing with the experimental issue, the posttest questionnaire included items designed to measure perceived choice, perceived audience attitude, and degree of difficulty in en— coding the messages. Ss in the control condition completed pretest and posttest instruments. Pretest Students in undergraduate communication courses at Michigan State University were administered a pretest questionnaire, ostensibly to solicit student opinion on possible issues to be used in a class project. (See Appendix A) The instrument contained several issues believed to be salient and somewhat ego—involving for undergraduate students. Each issue was followed by four, seven~interval, semantic differential—type scales bounded by the adjectives good—bad, valuable—worthless, pleasant— unpleasant, and fair-unfair (Osgood, Suci 8 Tannenbaum, 1957). Issues were scored by summing across the attitude scales on a one low, seven 19 20 high basis. Thus, with four scales per issue, the possible range was from four to 28. The issue producing the most skewed distribution called for compulsory on—campus residency fer all college students; thus, it was chosen as the experimental issue for the investigation. Independent Variables TWO independent variables were manipulated in this investigation: 1. Perceived Choice: This variable was dichotomized as Free Choice and No Choice. §s in the Free Choice conditions were informed of their freedom to comply at specified points in the experimental in— duction. Ss in the No Choice conditions were not given such freedom. 2. Audience Attitude: This variable was dichotomized as Committed Audience and Uncommitted Audience. Ss were told that the Committed Audience was composed of individuals who held attitudes con— sistent with the counterattitudinal position advocated by the Ss. Ss in the Uncommitted Audience conditions were told that the audience was Composed of individuals who held no opinion on the issue. Inductions and Posttest Approximately two weeks after the pretest had been administered and the experimental issue had been chosen, Ss were induced to perform the counterattitudinal task. E_gave the following instructions to Se in the Free Choice conditions: The Department of Communication at Michigan State is currently conducting research aimed at finding out what it takes to strengthen or change people's attitudes. What we do is to get some indication of the attitudes people have on certain issues and then, at some later date, we present certain arguments in the form of speeches, essays, debates—~various methods——which are designed to get them to change or strengthen their new-wee .351” trim may} w:- Err-mi .-:*- wit suz:.1’-..rr.- r'iz-app 'd‘!’ an (Bub ' I l I I I. ‘. I .- 5 '.. v - 3-: -.',.-'.' .._._...__.:.'_- '. . 21 attitudes. Sometime ago, we administered questionnaires to students enrolled in various high schools in the Lansing-East lansing area asking them their attitudes toward a policy of required on—campus living for all college students. That is, we were interested in their opinions on the issue of whether or not college students should be required to live on campus during their attendance at college. Analysis of the questionnaires indicates that these high school students are generally undecided on (favorable toward) the issue. That 18, they really—have no 0 inion on (favor) the issue. We are interested 1n getting the most persuasive arguments we can supporting a policy that would require students to live on campus during their college attendance. Your arguments will be presented to these high school students who are undecided about (favorable toward) the issue and we will again measure _the1r attitudes toward the issue. In other words, you are being asked to write essays that present the most persuasive arguments favoring compulsory on—campus residency for college students. Those choosing to write the essays will be given 15 minutes. we hope that all of you will write the essays. However, if you do not care to participate, you may leave now. §s in the No Choice conditions were given the following instructions: The Department of Communication at Michigan State is currently conducting research aimed at finding out what it takes to strengthen or change people's attitudes. What we do is to get some indication of the attitudes people have on certain issues and then, at some later date, We present certain arguments in the form of speeches, essays, debates——various methods——which are designed to get them to change or strengthen their attitudes. Sometime ago we administered questionnaires to students enrolled in various high schools in the Lansing—East Lansing area asking them their attitudes toward a policy of required on—campus living for all college students. That is, we were interested in their opinions on the issue of whether or not college students should be required to live on campus during their attendance at college. Analysis of the questionnaires indicates that these high school students are generally undecided on (favorable toward) the issue. That is, they really have no 0 inion on (favor) the issue. We are interested in —gett1ng the most persuasive arguments we can supporting a policy that would require students to live on campus during their college attendance. Your arguments will be presented to these high school students who are undecided about (favorable toward) the issue and —I—‘—“——‘ we will again measure their attitudes toward the issue. 1+: - {-11.1 "x- nmadainmkdll'n .""al«'.~:--z-- .1. mini-1r: Mm ad 3113-154": -' ;i-. "5' '1 NH '. ‘ "i -:'- ". -‘fi5;£u‘ 3'5 Inh- ~ -_"‘._4':.-';-.':o If? P ?-3.'!'J..- - '- .. '-_: --:-.-"—'\.f1':_'9"'mM I. r. I.... I . ._. _...._.....__. gun—I-.. . -._._- l 22 In other words, you are to write essays that present the most persuasive arguments favoring compulsory on- campus residency for college students. You will be given 15 minutes to write your essays. Ss were then given paper on which to write the essays. After 15 minutes, the essays were collected and the post—encoding questionnaires were distributed. The instrument contained the same four semantic differential-type scales employed in the pretest to measure attitudes toward compulsory on—campus living. In addition, it contained items designed to tap Ss' perceptions of choice, audience attitude, and difficulty experienced in encoding the essays. (See Appendix B) Control Ss completed a posttest questionnaire which contained several issueS, among them the experimental issue with the same scales employed in the experimental conditions. (See Appendix C) Dependent Variables The major dependent variable in this investigation was attitude change: the difference between st attitude score on the pretest and posttest measures. In addition, a random sample of essays from all experimental groups was drawn and distributed to two varsity debaters who were asked to evaluate them and respond to items designed to measure the following dimensions: attitudinal direction, persuasive appeal, organizational quality, number of different arguments, and argument content. This essay analysis was similar to the method employed by Rosenberg (1965) and Linder et_al. (1967). The essays were randomly ordered and the raters worked independently and without knowledge of the experimental conditions. Finally, word counts were made of all essays and these were - a”? m4 as”... m'i... m m ml: an» "3:,- 1f-'_-._ .11:- ._. "-I‘TT". | 1:.