. . .OO. . . .. . .o......-.o..'.....: . . . .. . A. ... ...v. . . . . .9...........A.o... II. . ......n...v.‘...........o....o...¢... 1.31;... .-. .. .. .... .... .. A.... . .... ... . . .. .. . . r . I 0140 u. I A .. ... .\..o..o.... .. O.....I..a..1.......Ao...... ..I...... _. . ......L..._ . ... _... . . . ... . IA... . .... . . L . . .. _... .........Awhs.._.......\. o. naked ...)” O ...... .. .....o...u.v¢ outca....o.5r~O.-.o"o ... ......09.\. u. A .1.... .o . . . .IV...I.1.. . . o c. .. . .. ... ...... . t ., . _.. .. . . 4 . . r r .. .... o.\.l.L. .’ 4 4 ~10! ...-I.Qa.o to. . .w. ... . ......VA 9. .... I . 9.. .. ... . c. v. . I I . I. v . .. I o. A a . .p I... I a..... ’o....0’.'l- .Uoo'V ‘~ . .. .....r. I... .o .. v. ... . . o , .. . .... . ... . ... O. 0...... .4...¢o‘§. — 9.0' 'D.“( D . p...v..... ...-...... . . ... A .. .. . I. .. A . . . . A . . .. . o f.o..o ......o..v¢.afi»l4.y1 ’ f.‘/ l' I ..... . 0.. . o . . ..a.. _. :1. .. . A .... .. o .. .. . ..._. I... .. ..(...... .o.;¢. 01'. uooV, i 0 IO I f -. .6... . . . . .... . . . .. _ . _. . .A a . .. . . . . .......v.tl v.0 oo.’.;..pl,.'yda. . 0/. a 0" U ..... .. I. A. . . .... . . . . AA. . .. A... . . . . A. .A . .1.....p)o... .4... . .rlou .rflfu. a ...Q..... ... .. . . ... . . . ' ... . . . ... . . . .. .. .. . . . .. . . o .. ... . 09.....0-5'0...o.¢o. (”0.04.4 u .A. ‘ . ... . ..I.. .Q . o . ... .... .0 .—.. . . o o I . ... ...-.. . .... .q A ... Z . . A . .. . . ... .1 I... ... . . a . . A . . . g . ... ... I A .30... ..d DINO." I7 0 “0*.JD .... .. . I. .... .... . .. . o .. .... .0 .... A . _ .. , . _ . .. . . . ...A. ..q .\.a ..... . no)..|b ~—'.L ’0 .A s a . . . .. u . . . _. ...... I. I _.. . A9. . . . .... . A. . .. . . 9 .. . . .. . . . . . .....A . . . . . .. to. .9 _ . A.. . . _. ... .. _.. ... . .. .(:..-o.wn.a. (wofltwfovtnfofyo/ .. ... . . .... .. .. : . .....: U . . . . .. ...: ...... . «.... 1.1.15.4 V ...... . .. . ..- .. .. . . . a . . ....O . . . .. . .. A . .. ...-. .....1 nv¢o'. '..J . 0.. .. . . A . ...... . . .. . . . . ... . n A .. Ac 4f?a.o...0.0.oro o _.... . . .. .. . .. . A . . .. .... .. Il...;. I u . . . Q... ... . .. I s. . .. ... v . . . . . .A . _ . . .v .. . A. _ .. . . . . . n . . . A . .. . . . . . . r o . . 4. . . ...ocow._009.fld.'¢..4. . s. . ... . .- A. . . o A. . . A AA I I . I ...O. .. . ..A. A . A. . . . . . . A l...o. . l...'».0¢ OI . .....o. ... .... ..1. ... . . . .. . . . . . . . . . . . . I C...’ .llA.oF'¢.. Did-O bocrro ..«.. ........ . .. .A. . . . . . A. . .. .. . . . . . ..I.rn...p..' “4/..170 .. .... ..L ... . . ... . . . . .. .~ . . . ... . . . ... (if! f. ..‘.O 0.... . v. c. . ....Tl,. ,- . Q .. . A r .. . . . . fofi’l... 10......“ v.0 . ... . ... Ac. .. . A .. 0.. o . . . v..rl.v.0.l.. .0..oA-.J.b." .. y. .... . . . -..A. .. ... .A - .A .. . . . .... . . . . .. ov.¢.r .0, p . . \...s.. . .. ... . A A ... . .. u. . At. ». , . . . . . ... . . .. ._ .A .0# CI... 1 3 Q. s... A. ._ v. . . h v A O: . . ' _ a . ..II A . .. . A .. . .Ilutdtlo— 00/..nv' .. . . .. . . . . .. .. .. . . . or. a...» O i'hn . g u . . .... . . v I . . . ... C o. . ppm. 10. A .\..usI..o . . . . A . .rA..d._ ..o. I..-rw.o A. I . . . . A . o . . . O A . . . A . _ A . . .1 . .. .. . p . O . . ... . . .. . . . . ... . .. .. . . .. .. . . A . .. o . o \ . . 4. o . . . .. . . . . . . . . . . A . . . A .. . .r. 0 ..fl. .o .. 1.)” r.o.l..nlv.(J.J Al” . ... . .n. c . . .hn..... .. .. ...... ~. . O . l . . .. . . . .... o .1' . .... o .s \‘9 ......J. .9 .. a . u . . o . . . .. c ... o ... \. .I.Y.qu. ..v... . . .0. . . Iv . . .u . .. .. .. p . A. . . . . . . . o I .o ...:o.. nrtfyof /J n a». A.. If, ,’!-r . . . . . . I .. . .. . . A. . .. . . A . . . . . . .. _ . .. . I . . \ 9 . A..\.. . .. s., o. .. ... . o v. .. AV ... I. . . . . . ...I.A . ... . ... ... . ... .. p .. _. . .. . .A . . . _ . . . . .. . . . . .o . s y .. ....... a '0. .... . .... I.J..l..o'nlu .J\ ”(I . . . .... ... ... a . ..... ...9. .... . . .. ... .... . ... . . I; ...; .o .. . . . ... . . A. . . . . . . A... ......f ..r...t... a ....5..... 4.. 7.40.0.- . ‘ .x 5...... ...~ . ..Y. . .... . 9.. . o . r.. . . . . . . _ . - ..:.oI« .. . . , M? IA. I. ...flqu/ 1 I f . . .0 #3177 ..‘W m - , may!!! _ . . ... '- . c . . .. . '0 . I .O .C \ '- . . . ‘ . - . . . . . v I O . . . . . . C . , . O . . l . O . o p . . . A Q .. .. o.... . . .. _.. . . D . o .- . . . . . A a (7 ... cal. . . . . , n .. ' . . . v .0 .... u~ .... . 0 _ _ . . . . . . -. ... r p ('1. . . . I wt . . .. . c . . . . . . . . . . o . ... ... 4.. y . .. A. . . .. .. . .. . .. _ . . .A .....J A _ . . . . . . . . . . . . . . ...... . ,o ... ... ...... ... .... . ... . .. .. .. . _. . . A .. .. .l ..v.. r.. I... ..o. ..d ...”; on I .. v .. . . .. . . - - . . . .. l A . r. .. f . . s. o .. .... .\.- .. a. Q . . .A . c O . . . . ._ A; J. .o ... .Af.v-...r¢.o-..v.A/000001. I ... u . I I . o . . . ._ . ... 4.. I.o . .. .-.A 0.01.19. .,1.. n . . .. . .. . . . . A . . . . A . . . ... a .p . J‘.. ..d’ .. If?..' to 0". . ._ . . . . . .. . . . . A A . ..A . . n . . u .. .. .. . . .. . . . A . . d . v . . . . .44.! . .r I Joni}. .iwfl .. a .. o . . . . . . . I I .I .0 o . . . . .. . . ...... . _ ._ _ . A. . A... 22?. ..1 I . . ... r . I .. . I I Old... ..\L .d'0 A. v. . . .. ... . ._ . . . ..A.... It... .4..V..1.p...¢n . . . .. . . . .. A .. . ... A... .1... . . A . A . . .. . . A o... r/ ...... ... .f I. . . . . ... . AA ... . 1,... 