- ii!“ rug ‘1'“? hi'. Tl'ufl' ‘;!':‘|‘.1 915-? m __ _ ' "'1': L .. - - -- - -- --..-"_ humility . . - -’-.1---r*--[:::::.mm$ 23 examined for possible differences. Manipulation Checks The effectiveness of the Choice—No Choice nanipulations was deternfined by having §s respond to the following question: How much choice did you have in your decision to write the essay? §s responded on a seven—interval, semantic differential-type scale that used Complete Choice, Neither, and No Choice as anchors. Perceptions of the audience attitude manipulations were de— terndned by having §s respond to the following question: Before you wrote your essay, how did you think the high school students felt toward the issue of compulsory on-campus residency? §s responded on a scale consisting of the response categories Favorable, Undecided, Opposed, and Don't Know. The anoumt of perceived difficulty in encoding the essays was measured by having §s respond to the following question: How difficult was it for you to write the essay on this issue? §s responded on a seven—interval, semantic differential—type scale that used Difficult, Neither, and Easy as anchors. ‘1.'é-52‘4Llifiin aflod‘a 4:.V-zafzm an In - . . . . l "'1 “1‘ :n ,.:‘E-;_-‘,-.:. :-‘..!'.'WI '.n'. -.I._ J lira-1:! F! 34’ grave! fl? CHAPTER III RESUIJS Pretest After random assignment of subjects to experimental and control groups, two prelindnary analyses were perforned to ensure that there were no significant differences among the groups' pretest scores. First, Bartlett's test for homogeneity of variance (Winer, 1962) was perforned. The results indicated no significant differences existed in the variances of the five groups (x2: 3.57; p > .05). In addition, a simple analysis of variance was performed on the pretest attitude scores of the experimental and control groups to ensure that the pretest scores were not significantly different. The results of this analysis indicated that the group Heans did not differ significantly (Table 1). Table 1. Means, standard deviations, and analysis of variance of pre- test attitude scores for experimental and control groups. Condition Mean Standard Deviation Choice: Unconnfitted Audience 5.86 2.33 Choice: Committed Audience 6.38 2.92 No Choice: Uncommitted Audience 6.05 3.21 No Choice: Committed Audience 6.87 3.39 Control 7.M7 3.H6 21+ *1 Table 1 (Cont'd.) Source 88 df MS F Between 30.59 H 7.65 < 1 n.s. Within 916.82 97 9.45 Total 9H7.H1 IUI Experimental and Control Differences Attitude change scores of the experimental and control groups were submitted to a simple analysis of variance. The overall anount of change and analysis of variance results are found in Table 2. These results demonstrated that the five groups differed significantly on amount of attitude change. The specific differences between each experi- mental gxoup and the control group were computed by means of Scheffe's test (McNemar, 1969). Table 3 reports the results of these comparisons. Table 2. Attitude scores, amount of change, and analysis of variance of attitude change scores of experimental and control groups. Condition Pretest Posttest Change Choice: Uncommitted Audience 5.86 10.71 H.85 Choice: Committed Audience 6.38 8.81 2.93 No Choice: Uncommitted Audience 6.05 10.77 ”.72 No Choice: Committed Audience 6.87 9.78 2.91 Control 7 H7 7 80 33 ' 3.3.5.; ________ sé """""""" a} ' _ E43— _ ' _F— — Between 2u5.58 1+ 61.6Lt 3.88* Within 1,5u1.2u fl 15.89 Total 1,787.82 01 26 Table 3. Analysis of differences of attitude Change scores between each experimental group and control. Condition* Control Means .33 C:U ”.85 l2.82** C:C 2.u3 5.95** NC:U 9.72 l2.57** NCIC 2.91 7.H5** * C:U = Choice: Uncommitted Audience C:C = Choice: Conmitted Audience NC:U = Choice: Uncommitted Audience 'NC:C = No Choice: Conndtted Audience *9 ep < .05 Thus, the results of these analyses indicated that the amount of attitude change reported by each experinental group was significantly greater than that of the control group. Manipulation Checks Chgige_ In order to determine if subjects in each of the choice conditions accurately perceived the level of the variable to which they were assigned (Choice or No Choice) and to discover whether or not these perceptions were significantly different, a two—factor analysis of variance was performed on the choice perception ratings. This analysis (Table H) yielded a significant choice nain effect. Subsequent t_tests demonstrated that the two levels of the choice variable were perceived as significantly different (Table 5); thus, the manipulation of the choice variable was successful. 27 Table 9. Means, standard deviations, and analysis of variance of choice ratings.* Condition Mean Standard Deviation Choice: Uncommitted Audience 5.67 1.56 Choice: Committed Audience 5.71 1.85 No Choice: Uncommitted Audience 1.95 1.19 No Choice: Committed Audience 2.00 1.65 Source SS df MS F Audience 1 .09 < 1 n.s Choice 15.71 1 15.71 157.22** Interaction .06 l .06 < l n.s. Error 8.30 83 .10 Total 29.16 6 “Choice perception scores ranged from 1 (No Choice) to 7 (Complete Choice). The higher the mean, the greater the perception of choice. :‘n’: p < .05 Table 5. Comparisons of mean choice ratings of experinental groups. Condition* C:U C:C NC:U NC:C Means 5.67 5.7l 1.95 2.00 C:U 5.67 ---— < 1 10.109”? 8_56:'::': C:C 5.71 —-—- 9.19** 7.81** NC:U 1.95 --—— 1.51 NC:C 2.00 ____ * C:U = Choice: Uncommitted Audience C:U = Choice: Committed Audience NC:U = No Choice: Uncommitted Audience NC:C = No Choice: Committed Audience :‘n‘: p < .05 w't-ur‘: 35'; Had! :ssi' 28 Audience Attitude The effectiveness of the audience attitude manipulation was assessed by computing the percentages of individuals within each experimental group who accurately perceived the experimental induction. These percentages are reported in Table 6. The figures in this table indicate that 90 per cent of the subjects in the Choice: Uncommitted Audience condition perceived the target audience as intended; 61 per cent of those in the Choice: Committed Audience condition accurately perceived the target audience; 69 per cent of the subjects in the No Choice: Uncommitted Audience condition perceived the audience as intended; and 26 per cent of those individuals in the No Choice: Committed Audience perceived the induction correctly. With the exception of the latter experimental condition, the manipulation can be regarded as successful. Table 6. Percentage and frequency of subjects accurately perceiving e audience condition. Italics indicate intended inductions. Audience Audience Audience Audience Condition Favorable Undecided Opposed Unknown Choice: Unconndtted Audience --—- 90% (19) 10% (2) ---— Choice: Committed Audience 61% (13) 10% ( 2) 10% (2) 19% (9) No Choice: Uncommitted Audience 9% ( 1) 69% (15) 9% (2) 18% (9) No Choice: Committed Audience 26% ( 5) 22% ( 5) 35% (8) 17% (9) Test of the Hypotheses The three major hypotheses of this study were tested by use of two analyses. Hypothesis 1, which predicted a significant choice main effect, was tested by a two—factor analysis of variance. For Hypotheses 2 and 3, which predicted differential effects of the choice and audience attitude variables, multiple comparison t_tests were employed. Data used for analyses testing the hypotheses were the mean pretest to post— encoding attitude change scores of the experimental subjects. The significance level employed for all analyses in this investigation was .05. Hypothesis 1 The first hypothesis predicted a positive relation— ship between degree of perceived choice and subsequent attitude change. Specifically, this hypothesis stated: Persons encoding counterattitudinal communications under conditions of free choice will demonstrate greater attitude change in the direction of the position taken in the come umpication than persons encoding counterattitudinal come munications under conditions of no choice. Table 7 indicates the amount of pretest though post-encoding attitude change and presents a sunnary of the analysis of variance for the experimental groups. The results of this analysis indicated that freedom of choice had no significant effect on the attitude of the in— dividuals perforndng the counterattitudinal task. However, the analysis indicated that subjects in the Uncommitted Audience conditions changed their attitudes significantly more than those in the Committed Audience conditions. Table 7. Attitude scores, amount of change, and analysis of variance of attitude change scores of experimental groups. Condition Pretest Post—Encoding Change Choice: Uncommitted Audience 5.86 10.71 9.85 Choice: Committed Audience 6.38 8.81 2.93 No Choice: Uncommitted Audience 6.05 10.77 9.72 No Choice: Committed Audience 6.87 9.78 2.91 Source SS df MS F Audience 9 50 1 9.50 9.19* Choice 0 3 1 . 0 3 1 n , 8 Interaction .09 1 .09 1 n.s Error 90.97 83 1.09 Total 9 5 . 09 ‘85 :‘e p < .05 Hypotheses 2 and 3 These hypotheses specified those cells when the greatest and least amounts of attitude change would occur. Hypothesis 2 predicted: Attitude change in the direction of the position taken in the counterattitudinal communication will be greatest when the communication is encoded for an uncommitted audience under conditions of free choice. Hypothesis 3 stated: Attitude change in the direction of the position taken in the counterattitudinal communication will be least when the communication is encoded for a committed audience under conditions of no choice. Multiple comparison t_tests were performed for all possible two treatment combinations of the experimental groups. The results of these comparisons indicated no significant differences existed among the ul' {'.JIIR'A M: =4 . MA tantrum? ' conditions (Table 8). Table 8. Comparisons of mean change scores of experinental groups. Condition* C:U C:C NC:U NC:C Means 9.85 2.92 9.73 2.91 C:U 9.85 ———— 1.72** < 1 1.29 C:C 2.92 ——-- 1.60 < 1 NC:U 9.73 ---— 1.19 NC:C 2.91 ____ * C:U = Choice: Uncommitted Audience C:C = Choice: Committed Audience NC:U = No Choice: Unconndtted Audience NC:C = No Choice: Connfitted Audience P < .05 (t = 1.68; df = 90; one-tailed) Thus, the three major hypotheses of this investigation were not supported. Supplementary Analyses In addition to the analyses of the attitude change data, analyses were performed on neasures related to the encoding behavior of the sub- jects and to some of the characteristics of their essays. Table 9 reports the means, standard deviations, and analysis of variance of subjects' perceptions of difficulty in encoding the counter- attitudinal essays. These data were subjected to a two-factor analysis of variance. Results of this analysis indicated no significant differences anong the experimental groups on the perception of difficulty measure. 32 Table 9. Means, standard deviations, and analysis of variance of difficulty ratings.* Condition Mean Standard Deviation Choice: Unconndtted Audience 3.76 2.07 Choice: Conndtted Audience 3.86 1.98 No Choice: Unconndtted Audience 9.19 2.00 No Choice: Conndtted Audience 3.22 1.76 Source 88 df MS F Audience .17 1 .17 < 1 n.s. Choice .02 l .02 < l n.s. Interaction .26 1 .26 1.99 n.s. Error 19.99 83 .18 Total 15.39 6 *Difficulty ratings ranged from 1 (Difficult) to 7 (Easy). The greater the mean, the easier the task was perceived to be. A random sample of five messages from each experimental group was drawn and these messages were given to two independent raters for judgments on the following criteria: persuasiveness, organization, number of different argunents, and attitudinal direction of the essay. The two judges rated each essay by responding to a seven—interval, semantic differential-type scale for each criterion (persuasiveness or organization) bounded by the adjectives gggd and bad. The judges agreed or were only one point discrepant on 55 per cent of the persua- siveness ratings and 65 per cent of the organization ratings. Each rater was asked to count the number of different arguments for each essay and these judgments yielded a correlation of .60. The judges agreed or were only one argument discrepant in 85 per cent of the essays. f t ., . ‘ fl 1_. PI? I I. r: '-'Fi‘!'—"T'Ma . aw!- _ _ '.i- " .‘I. v _ _-——- r 33 The judges were also asked to evaluate the essays as favorable, neutral, or opposed to the experimental issue. They agreed that 19 of the 20 essays (95%) were in favor of required on—campus living for all college students. Finally, the essays of all conditions were submitted to a word count analysis. Table 10 indicates that individuals who encoded counter— attitudinal messages in the Uncommitted Audience conditions wrote signifi- cantly longer essays than those who wrote in the Committed Audience conditions. Table 10. Means, standard deviations, and analysis of variance of essay length. Condition Mean Standard Deviation Choice: Uncommitted Audience 172.33 56.60 Choice: Committed Audience 156.76 97.83 No Choice: Uncommitted Audience 179.68 52.96 No Choice: Committed Audience 199.57 56.93 Source SS df MS F Audience 521.85 1 521.85 3.96* Choice 29.25 1 29.25 < 1 n.s. Interaction 52.89 1 52.89 < l n.s Error 10,932.76 83 131.72 Total 11,561.75 8 CHAPTER IV DISCUSSION In light of the strong theoretic support for the prediction, failure to confirm Hypothesis 1 is disappointing. Prior findings concerning the effects of varied choice conditions indicated that perceived freedom of choice in perfornance of a counterattitudinal task is a significant deternfinant of attitude change. The non— significant results are even nore perplexing in light of the highly successful inductions of the choice variable. Nevertheless, there are several plausible explanations for the failure to confirm Hypothesis 1. 