5.2., ...: . I . . . . n . . ..I. .I. 4‘ . . .J: o . . . f . . . . o. . 4 tJVd...-; . . 0.?IO’I.A . . . . . .0 . .. . . A. .A . . ... A. ..ot a (.....A’. .p _ . . . . . ., .. . . .Oal .... 'bl.yl if A' I . . . . . a... ... o . ... I...vcrc:~ra I. A o . . . . v .1 ..rr ... .Ir/.O.Iol O . ..A . .. . _ .o . r. . . .. ..,. 0.1.0 ... . u v . A .. ,4 . . ... a . o: . r1 . V. .J . roo..la. . _ . . . .f. . o I a .v I. IfiI.’ .- . . . . . . n7n.l ....0 .c . u . 5 . A _ . .. .... . . .I:. .A. ....v . AVA.'0¢ v . . F . . . o o . a . . A. .../. ...I..A’ 0 . . . . . . . O I o 01' ..v) I:- . .. .. .. r. . . . . ... Jpn:— v..." o _ . ;. . a o . .1 ’A. .... u o _ . A. v ...a. ...... . . . ... .v r. or . c . .. . . . . . ... . cor '. ‘ . o. .o . .I. o ....o ' vivo’I-iflr’ r . .. . . . . . ..... 1.. . .! Iv. . . A.... ../.., ....I..a ...... u . .- n . . . . . ..A . . ... .n I . . . I. .!.Iol..‘. ...r'lllr . . O 8 c. It. _ . A.. I . A . r. .... .4r....O.Ar. . . ... . ,rr.ot.ArP9.. q . . . o/.o ‘ .001... I u . . I . .I I . n t .I ... _ . r lacuvv l.lo .. ‘ . I o . a '.’,'DOI. . A . . .. ..I.r .. ..o. s. O.‘ '5 O 0.4 .l. . u . .A. . .11.: ....~... IV. (03. o .1 ~... .A r . o . . o . 0.. A dc . . v. . .. . . (”all 9 {fin . .t .. . . . A . .. ..u . A .r (IOIIIAFI }.n\(.....J!J}r.JIO.... A . . ran... . .. . . .a .-.-91"). ...)..I .3. _. . _ v . . . . .0..I . .ofr.c.r r tv’.'v’o..l\. . l , I f O . o. . .. pt! 10'..I.. 0.4,. 5' _w. , . f . A. . . 4,. . .I.’ 01.. 0“. OlvaA_f/r’vl'b N.“ . . . ... .rpn tr: '. . .r... >I'I.A..lou- . i . . - . a . . . v I, I ..4..’.. 1.. ’Oo—v. cl” 0 S . . . .... .o ... n... ...r... . v... ..Ilt‘ . . . .A . , . . . . . r . .... Id': .. r. 0 9-4!. . o'Ol . . A . . . . . .I... .. O. l}...(.) Of.'.(lP..' . E N ._ ...... ¢ . . . . y . .I. . . . . I ..rtl."( . o \ . .. . . o I p A p . . ...r.ao.L....o'A/L .. . A . . . . .. . .... .......J..!... - A. . . .I I o 9 o; ..I;f¢la...l(.lt'a _fi , . . t... . ... or . 1'0! .. .. . ’o. .. .. . .. ' .;I:IID‘JIII (J'II.I.—!A:l . . . ... A. ... ... p . II . r.’ . .. . _. ......- I ...lo'.'.lo ..oo. D ... -... .. s T. . ... .. a. ... . . a all. I ...r.' P'.. I v..,. I. .4... ‘ . .. l I p: I. .c - '.- .v 5‘ . . . . ... . .01....4r.lfl5... v... I . . . . .. . . , . y .. .. ,. .1 . . o . r o ... .’.r.l'6.,0..l I .ff...’ .90. . . . . ..v u . .. I . . v 1 I... . ...!Io ...(.I‘I ..l A. o .. .. .3 I. p .l .: . I . . . .... .... ...-0.0.. 9.. I .A . . . . .... I .vo . 4 .. .. 5 . . A... , v .. . o o'.. . U'bo .. . c . .. o .I 11.. . .u . .... I. U . o. . . . ... a . . c . A . . n . o D ' . a» .vo o 1 . . I .o 1.21.. IA.’ 0 . .. .Dv..v 1 I . A . . O I... an. 0 U'. . . . . I..v u. or . . c . a .on.. '9 , .0 pl . a A I l. o . I a. a. . O . a! .u .... 9' .o . .0 'p- . a a . PI .f.-. . . I . .. . . p...(.. I. o. . . . . . .4. .10! I . _. . .I .f’.. r 4r.... r .. 4 on . o I. incl. .- A." p . . . I.’COI. . I 710. .r . c .. Orlt' . « .. F . . .O.o' . . O a o . l . A |. 1A . .I . a o . ...... . O. I . .. O . o . o. 'o 'r' . Q . . ... ..v.. I. I - 'r . .o ; . . . ... I . . I An. .. u , . .II I ~ . r .. .. I . I. l . . io'..u.'l.'. . . . . III .7 I .. O . . . . u .'A‘. to. V _‘ .. _ I d'. .o. I _h c. r. u! y .. O u o . v I- .0 _ n I no . . - .. .0 .I . A . .A . . - . ..r r. ... . .44 o. . .. o.A I "u _ . .Io. .. n . ... p. . - A l. {.0 . . . o. . s 0 .Ir . . . . . PO. . .ofol.‘. .. . .. ... . . . on DI.- . o . O .I A . . w . . . . . .‘ v .. p I . . . W . . o I rt 0 " a 0- . . .’ . .r . . I . v o . . u . . I. . f O Q - c,» Q I 0 9.. D 'l . . _ '0 ft I ._ n p», I . .. ..- . . . .I .o o l .. o . - . .\-oAu‘| ‘ . .0 D ID... .I .‘Io‘- — .‘t. a..- . . \V... ‘. .... .. o .. '0... O .. .......l|o::\.. A... ...... .I .921. . t .....1.‘ 2...... . A. .. ..... ........ o A O . . \ . o . a . I O p 0.! . | I t O . t. v I I .I\ I . toiobofi'iw Ceu.’v..‘ .1W.‘%“O.fi91 .. .1... .. .. - ...: . . . -. A . . . . p)? 0... .. I 6. ..v’aLmo L... ‘11:. ,\r . Gift-’23.. ..vl...u...... .0..t.l....c. s....\! 3560 .C.7?..... .Iera..... .... o .4... -. -.. A. ....hrt.(?8!:....r\.63... . ...v.. . .....rfofie .. ... .... o) ... .9. I . ......V‘ ..f..f1fly.i.(-~hlu£.. «KEN-A ....:.Ihou.a.4nl.$..l01:.r’lié In 06. ' “hf-.1;- RETURNING MATERIALS: TViESI.) P1ace in book drop to LIBRARIES remove this checkout from M your record. m5 W1” be charged if book—is returned after the date stamped below. COMMUNICATION AND SELF-PERSUASION: SOME FURTHER EXPLORATIONS By Richard L. McGraw A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Communication 1968 5:034 .hxvr. ..ro Ir 4. :1 h ABSTRACT COMMUNICATION AND SELF-PBRSUASION: SOME FURTHER EXPLORATIONS by Richard L. McGraw This study was prompted by the existing research dealing with self-persuasion in terms of a dissonance prediction or an incentive prediction. Both of these theories were employed in this study to make predictions of differential amounts of attitude change both among encoders as well as receivers. One of four hypotheses was confirmed and is consistent with much of the prior research in the area: under conditions of simple commitment, individuals provided with relatively low justification for engaging in belief-discrepant communication will demonstrate significantly greater attitude change in the direction of the discrepant position than will individuals provided.with relatively high justification. Earlier research had also indicated that after message encoding, relatively high justification would