5 The c1assroom.experinental setting is one factor which could account for individuals correctly responding to the manipulation check items but failing to denonstrate the predicted amounts of attitude change. Comments by the Choice experinental subjects during debriefing indicated that even though they had been given an opportunity to leave, they felt that the instructor would expect them to remain. Moreover, Host of the subjects had participated in prior classroom experimental sessions and conforndty with experimenter requests was perceived to be normative behavior. Debriefing infornation given by the Choice subjects is at odds with the position held by Holnes and Strickland (1970) concerning the desired situation for an optimum test of the choice variable. They argue that confronting subjects in a classroom situation allows them 39 more freedom to leave the experdnent than if they are involved in a one-on—one confrontation with the experimenter. In addition, they note that a group situation eliminates the problem of a sequence of decisions; that is, the subject does not have to first decide whether to attend an experimental session and then later decide whether to stay and complete the task. In the present study, the subjects indicated that they perceived the choice to be a very limited one, even though their choice ratings were high. In an earlier study (Bodaken, 1969), subjects were invited to a special night session to participate in a research project. Indi— viduals in the Choice conditions indicated that they remained to perform the task because they had taken the trouble to attend and that as long as they were there they could afford the tine necessary to complete the task. Thus, it is clear that regardless of the method used to confront subjects, there is going to be some tempering of the choice perception. It nay well be that the optimum test of the choice factor resides in a methodology which calls for confronting subjects in their residence halls or some similar extra-classroom environnent. For the present study, it seems plausible that individuals perceived that they were given decision freedom. At the same time, they seemed to be aWare of the situational requirements that they generally do not leave the classroom without the instructor's pernission. Another possible explanation for failure to confirm Hypothesis 1 involves the perceived importance of the subject's decision to perform the task. According to Festinger (1957), the variable of importance is ' v 99:41" nth .‘fi‘n ' : r - - ~=- 6 :31? m: " . . mm um -‘ -' :thrmfx‘. :IrI ' -.;--4 .1. ' I.- .. . .. .- .I.;. A“: .'- " 311$}: 53 m: .1 I. "T - 'r'i r——————————————————____________________________________________________________________________________::!Ill 36 a crucial determinant of the magnitude of dissonance. He writes: The importance of the decision will affect the magnitude of the dissonance that exists after the decision has been made. Other things being equal, the more important the decision, the stronger will be the dissonance. (p. 37) The importance of the choice as a relevant variable was determined in pretesting of the manipulation prior to the experimental induction. Students in several undergraduate communication courses were asked to react to a request to performlthe counterattitudinal task. Generally, these students indicated that the choice would be relatively easy to make: they would refuse to comply. On the basis of further reactions, the choice inductions were rewritten to include a sponsor (the Department of Communication) and some clarification of the bogus research project. It was reasoned that the addition of these elements would make it more difficult for experimental subjects to refuse. Furthermore, it was felt that the experimenter's explicit statements reminding the subjects of their freedom not to comply would provide a perception of real decision freedom. Finally, the absence of the regular instructor during the entire experimental session was a factor that should have presented the subjects with an illusion of complete freedom. Thus, the choice to perform or not perform the task should have been perceived as an important one and, additionally, a decision free of external pressure. Nevertheless, the subjects' comments indicated that the decision to write or not write the essays was perceived as relatively unimportant. After all, students perform many routine encoding tasks aim-1mm- mam- 1m Lliu um: be!) it!!! 5‘4 313’ dual-ash not) «9'2”. ' «untur- and? "w! r'.-.-.~:_.mi “on? J?! Le..;- *7. -:' mum!" M‘O .M i" sumr: :; ' .h" -'-- “v an 11"»: *u': ..th 37 in the course of completing class requirements and, as noted earlier, conforming to an experimenter's request was perceived to be just another classroom exercise. In sum, while the task may have assumed importance during the actual encoding process, the choice per §§_was perceived as relatively unimportant. Another possible reason for failure to confirm Hypothesis l—— one related directly to the perceived importance of the task——is concerned with the relative attractiveness of the unchosen alternative. Festinger (1957) specifies the importance of such relative attractive- ness when he says: . dissonance exists because, following the decision, the person continues to have in his cognition elements that, if considered alone, would lead to an action other than the one he has taken or is engaged in. . . . Con— sequently, the greater the relative attractiveness of the undhosen alternatives to the chosen alternative, the greater will be the proportion of relevant elements that are dissonant with the cognition corresponding to the action. (pp. 37-38) Subjects in this study were not presented with mutually positive or negative aspects of the alternatives. Generally, subjects seemed to be more concerned with the potential effects on the target audience than with the relative merits of whether or not they made the right decision. Brehm and Cohen (1962) suggest the importance of presenting attractive alternatives in a choice paradigm: The subject in these experiments presumably wants to make a choice because of the benefits or potential benefits accruing to the choice, and little or no external pressure is needed to roduce the choice. Dissonance in these studies is us 1y conceived of as a function of the relative number of cognitions favoring the unchosen alternative. (p. 21) (Under— 1ining Added) h -.: II:- I. 7-. 1 him" ’* hut-um I'll-“:1? :19; -.--?':.Tu we ill! g \— 4. 5: ; ":'.".';“:9 m. 1:105 sue 38 They later note that: . . the more equal the distribution of relevant attributes (the higher the choice), the greater the dissonance and consequent attitude change in line with one' s discrepant commitment. (p. 91) Thus, assuming that the subjects perceived some pressure to perform_the task, and fUrther assuming that the manipulation of the Choice conditions did not present the alternatives as mutually at— tractive, it is reasonable to conclude that decision freedom was not a mediating variable in this study. Perhaps the most relevant factor in the failure to confirm Hypothesis 1 was the subjects' concern for the potential effects of the counterattitudinal task. Invariably, subjects in the Uncommitted Audience conditions expressed concern for the possible effects that their essays might have had on the high school students' attitudes. This concern is supported by the significant results indicating that subjects in the Uncommitted Audience conditions demonstrated greater attitude change in the direction of the counterattitudinal position than those in the Committed Audience conditions. While this effect may have militated against confirmation of Hypothesis 