produce significantly more attitude change than low justification. This hypothesis was tested but was not confirmed. It was also expected that among those messages encoded by belief-discrepant §s, those which were written under high justi- fication conditions would produce significantly more attitude change among another group of §_s than would those written under conditions Richard L. McGraw of low justification. Further, it was expected that across justification, belief-congruent messages would elicit more attitude change than belief- discrepant messages. Neither of these two hypotheses were confirmed. Possible explanations for the lack of significant findings in three of the four hypotheses are discussed. One explanation may be that the manipulation of high justification failed, the result being that instead of high and lo: justification, there were actually lo: and moderate levels of justification. Accepted by the faculty of the Department of Communication, College of Commmication Arts, Michigan State University, in partial fulfillment of the requirements for the Master of Arts degree. Director 0% Thesis ACKNOWLEDGMENTS Special thanks go to Dr. Gerald R. Miller,thesis director, whose constant encouragement, assistance, patience and constructive criticisms made this thesis possible. Special thanks also go to Dr. Verling C. Troldahl and Dr. Bradley S. Greenberg fer their assistance and cooperation. The Michigan Department of Corrections also deserves a special note of appreciation. Without the aid of several staff members of that department, this study might never have gotten off the ground. For their help in administering questionnaires, the author would like to express his gratitude to Mr. Gary Wilson, Mr. Jack Baseheart, Mr. Elliot Siegel and Mr. Chuck Berger. Commendations are also in order for Mrs. Shirley Sherman and Mrs. Donna Kokx for typing the drafts and manuscript under a great deal of time pressure. Finally, the author wishes to thank his wife, Barbara, who knew just what to contribute at the right time. TABLE OF CONTENTS Chapter I—. INTRODUCTION 0 O O O O O O O O O O O O O O O O O O Justification, Counter attitudinal Commmunication, and Subsequent Self-Persuasion . . . . . . . . Magnitude Of JUStificatiOn o o o o o o o o o o 0 Relative Persuasive Efficacy of Belief-Congruent and Belief-Discrepant Messages . o o o o o o 0 II. PIBTHODS MD PROCEDURES 0 I O O C O O C O O O O O C PhaSeIoooooooo General Procedures . . . . . . . . . . . . . . Phase II o o o o o o o o o o o o o o o o o o o 9 General Procedures 0 o o o o o o o o o o o o 0 III. RESULTS Phase I o o o o o o o o o o o o o o o o o o o o Perceived.Justification . . . . . . HypOtheSiS I o o O o o o o o o o o o o o o o o HYPOtheSiS 2.. o o o o o o o o o o o o o o o o Overa11.Self-Persuasion o o o o o o o o o o 0 Phase II o o o o o o o o o o o o o o o o o o o o HYPOthGSiS 3 o o o o o o o o o o o o o o o o o HYPOtheSiS u o o o o o o o o o o o o o o o o 0 Message Quality 0 o o o o o o o o o o o o o 0 IV. DISCUSSION 0 O O O O O O O O O O O O O O O O O O O BIBLIOGRAPHY 5° 9 o o o o o o o o o o o o o o o o o o o 0 iii 0 O O O O Page UiF' 10 12 17' 18 21 21 2M 26 26 27 28 29 29 32 38 LIST OF TABLES TABLE 1. Summary of Findings Concerning Message Variables Obtained in Prior Self-Persuasion Studies . . . . 2. Distribution of Phase 1 Se in the Four Conditions and the Mean Pretest Attitude Scores . . . . . . 3. Distribution of Phase 2 $8 in the Four Conditions and the Mean Pretest Attitude Scores . . . . . . It. Number of Subjects in Bach Condition Who Chose Not toWriteArguments ............... 5. Attitude Scores and Amount of Change at Three Measmment Times 0 C O O O O O O O O O O O O O 6. Pretest-Posttest Attitude Measures for Phase 2 Se in the Four Conditions . . . . . . . . . . . . 7. Mean Message Ratings for Phase 2 _S_s in the Four cwditions O O O O O O O O O O O O O O O O O 0 iv H ._ A APPEN E APPEN D APPEND APPENDI APPENDIX A: APPENDIX B: APPENDIX C: APPENDIX D: LIST OF APPENDICES OPINION PROFILE QUESTIONNAIRE . . . . OPINION POLL ON CURRENT TOPICS . . . OPINION PROFILE QUESTIONNAIRE . . . o MESSAGE EVALUATION AND QUESTIONNAIRE 48 51 CHAPTER I INTRODUCTION Considerable recent attention has been directed to the general problem of establishing optimum conditions for self persuasion (e.g., Pestinger and Carlsmith, 1959; Cohen, 1962; Janis and Gilmore, 1965; Elms and Janis, 1965; Aronson, 1966; Rosenberg, 1966). The present study examines several issues relevant to this problem: One purpose was to attempt replication of prior findings regarding the relationship of magnitude of justification for engaging in counterattitudinal communication to subsequent self-persuasion, or attitude change. A second, and major purpose was to compare the relative persuasive efficacy of belief-discrepant communications prepared under conditions of high and low justification. Finally, a third purpose was to examine the relative persuasive efficacy of belief-discrepant and belief-congruent communications, a question apparently untouched by prior research. Justification, Counterattitudinal Communication, and Subsequent Self-Persuasion There are two competing positions regarding the relationship between justification fer engaging in belief-discrepant communication behavior and subsequent selfepersuasion. Dissonance theory (Festinger, 1957) posits a negative relationship between the two variables: the less the justification, the greater the subsequent self-persuasion. ' Conversely, incentive theory (Janis and Gilmore, 1965) holds that the two variables are positively related: the greater the justification, the greater the subsequent self-persuasion. Support for the dissonance position finds its genesis in a study by Pestinger and Carlsmith (1959). After undertaking a dull, boring task, subjects were "hired" to perform the belief-discrepant task of convincing the next subject -- actually a confederate of the experimenter -- that the task was interesting and enjoyable. Subjects in the Low Justification condition were paid one dollar for’praising the dull task, while those in the High Justification condition were paid 20 dollars. Consistent with theoretic predictions, the one dollar subjects increased their post-communication ratings of the dull task significantly more than did the 20 dollar group. Rosenberg (1965) has questioned the dissonance interpretation of Festinger and Carlsmith's findings and has suggested that the failure of the 20 dollar group to increase their ratings of task attractiveness can be attributed to a variable labeled evaluation apprehension: a fear that the experimenter was trying to engineer a successful bribe. However, Cohen (1962) has replicated Pestinger and Carlsmith's results using 50 cent and one dollar conditions, a difference hardly calculated to produce extreme inter-condition variations in evaluation apprehension. Thus, there is at least some evidence to buttress the dissonance prediction of a negative relation- ship between amount of justification and subsequent self-persuasion. Several studies support the incentive theory prediction that justification and selfepersuasion are positively related (Scott, 1957; Scott, 1959; Bostrom, Vlandis, and Rosenbaum, 1961; Elms and Janis, 1965). Of particular interest is a study by Janis and Gilmore (1965). College students were asked to write arguments espousing the value of a disliked undergraduate course sequence. Justification was varied by attribution of sponsorship: in the High Justification condition, subjects were told the arguments were being written fer a national research agency employed by a number of universities to study the undergraduate curriculum; while in the Low Justification condition, they were told the arguments would.be used as promotional material by a new publishing company. Measures taken after the arguments were written indicated that subjects in the High Justification condition expressed more favorable attitudes toward the previously disliked courses than did subjects in the Low Justification group -- a finding consistent with incentive theory expectation and at odds with dissonance theory. Another aspect of Janis and Gilmore's results should, however, be considered. The process of engaging in belief-discrepant communi- cation behavior can be divided into at least two distinct phases. Phase 1 involves only a simple public commitment to engage in counter- attitudinal communication; i.e., the individual agrees to encode a message that is at odds with his prior beliefS. Phase 2 involves the actual performance of encoding behavior; i.e. , the individual goes through the complex.process of cognitively assessing, and subsequently encoding belief-discrepant arguments. In the Janis and Gilmore study, Phase 1 (if I was represented by two groups of subjects who_agreed to write agruments, but who did not acutally engage in encoding behaviors. For these groups, the findings, although not significant, favor a dissonance interpretation: Subjects who agreed to write arguments for the relatively unfavorable Sponsor expressed somewhat more attitude change than did subjects who agreed to write for the favorable sponsor. These results have led Rosenberg (1966) to argue that a negative relationship between magnitude of justification and selfepersuasion (dissonance prediction) may hold at the simple commitment level, while a positife relationship between the two variables (incentive prediction) may hold for situations in which the individual has actually encoded the belief-discrepant communication. This possibility served as a basis for the replication hypotheses tested in the present study. Given only agreement to engage in counterattitudinal communication, a dissonance effect was predicted, Specifically: Hypothesis 1: Under conditions of simple commitment, individuals provided with relatively low justification for engaging in belief- discrepant communication will demonstrate significantly greater attitude change in the direction of the discrepant position than will individuals provided with relatively high justification. Given actual perfOrmance of encoding behaviors, an incentive effect was predicted, specifically: Hypothesis 2: Following actual encoding, individuals provided with relatively high justification for engaging in belief-discrepant communication will demonstrate significantly greater attitude change in the direction of the discrepant position than will individuals provided with relatively low justification. In other words, if Tl represents the pretest measurement, T2 the simple commitment measurement, and T3 the post-encoding measurement, then Hypothesis 1 predicts greater attitude change for the low justification . group from T1 to T2. Conversely, Hypothesis 2 predicts greater attitude change fer the high justification group from T2 