1, it certainly supports earlier theoretic fOrmulations and research. As noted earlier, Collins (1969) has suggested that an individual's concern about his responsibility for potentially harmful consequences growing out of his act may be dissonance producing. By contrast, as might be expected, subjects in the Committed Audience conditions were more concerned with the fact that their task was seemingly meaningless; they perceived theme selves to be trying to persuade others (high school students) to adopt a position already held. Taken as a whole, these effects of the w JMu ”WW: m {If _:‘ . TIP-‘HIT'Imm M ..v. ' J 'h 39 audience variable may have over—ridden any effects resulting from the choice variable. The significant audience efféct not only supports Collins' theorizing, it is consistent with research by Nel, Helmreich, and Aronson (1969). TWO major implications of this successfu1 replication are apparent. First, it appears that manipulating audience attitude is a useful means of tapping the aversive consequences dimension. Unlike Nel §t_§l., the present study did not manipulate a "neutral" audience attitude; however, the findings suggest that such a manipulation is unnecessary. Moreover, the procedures used by Nel §t_al, involved presenting the subjects with several major arguments on the experimental issue. By contrast, subjects in this study were not given any arguments on the issue; all arguments used in the essays were the subjects' own ideas. Thus, subjects were not allowed the option of cognitively shifting the responsibility of their task to the experimenter. In addition to the methodological implications, the significant audience effect provides fUrther support for theoretic expectations about the aversive consequences of counterattitudinal advocacy. The finding suggests that an individual's overriding concern in the per— formance of counterattitudinal tasks is the effect, or potential effect, that such messages have on others. Aronson (1968) and Collins (1969) have both indicated that individuals do not perform counterattitudinal tasks in a cognitive vacuum, that subjects are concerned with the per— ception that they have ”conned" someone into believing some thing which is not true. Furthermore, the absence of an option to explain the be- havior to the target audience is even more dissonance—producing. 6‘ 5r?"- .. 9:731 infirm It" . T ..u- u ‘. 'u' _'.'i "nib? rm‘Y' 90 Subjects in the present study were told that they would not_be able to explain their private opinions to the high school students, thus pre— cluding a take-back mode of dissonance-reduction. The significant audience effect also suggests the importance of self—concept and the part it plays in the arousal of dissonance. As noted earlier, Bramel (1968) posits that the cognition one has that he is immoral or has performed an immoral act can be instrumental in dissonance—arousal. Subjects in this experiment repeatedly commented on the unethical nature of performing the task because of the subsequent harmdul effects such behavior could have in persuading high school students who held no opinion on the issue. However, since Bramel‘s formulation places emphasis on the voluntary nature of the task and since the results of this study fail to demonstrate a significant choice effect, the self-concept interpretation is equivocal. While the results of this study failed to confirm Hypotheses 2 and 3, the pattern of change scores is encouraging. Taken together, these hypotheses predicted that the greatest and least amounts of attitude change would take place in the Choice: Uncommitted and the No Choice: Committed conditions, respectively. While no predictions were made for the Choice: Committed and No Choice: Uncommitted conditions, it was assumed that subjects in those conditions would experience moderate dissonance and would demonstrate amounts of change ranging be— tween the Choice: Uncommitted and No Choice: Committed conditions. The findings do indicate that persons in the Choice: Uncommitted conditions demonstrated a relatively greater amount of attitude change than subjects in the other conditions. However, the condition yielding the least amount of change was not the No Choice: Committed condition; .1 .i '7'": '.t. "1,553.01”. -. :. " I'; ' ’3'“- 'J" I ‘3'?- d,- .l: - F“ 'l I . - ‘ - . '. .. - . fiw -’m H199. " all . -'_ Ir'g-gnr‘ ,.mj_:tg!'av--Hr_.-a‘l=fi7hd5 ,5 ‘ I .- '- I 7'3 . .‘cnslmms 91 rather, it was the Choice: Committed condition. As indicated in Chapter III, the only significant difference among the experimental groups was between the Choice: Uncommitted and the Choice: Committed conditions. Theoretically, the No Choice: Committed condition should yield minimal attitude change. For not only are subjects being forced to encode counterattitudinal messages, they are doing so for an audience that already holds the counterattitudinal position. The No Choice: Committed condition was not expected to be significantly different than the control group. In fact, all conditions differed significantly fawn the control group, with the Choice: Committed subjects demonstrating less attitude change, although not significantly different, than all other experimental groups. It is interesting to note that this pattern of changes replicates an earlier study (Bodaken, 1969). However, in the present study, the difference between the Choice: Committed and No Choice: Committed conditions was less than one-half a scale unit, thus preventing clear interpretation of the findings. A possible contributing factor in the failure to confirm Hy- potheses 2 and 3 was the partial failure of the audience attitude manipu- lation. Only 26 per cent of the subjects in the No Choice: Committed condition perceived the audience attitude as intended, while by contrast, 90 per cent of the subjects in the Choice: Uncommitted condition per— ceived the audience attitude consistent with the induction. Also, the induction was Successful for over 60 per cent in each of the other two experimental conditions. These figures indicate that while the induction was generally successful, it did not work as intended in the No Choice: hJfi‘énm'f" :s'l‘mfl an: Fills ('.:- -.--:=.am 2M- 92 Committed condition. The attitude change demonstrated by these in— dividuals cannot be attributed to the audience induction. Considering this fact in conjunction with the failure to find a significant choice effect, it appears that the attitude change may be attributed to ex— traneous factors. For example, unlike the other conditions, subgroups of subjects in this condition were given instructions by different instrmmtors. More investigation must be undertaken to explain the change demonstrated by No Choice: Committed subjects. In presenting the rationale for Hypothesis 2, it was suggested that for subjects in the Choice: Uncommitted Audience condition, the voluntary nature of the task and its potential harmful effects would yield an amount of attitude change significantly greater than that observed in other conditions. While a significant audience effect was observed in this study, the lack of confirmation for the choice pre— diction prevented a powerful test of Hypothesis 2. Hypothesis 3 was based on the assumption that lack of choice in performing a counterattitudinal task plus the seeming absence of harmful effects would combine to yield an insignificant amount of change in the No Choice: Committed Audience condition. Failure to find a significant choice effect, the apparent weakness of the experimental induction, and possible experimenter errors prevented confirmation of this hypothesis. It is, however, interesting to note that in conditions where the most attitude change was reported, the subjects wrote significantly longer essays. That is, subjects encoding counterattitudinal messages directed toward an uncommitted audience wrote significantly longer essays than those subjects who wrote counterattitudinal messages 1"“ . '.