to T3. It should be noted that the general aura of controversy surrounding the relative merits of the dissonance and incentive positions, as well as the conflicting findings obtained in prior research, allowed for the definite possibility of alternative outcomes. Even so, any added evidence concerning the two competing positions should assist communication researchers in attempting to arrive at conclusions con- cerning the relationship of justification to subsequent selfapersuasion. Magnitude of Justification and the Persuasive Efficacy of Belief-Discrepant Communications As indicated above, a major purpose of this study was to examine the relative persuasive efficacy of belief-discrepant communications prepared under conditions of high and low justification. Proponents of incentive theory hold that a condition of relatively high justification should result in better quality belief-discrepant communications than a condition of relatively low justification. The rationale for this prediction can be summarized as follows: According to "inceitive theory," the attitude changes produced by role-playing are mediated by intensive "biased scanning" of positive incentives, which involves two types of verbal response: (1) fulfilling the demands of the role-playing task by recalling and inventing arguments that are capable of functioning as positive incentives for accepting a new attitude position, and (2) appraising the recalled and improvised arguments with a psychological set that fosters Open-minded cognitive exploration of their potential incentive value, raEHEPIthan a negativistic set of the type engendered by the arousal of feeling of hostility, resentment, or suspicion. (Elms and Janis, 1965, p. 59). High justification for engaging in belief-discrepant communication best satisfies the two criteria mentioned by Elms and Janis, for low justification serves as one negative incentive that both interferes with the open-minded cognitive exploration of opposing arguments and impedes the invention and recall of such arguments. Thus, one would expect a better quality belief-discrepant message if the negative incentive of low justification were eliminated, or if the positive incentive of high justification were added. While several studies have dealt with differences in belief- discrepant messages produced under conditions of high and low justification, the findings are relatively inconclusive. As Table 1 indicates, such variables as quality of arguments, length of essays, number of belief-discrepant arguments, and self-ratings by subjects of their own belief-discrepant communications have been examined. Although some differences have been obtained, the results are liberally sprinkled with non-significant outcomes. The message quality measure of central interest of this study was the relative persuasive efficacy of belief-discrepant messages prepared under conditions of high and low justification. If there are systematic differences in quality between persuasive communications prepared at the two levels of justification, these differences should be reflected by the messages' relative persuasive impact. At least three studies (Rosenberg, 1965; Carlsmith, _e_1_:_. ;a_l_., 1966; Linder, 33, 31,, 1967) have examined the persuasiveness variable, but only Rosenberg found significant differences: the high justification messages were rated more persuasive than the low justification. A major limitation of these studies lies in the technique used to assess persuasiveness. Both Rosenberg and.Linder, et,_al, used two trained judges to rate the variable, while in the Carlsmith st? 21, study, the authors themselves rated.the messages. Several apparent difficulties are festered.by this method: first, Linder’etf‘al. report severe problems in achieving high inter-judge reliability. Second, even though judges supposedly engaged in "blind" rating, there are still possibilities for experimenter bias (Rosenthal, 1966). This problem is particularly pronounced in the case involving ratings by the authors themselves. While Carlsmith gt. al_.'s ratings did not produce significant results, differences might have occurred in the opposite direction had unbiased ratings been obtained. Finally, use of trained raters -- e‘.g., Linder 3:. all. used varsity debaters -- increases the possibility that criteria other than those normally used by relatively naive audiences might have served as the bases for judgments of persuasiveness . Table 1. Stmunary of findings concerning message variables obtained in prior self-persuasion studies.* Study Variables Results Festinger- 1. "Strength" of S's positive Carlsmith statements about the task NSD 2. Rating of overall content NSD 3. Rating of persuasiveness of _S_. NSD 1:. Amount of time S spent dis- cussing task (not actual time, but rated) NSD Janis- l. Arguments supporting position S's in public sponsor- Gilmore $20.00 condition gave largest number 2. Arguments against position NSD Elms- 1. Quality of arguments Overt Role-Players-- Janis higher quality argu- ments in the favorable Sponsorship condition. No differences among mmetary reward groups. 2. length of essay Higher paid §_'s produced longer essays Rosenberg 1. Length of essay Higher paid S's produced longer essays 2. Basic persuasiveness of Higher paid S's essay produced more per- suasive essays Carlsmith l. Role-play performance: gt. al. a. Persuasiveness NSD b. Overall positiveness NSD c. Overall positiveness 8 conviction NSD (1. Time Spent on assigned topic NSD e. Dissociation of self from content NSD Table l--continued Stufl Variables Results Carlsmith 2. Accomplice's ratings of S. et. a1. a. Rated S's on 1. a., b., c. NSD _ — b. Apparen-t conflict 50 cent S's diSplayed more conflict c. Signs of discomfort NSD 3. Essay PerfOrmance a. Emphasis used in making points NSD b. Elaboration of general theme NSD c. Overall quality and per- suasiveness NSD d. Apparent effort NSD Linder 1. Experiment #1 gt. _a_.]._. a. Number of words per essay NSD b. Degree of organization NSD c. Intent to persuade NSD 2. Experiment #2 a. Persuasiveness NSD b. Degree of organization NSD c. Number of words NSD d. Extremity of position advoca- ted NSD *Berger, 1968, pp. 16-17. These problem can be largely eliminated by using the high and low justification messages as experimental stimuli in a typical attitude change paradigm. In other words, the present study assessed the relative persuasive efficacy of belief-discrepant messages prepared by subjects under conditions of high and low justification by presenting 10 the messages to subject audiences and measuring attitude change fOr each audience. Consistent with incentive theory predictions, the following major hypothesis was tested: Hypothesis 3: Belief-discrepant messages prepared under conditions of high justification will elicit greater audience attitude change in the A advocated direction than will messages prepared under conditions of low justification. The Relative Persuasive Efficacy of Belief-Congruent and Belief-Discrepant Communication A third purpose of this study was to examine the relative per- suasive efficacy of belief-congruent and belief—discrepant communi- cations. While considerable attention has been directed at identié fication of the variables that impede or facilitate the persuasiveness of belief-discrepant messages, the researcher knows of no instance in which the relative impact of belief-congruent and belief-discrepant messages has been assessed. The present study provides an Opportunity for such assessment. Extent theory and research points to the likelihood that belief- congruent messages will be more persuasive. Such notions as frame of reference (Levine and Murphy, 19143; Jones and Aneshansel, 1956) and selective exposure (Freedman and Sears, 1965) imply that individuals should be able to marshall more belief-congruent arguments. Assuming that the availability of information and arguments is related to persuasive success, the following hypothesis was tested: Hypothesis 4: Audiences eXposed to messages encoded under belief-congruent conditions will report greater attitude change in the advocated ' direction than will audiences eXposed to messages encoded under blief-discrepant conditions. 11 It was expected that the predicted difference, while perhaps more pronounced in the high justification condition, would hold for both levels of justification. CHAPTER II METHOD AND PROCEDURES This study was conducted in two phases. During the first phase, data for testing Hypotheses l and 2 were collected and the Phase 2 stimuli were written by Phase 1 Se. Phase 2 provided the data for testing Hypotheses 3 and u. Phase 1 SS for Phase 1 were 178 inmates at the Michigan Training Unit (MTU) in Ionia, Michigan. All Se were obtained from a group of educationally advanced inmates, and ranged in age from 17 through 23. They were either enrolled in an accelerated high school curriculum at MTU, or were studying one of three trade programs: data processing, drafting, or computer programming. General Procedures: In order to determine the issue to be used in the study, a pre- test containing 12 tOpics was administered (see Appendix A). All SS rated each of the topics on four, seven-interval, semantic differential-type scales bounded.by adjectives loading high on the evaluative dimension. To determine an Sjs attitude toward a given tOpic, the sum of his scores on each of the four pairs of adjectives was used. Thus, a score of four (the lowest possible) would indicate l2 13 a strong negative attitude toward the particular topic, while a score of 28 (the highest possible) would indicate a strong positive attitude toward that topic. As a result of the pretest, the issue of a lottery draft system was chosen for use in the study. Since attitudes were generally bimodally distributed on this topic, it afforded an Opportunity to obtain the belief-congruent and belief-discrepant messages used in Phase 2. SS were assigned to Belief-Congruent or Belief-Discrepant conditions on the basis of their pretest attitude scores on the lottery draft system issue. SS with pretest attitude scores ranging from 17 through 28 were assigned to the Belief-Congruent conditions, while SS with pretest scores ranging from u through 10 were placed in the Belief-Discrepant treatments. SS with pretest scores of 11 through 16 were not used in the study. Within Belief-congruent and BeliefeDiscrepant conditions, §s were assigned to High Justification and Low Justification treatments on a stratified random basis, controlling for pretest attitudes. These stratification procedures stemmed from a desire to have approximately the same number of SS at a given pretest attitude intensity within the High Justification and Low Justification con- ditions. For example, of the 35 SS with a pretest attitude of four (maximally