- If" 1‘"; , . f ‘.: .{....- -- - .-" . ""Ji:\r=.‘,- Sinai-151393" '* Mg}. II"; 'Hultfi ih'l.’ W fl' . .~ -=-5-r.-ri:-u. 21!: 3'51! . ' -- firm-I an“! ——_.! 93 directed toward a committed audience. This finding contradicts Rosenberg's (1965) conclusion that verbal productivity does not cor— relate with attitude change. Also worth noting is Freedman's (1963) failure to find any differences in essay length and his conclusion that subjects did not exert any more effort in encoding essays across con- ditions. Confidence in the finding of the present study is tempered by the failure to find that subjects who showed greater attitude change wrote a significantly greater number of different arguments. Moreover, with the lack of a suitable control for individual differences in writing speed and verbal skills, the importance of this finding should not be overemphasized. Implications For Future Research While the major hypotheses of this study were not confirmed, several implications for future research in the area of self—persuasion can be suggested. First, present operational procedures for manipula- ting the choice variable need to be re—examined. Researchers investi- gating this variable have strayed considerably from Festinger's (1957) fOrmulation of pre— and post-decisional conditions. Future researchers should concentrate on presenting the subject with attractive alterna— tives, rather than a sterile forced—choice paradigm. Not only should the subject feel that there are obvious benefits or harms in either alternative, regardless of his choice, he should also perceive the choice as important and one that is difficult to make. The potential aversive consequences dimension of counter— attitudinal advocacy is a worthwhile area for future research. The methodology employed in this study seems to be an adequate means of (tam? swash-- : :2. 912'}!!! im his"! W47; _ '51:.» -. -. -1'_I'_'.f"- vs: '-- . stnw'nfifl I . p_ .. 1' 1 '. . --~. n. ‘(M Tara's 3m ' . if .3 _-. '.: Hi: ' ’13.!) nu tapping this variable, but as indicated by the failure of the induction in the No Choice: Committed Audience condition, more can be done to successfully manipulate perceptions of audience attitudes. One option might be to present the experimental subjects with bogus results of a questionnaire presumably administered to a target audience. Another approach would be to make the encoding session an oral one in which subjects give speeches to an uncommitted audience composed of confed— erates. At predetermined intervals the confederates would signal agreement with the speaker, thus presenting him with immediate feedback. Measures of source attitude could then be taken at various times during the encoding process. Berger (1970) is currently conducting research employing this methodology. Finally, efforts should be concentrated on analyzing the content of the counterattitudinal messages. While subjects in the present study supposedly did write counterattitudinal essays, nothing is known of the intensity or complexity of the language used. Burgoon (1970) has de— vised a methodology that, if successful, will provide information relevant to the correlation between language intensity and attitude change follow— ing counterattitudinal advocacy. Content analytic procedures also offer potentially useful measures for the comparison of counterattitudinal messages written by subjects in different experimental conditions. The area of counterattitudinal advocacy offers the communication researcher the opportunity to perform insightful research into variables influencing attitude fermation and change. This study has attempted to specify the relationship between two variables considered to be relevant to the perSuasive process. Hopefully, the results will contribute to 3 It'- I i ‘ :I l I. ‘7'. --‘. I -' .«._ n- ”‘im‘fl WNW r 9'", . :17“ f" «‘5' -.‘.~:r":r=. await-m. 5:: -. swim. . w. tan-1m as m. _ 1" 1'5 "Mime-mg :‘; "3...: ‘.., '_{ 232W?” I I I ‘ J r E I .. 1 . . .,:_ 'i g, E . I I I . I i . i 95 existing knowledge about self—persuasion phenomena and will lead to further investigation of the process of counterattitudinal advocacy. Summary This study examined the effects of choice and audience commitment on attitude change following encoding of counterattitudinal messages. Hypotheses were based on prior self—persuasion research which has established that decision freedom is a determinant of cognitive dis— sonance and subsequent attitude change, as well as research relevant to the selftconcept and dissonance aroused when an act is perceived to have potentially aversive consequences. It was predicted that persons encoding counterattitudinal communications under choice conditions would demonstrate greater attitude change than persons encoding such messages under conditions of no choice. Furthermore, it was predicted that the amount of attitude change demonstrated by individuals freely Choosing to encode counterattitudinal essays for an audience holding no opinion on the issue would be significantly greater than the change demonstrated in all other treatment conditions. Finally, it was hy— pothesized that the amount of attitude change demonstrated by subjects who were forced to encode counterattitudinal essays directed toward a committed audience would be significantly less than the change in all other treatment conditions. The results of this study failed to confirm the major hypotheses; however, a significant audience effect was observed, indicating that persons who encoded counterattitudinal communications directed toward an uncommitted audience demonstrated significantly greater attitude change than individuals who encoded such messages for a committed audience. I. . '! ":L'IE' .__ . en- _ > r . .- l "-. _ .. .r.