unfavorable), 17 were randomly selected for the High Justification group and 18 were randomly assigned to the Low Justification group. The procedure therefOre ensured a relatively 1H comparable distribution of pretest attitude intensity within the four experimental groups. Table 2 contains a breakdown of the dis- tribution of SS in the fbur groups as well as the mean pretest attitude scores for each group. Table 2. Distribution of the Phase 1 Se in the Four Conditions and the Mean Pretest Attitude Scores for Se in Each Condition. Condition BeliefeCongruent Initial High Low Attitude Justification Justification Score Number of SS Number of SS 28 12 12 27-26 2 2 25-2“ 5 5 23-22 2 2 21-20 u M 19-17 3 3 Mean pretest attitude score: 2n.28 2u.61 Condition Beliefeniscrepant Initial High Low Attitude Justification Justification Score Number of 83 Number of SS 10 5 u 9 2 2 8 2 2 7 u u 6-5 3 3 n 17 18 Mean pretest attitude score: 5.97 5.79 .‘12- 2.. In. H' 15 Three days after the pretest, _S_s were sent to the appropriate experimental room. Originally, it had been decided to put all the High Justification SS in one room and the Low Justification _S_s in another. However, Since physical capabilities of MTU allowed a maximum of 30 people in any room, the four conditions were assigned separately to rooms. All _S_s were asked to write arguments in favor of a lottery draft system. Justification was manipulated in terms of the Sponsorship of the request. _Ss in the High Justification conditions were given the following instructions: A committee of college teachers and presidents is working in cooPeration with Michigan State University on a study of the merits of a lottery draft system. AS you know, a lottery draft system would involve choosing people for military service on the luck of the draw. There are many arguments for and against such a system. It would be helpful to the committee if you would write some of your arguments down for them. You are all to write arguments and points in favor of a lottery draft system; that is you are to argue in favor of drafting solely on the luck of the draw. Your arguments will provide valuable assistance to this comittee in developing their report and they hope you are all willing to write the arguments. SS in the Low Justification conditions received the same instructions, with the single exception of sponsorship attribution. Low Justification is were given the following information about the Sponsor: A comercial publishing conpany is preparing a book explaining how young men can get out of military service, that is, how they can avoid the draft. In one chapter of this book, they wish to write about the merits of a lottery draft system. As you know, a lottery draft system would involve choosing people for military service on the luck of the draw. There are many arguments for and against such a system. It would be helpful to the publishing company if t‘xf‘T" E l6 you.would.write some of your arguments down for them. You are all to write arguments in favor of a lottery draft system; that is, you are to argue in favor of drafting solely on the luck of the draw. Your arguments will provide valuable assistance to the publishing company in preparing their book and they hope you are all willing to write the arguments. After the initial instructions were given, the §_gave SS the choice of writing or not writing the arguments favoring a lottery draft system. Those who chose not to write were told to report to the superintendent's office and were subsequently sent back to class. It was assumed that those who remained had committed themselves to writing arguments in favor of a lottery draft system. At this point, a second attitude measure was obtained in order to gather data relevant to Hypothesis 1, the hypothesis regarding simple commitment self- persuasion effects. In order to avoid such possible contaminating effects as evalua- tion apprehension (Rosenberg, 1965), an attempt was made to link this second measurement with a different study. The E made the fellowing remarks to all Se: Before you.write them (the arguments), there is one other thing we would like you to do. Another professor at Michigan State is doing a poll on peeple's feelings about current topics. He has a.brief questionnaire he would like you to fill out. I'll distribute them now and _ give you some instructions on how to fill them out. After they are done, we'll go on with the writing. In addition, while the same four scales were used to measure attitudes toward a lottery draft system, the format of the second questionnaire varied from that of the pretest instrument and several new masking topics were added (Appendix B). 17 After completing the second questionnaire, all Es were given a third form containing paper for writing their arguments and the third, post-encoding questionnaire. _S_s were given two blank sheets of paper, with the third questionnaire following the blank pages. The Es were instructed not to go beyond the second blank page until told to do so by E. They were given 20 minutes to write arguments. At the end of 20 minutes, the S_s were instructed to turn the page and to complete the third questionnaire. This third questionnaire (Appendix C) was similar in format to the pretest questionnaire, but was shorter. It contained only four items, the last of which was "A Lottery Draft System," with the same four pairs of adjectives used on the preceeding measures. Data from this measurement were used to test Hypothesis 2, the hypothesis regarding post-encoding self- persuasion effects. Following completion of this questionnaire, _S_s were thanked for their participation in the study. Casual conversation with a number of _S_s revealed that, unlike some student subjects, they were unfamiliar with the theoretic notions being tested in the study. Phase 2 _S_s for Phase 2 were 153 students enrolled in a basic under- . graduate social science course at Michigan State University. The purpose of Phase 2 was to gather data for testing Hypotheses 3 and u of the study . 