= . i I ¥ "'fiiaifi REFERENCES Abelson, R. P. Psychological implication. In Abelson, R. P., licGuire, W. J., Newcomb, T. R., Aronson, E., and Tannenbaum, P. H. (Eds.), Theories 9f_Cog§itive Consistency: A Sourcebook. Chicago: Rand McNally and Co., 1968, 112—139. " "" " Aronson, E. Dissonance theory: Progress and problems. In Abelson, R. P., McGuire, W. J., Newcomb, T. R., Aronson, E., and Tannenbaum, P. H. (Eds. ), Theories of Co 'tive Consistency: A Sourcebook. Chicago: Rand MCN aIIy and C_., , --27. _ Ashmore, R. D. and Collins, B. E. Studies in forced compliance: Attitude change and commitment to maintain publicly a counter— attitudinal position. Psychological Reports, 1968, 22, 1229-1239. Berger, C. Use of feedback in self-persuasion paradigms. Unpublished research, Department of Psychology, Illinois State University, 1970. Bodaken, E. M. Choice and perceived audience attitude as determinants of cognitive dissonance and subsequent attitude change following counterattitudinal advocacy: A pilot study. Unpublished manu- script, Department of Communication, Michigan State University, 1969. Bramel, D. Dissonance, expectation and the self. In Abelson, R. P., McGuire, W. J., Newcomb, T. R. ,Aronson, E., and Tannenbaum, P. H. (Eds. ), Theories of Cognitive Consistency: A Sourcebook. Chicago: Rand McNalIy ®., 1968, 35 —3 Brehm, J. W. and Cohen, A. R. Choice and chance relative deprivation as determinants of cognitive dissonance. Journal of Abnormal and Social Psychology, 1959, 58, 383-387. ~— Explorations ig_Cognitive Dissonance. New York: Wiley, 1962. Brock, T. C. Cognitive restructuring and attitude change. Journal 9: Abnormal and Social Psychology, 1962, 69, 269—271. Burgoon, H. M. Prior attitude and language intensity as predictors of message style and attitude change following counterattitudinal communication behavior. Unpublished doctoral dissertation, Department of Communication, Midhigan State University, 1970. 96 97 Carlsmith, J. M., Collins, B. E., and Helmreich, R. L. Studies inforced compliance: I. The effect of pressure for compliance on attitude change produced by face-to—face role playing and anonymous essay writing. Journal g Personality an_d Social Psyc‘holog, 1966, 5_2, 256-261. Carlsmith, J. M. Varieties of counterattitudinal behavior. In Abelson, R. P., McGuire, W. J. ,,Newcomb T. COR., Aronson, E., and Tannenbaum, P. H. (Eds. ), Theories ofCo tive Consistency: A Sourcebook. Chicago: Rand McNally and —Co., Cohen, A. R. Attitudinal consequences of induced discrepancies between cognitions and behavior. Public Qpinion Quarterly, 1960, 29, 297-318. . An experiment on small rewards for discrepant compliance and attitude change. In Brehm, J. W. and Cohen, A. R. Explorations 2 Cognitive Dissonance. New York: Wiley, 1962, 73-78. Cohen, A. R. , Breth, J. W., and Fleming, W. H. Attitude change and justification for compliance. Journal g Abnormal and Social Psycholog, 1958, g, 276—278. " —_ '——“‘——~ Collins, B. E. The mediation of change due to counterattitudinal behavior. In Abelson, R. P. ,McGuire, W. J. ,,Newcomb T. R. Aronson,E . , and Tannenbaum, P. H. (Eds. ), Theories of Cognitive Consisten : A Sourcebook. Chicago: Rand McNally and—CE,— 1968 819 826. The effect of monetary inducements on the amount of attitude change produced by forced compliance. In Elms, A. C. (Ed. ), Role Playing, Reward, an_d Attitude Change. Cincinnati: Van Nostrand, 1969, 2 Festinger L. A Theory of 00%itive Dissonance. Stanford Calif.: St dnford Univ. Press, , Festinger, L. and Carlsmith, J. M. Cognitive consequences of forced compliance. J___ourna1 g Abnormal aid Social Psycholog, 1959, 58, 203—210.—— ‘— Helmreich, R. L. 8 Collins, B. E. Studies in forced compliance: Commit- ment and magnitude of inducement to comply as determinants of opinion change. Journal 9_f_ Personality and Social Psycholoa, 1968, E, 75—81. Holmes, J. G. and Strickland, L. H. Choice freedom and confirmation of incentive expectancy as determinants of attitude change. Journal 9: Personality an_d Social Psychology, 1970. Janis, I. L. and Gilmore, J. B. The influence of incentive conditions on the success of role playing in modifying attitudes. Journal g Personality an_d Social Psycholog, 1965, _1_, 17—27. .c .... .. “I I W. v 0‘ - n H . _ a a. .J ...... . .... . .....H .... .. . .. _ ......m. . l {0 - ma... _. t. _ r _ is. a .9 "e":- I O .1.- 98 Linder, D. E., Cooper, J., and Jones, E. E. Decision freedom as a determinant of the role of incentive magnitude in attitude change. Journal o_f_ Personality and Social Psychology, 1967. 6, 295-259. _ "— _"‘—’ McNemar, G. Psychological Statistics. New York: John Wiley and Sons, Inc. , 1969. Miller, G. R. and Bodaken, E. M. Justification and self—persuasion follming commitment to encode, and actual encoding of counter- attitudinal communication: A replication and extension. Paper presented at the Speech Association of America Convention, New York, 1969. Nel, E. , Helmreich, R. , and Aronson, E. Opinion change in the advocate as a function of the persuasibility of his audience: A clarification of the meaning of dissonance. Journal of Personality an_d Social Psychology, 1969, 12, Tm.— Osgood, M.E E., Suci, G. J ., and Tannenbaum, P. H. The Measurement o_f eanin.g Urbana, Illinois: Univ. of Illinoijress, Rosenberg, M. J. When dissonance fails: On eliminating evaluation apprehension from attitude measurement. Journal g Personality and Social Psychology, 1965, p, 28-92. Steiner, I. D. and Darroch, R. K. Relationship between the quality of counterattitudinal performance and attitude change. Journal of Personality and Social Psychology, 1969, 1_1, 312- 320. —__———___._._—. Winer, B. J. Statistical Principles in Experimental Design. New York: McGraw Hill Book Co. 71962. Zajonc, R. B. Cognitive structure and cognitive tuning. Unpublished doctoral dissertation, University of Michigan, 1959. The process of cognitive tuning in commmication. Journal g Abnormal and Social Psycholog, 1960, 6_l, 159—167. APPENDICES APPENDIX A PRETEST ATTITUDE QUESTIONNAIRE 51 COMMICATION 101 OPINION PROFILE Name Student Number 101 Section Number Time Instructor As you know, one of the activities you will engage in this term is an 5 analysis of a significant social problem. The purpose of this survey g is to solicit the opinions of Communication 101 students on a variety of current issues—-campus and national--which may serve as the foci of the problem analysis. On each of the following pages you will find a number of issues each followed by a series of descriptive scales. For example, here is an item like those you will see: The United States should withdraw fnmm the United Nations. Your job is to place a check—mark (X) above the line that best indicates your judgment about the proposi— tion. For example, if you feel that U.S. withdrawal should be very bad, you would check as follows: _—.—._—._—_—__—_—_— If you feel that such a move (withdrawal) would be quite bad, you should check as follows: If you feel neutral or indifferent about the proposi- tion, or if you feel that the scale is irrelevant to the proposition, you would check as follows: Remember: Never put more than one—check—mark on a single scale and be sure that each check is in the middle of the line, not on the boundries. ALL SCALES SHOULD BE CHECKED —— DO NOT OMIT ANY. WORK RAPIDLY -- RECORD FIRST IMPRESSIONS -- DO NOT CHANGE MARKS Worthless Pleasant Unfair 52 and possession of marijunana should be legalized. Bad Valuable Unpleasant Fair Bad Fair Valuable Unpleasant All universities and college should establish Black Studies programs. Unpleasant Bad Fair Valuable Pleasant All students should be required to live on campus during their college attendance. Pleasant Good Unfair Worthless Worthless Bad Fair Unpleasant ..--~---— '... .. n _ u _ _ 53 Psychological testing should be used to determine a student's academic major. Good : : : : : : Bad worthless : : : : : : Valuable Unpleasant : : : : : : Pleasant Fair : : : : : : Unfair APPENDIX B POSTTEST ATTITUDE QUESTIONNAIRE Name Student Number Now that you have written your essay, we would like to know how you feel toward the issue of requiring all students to live on campus during their college attendance. ALL STUDENTS SHOULD BE REQUIRED TO LIVE ON CAMPUS DURING THEIR COLLEGE ATTENDANCE __Very Pleasant Quite Pleasant __Slightly Pleasant __Neutral __Slightly Unpleasant Quite Unpleasant __Very Unpleasant ALL STUDENTS SHOULD BE REQUIRED TO LIVE ON CAMPUS DURING THEIR COLLEGE ATTENDANCE __Very Good __Quite Good __Slightly Good __Neutral __Slightly Bad _Quite Bad Very Bad |I| ll'ilil‘l‘ 56 ALL STUDENTS SHOULD BE REQUIRED TO LIVE ON CAMPUS DURIN ING THEIR (DILEGE ATTENDANCE Very Valuable Quite Valuable Slightly Valuable Neutral Slightly Worthless Quite worthless ALL STUDENTS SHOULD BE REQUIRED TO LIVE ON CAMPUS DURING THEIR COLLEGE ATTENDANCE Very Vkmthless Very Unfair Quite Unfair Slightly Unfair Neutral Slightly Fair Quite Fair Very Fair How much choice did you have in your decision to write the essay? Neither No Choice 57 Before you wrote your essay, how did you think the high school students felt toward the issue of compulsory on—campus residency? They were favorable toward compulsory on—campus residency. They were undecided toward compulsory on—campus residency. They were opposed to compulsory on-campus residency. I don't know what their attitude was toward compulsory on—campus residency. How difficult was it for you to write the essay on this issue? Difficult Neither . ' ' Easy APPENDIX C CONTROL POSTTEST ATTITUDE QUESTIONNAIRE ‘ g 59 Name Student Number 101 SectiOn Recently, the activities of university students across the country have provided a multitude of news stories for the mass media. Yet with all the publicity surrounding student revolt and unrest, little is really known about student attitudes toward a number of issues related to student life in the university community. On the following pages are several issues that are of concern to university students. Each issue is followed by a set of descriptive scales. We would like to have you place a check (/) on the line that you think best indicates your opinion toward the issue. Each scale should be checked—-do not omit any. 60 The voting age should be lowered to 18 years. Very Bad 'Quite Bad'SIi—gffily. NeuFaI .SlifitIy Date—Gad Very GEO—5d Bad Good QUite ' Slightly ' Very Pleasant Pleasant Pleasant Neutral Slightly Pleasant 'EEQuite I Very Pleasant Pleasant Very Quite Slightly Neutral Slightly Quite Very Fair Fair Fair Unfair Unfair Unfair Very Quite Slightly Neutral Slightly Quite Very Worthless Worthless Worthless Valuable Valuable Valuable The sale of cigarettes should be harmed on campus Very Quite Slightly Neutral Slightly Quite Very Unpleasant Unpleasant Unpleasant Pleasant Pleasant Pleasant Very Bad QTite Bad Slightly Bad Neutral Slightly Good 'Quite Good Very Good Quite Slightly Valuable Valuable Valuable Neutral Slightly Worthless Quite Very Worthless Worthless Very Quite Slightly Unfair Unfair Unfair Neutral Slightly Fair Quite Very Fair Fair __—————__- 61 Draft deferments should be eliminated for college students. Very Quite Slightly “manual Slightly Quite Very Fair Fair Fair Unfair Unfair Unfair Very Quite Slightly Neutral Slightly Quite Very Worthless worthless Worthless Valuable Valuable Valuable Very ' Quite . SIigHtIy. Neutral ' Sligfitfiy' Qfiite .Very Bad Good Good Good Bad Bad Very Quite Slightly Neutral Slightly Quite Very Pleasant Pleasant Pleasant Unpleasant Unpleasant Unpleasant All students should be required to live on campus during their college attendance. ‘Very IIQuite Slightly —Neutral Slightly Quite Very Valuable Valuable Valuable Worthless Worthless Worthless Very Quite Slightly Neutral Slightly Quite Very Bad Good Good Good Bad Bad Very Quite Slightly Neutral Slightly. Quite .Very Fair Unfair Unfair Unfair Fair Fair very Quite Slightly Neutral Slightly Quite ’Very Pleasant Pleasant Pleasant Unpleasant Unpleasant Unpleasant 62 Military activity in South Viet Nam should be significantly increased. Very Quite Slightly Neutral Slightly Quite Very worthless Worthless worthless Valuable Valuable Valuable : : : : ' Very Quite Slightly Neutral Slightly Quite Very Fair fair Fm Unfair Unfair Unfair Very Quite Slightly Neutral Slightly Quite Very Unpleasant Unpleasant Unpleasant Pleasant Pleasant Pleasant Very Bad Quite Bad Slightly Neutral Slightly Quite Very Bad Good Good Good Draft Card burners should automatically be reclassified l—A by their draft boards. Very QUite Slightly Neutral Slightly Quite Very Fair Unfair Unfair Unfair Fair Fair Vew QHite Slightiy Neum a1 Slightly . Quite I very Valuable Valuable Valuable Worthless worthless Worthless Very Quite Slightly Neutral Slightly Quite Very Bad Good Good Good Bad Bad W—‘ery .Tte Tightly N'eu" Hal— Slightly Quite Very Pleasant Pleasant Pleasant Unpleasant Unpleasant Unpleasant 63 The letter grading system should be abolished Very Bad Quite Bad Slightly Neutral Slightly Quite Very Good Bad Good Good Quite Slightly Neutral Slightly Quite Very Very Pleasant Pleasant Pleasant Unpleasant Unpleasant Unpleasant Very Quite Slightly _Neutral “Slightly Quite ,' Very Fair Fair Fair Unfair Unfair Unfair Very . Quite Slightly Neutral Slightly QUl‘te ' Very Worthless Worthless Worthless Valuable Valuable Valuable The United States should withdraw from the United Nations. Very QUite Slightly Neutral Slightly Quite Very Unpleasant Unpleasant Unpleasant Pleasant Pleasant Pleasant Very Bad ‘ finite Bad S IlfitIy NeutraISlightly . the .Very Good Bad Good Good Very Quite Slightly Neutral Slightly Quite Wery Valuable Valuable Valuable Worthless Worthless Worthless Very (XE-{e :SI fifI: Neutr E: Slightly . Quite .V'e_ry Fair Unfair Unfair Unfair Fair Fair 6L} Presidential nominations should be based on a direct vote of the people. Very Quite Slightly Neutral Slightly QUite Very Fair Fair Fair Unfiar Unfair Unfair Very Qu1te Slightly Neutral Slightly ite Very Worthless Worthless Worthless Valuable Valuable Valuable Very . QUite Slightly NeufiI Slightly Quite VEry Bad Good Good Good Bad Bad Very . Quite Slightly. Neutral Slightly . Quite - Very Pleasant Pleasant Pleasant Unpleasant Unpleasant Unpleasant Sumner school students should pay higher tuition than those enrolled during the regular academic year. Very Quite Slightly Neutral Slightly Quite Very Valuable Valuable Valuable Worthless Worthless Worthless Very Quite Slightly Neutral Slightly Quite Bad Very Bad Good Good Good Bad Very QUite Slightly Neutral Slightly Quite Very Fair Unfair Unfair Unfair Fair Fair Very Quite SlightIy Neutral Slightly Quite Very Pleasant Pleasant Pleasant Unpleasant Unpleasant Unpleasant 65 All firearms should be licensed and registered Very Quite Slightly NeutraI Slightly jute Very Worthless Worthless Worthless Valuable Valuable Valuable Very Quite Slightly ’Né'Tutr .Slightly. Quite. Very Fair Fair Fair Unfair Unfair Unfair Very I Quite Slightly. Neutral—SHEET. Quite . Very Unpleasant Unpleasant Unpleasant Pleasant Pleasant Pleasant Very Ed I ifiiite . SIifitIy. NeutraI .Slightly . Quite Very Good Bad Bad Good Good ‘IlllllllllllllllES