18 General Procedures: A pretest measure of attitudes toward a lottery draft system was obtained during a regular class period. The pretest questionnaire for Phase 2 was identical to the post-encoding questionnaire used in Phase 1 (Appendix C). ES used in Phase 2 reported pretest attitudes ranging from strongly opposed to moderately in favor of a lottery draft system. Es were assigned to one of four conditions: Belief-Congruent, High Justification; Belief-Congruent, Low Justification; Belief- Discrepant, High Justification and Belief-Discrepant, Low Justification, using the same stratified randomization procedures employed in Phase 1. Since it was impossible to predict how many pretest _S_s would be in class on the day of the experiment, surplus Es were assigned to each condition. Table 3 contains a breakdown of the pretest attitude scores of all Es originally assigned to the four conditions as well as the mean pretest attitude scores for _S_s in each condition. 19 Table 3. Distribution of the Phase 2 Es in the Four Conditions and the Mean Pretest Attitude Scores for Subjects in Each Condition . Condition Belief-Congruent Initial High Low Attitude Justification Justification Score Number of Es Number of Es u 3 2 5 1 1 6-7 2 l 8 2 1 9-10 3 3 11-12 3 2 13-15 M 3 16 6 6 l7 2 2 18-19 2 l 20 2 1 Mean pretest attitude score: 12.uo 12.88 Condition Belief-Discrepant Initial High Low Attitude Justification Justification Score Number of Es Number of Es u 3 3 5 1 1 6-7 1 l 8 1 1 9-10 3 3 11-12 2 2 13-15 3 3 16 6 6 17 2 2 18-19 2 2 20 2 2 Mean pretest attitude score: 12.73 13.16 20 The Phase 2 experiment was conducted during class two days later. Since the class was in a large lecture auditorium, Es were divided into four groups seated in the four corners of the auditorium, with an E assigned to each group. Individuals not assigned to any of the experimental groups were seated in the center of the auditorium and were given the task of completing an irrelevant questionnaire. Each E_ read his list of names from top to bottom, handing a message to each E present. Those Es with the most unfavorable pretest attitudes toward a lottery draft system thus were the first to receive messages and those with more moderate attitudes received the remainder of the messages. Those surplus Es who did not receive a message were sent to the fifth group. In order to ensure that the Es read the arguments thoroughly, they were told to underline the major points in the message. After they had read the arguments, the Es filled out a questionnaire containing items fer evaluating the messages' arguments, information, organization, and overall quality as well as the same four semantic differential type scales for measuring posttest attitudes toward a lottery draft system (Appendix D). The message rating items were scored on a one (Very Poor) to five (Very Good) basis, while the attitude scores were obtained by summing across the four scales. CHAPTER III RESULTS Phase 1 For all statistical tests, the .05 level of Significance was required. Analysis of the data yielded the following results. Perceived Justification: In prior self-persuasion research, there has been considerable discussion and controversy concerning manipulation of the level of perceived justification for engaging in counterattitudinal communication (e.g., Aronson, 1966; Rosenberg, 1966). For particular studies using monetary incentives, some writers have argued that the amomt of money offered to high justification subjects is frequently so great that negative incentives are created; and conversely, that the amount of money offered to low justification subjects--rather than constituting minimal justification--is about the payment they would normally expect (Rosenberg, 1965). Moreover, certain theorists have pointed to extraneous sources of justification that may confound manipulation attempts. For instance, Rokeach has suggested that many college students may perceive that participation in a scientific experiment justifies their belief-discrepant behaviors, apart from any monetary payment extended for those behaviors. 21 22 The present study sought to provide low justification subjects with quite minimal grounds for engaging in belief-discrepant behavior. To avoid problems of assessing perceptions of varying monetary amounts, the sponsorship technique employed by Janis and Gilmore (1965) was used to manipulate justification. It was expected that the task of writing arguments about a lottery draft system fer a publisher pre- paring a book on procedures for avoiding the draft would be repugnant to most subjects, and would constitute minimal behavioral justification. Conversely, the task of preparing arguments for an altruistic, scholarly committee engaged in studying means for ensuring an equitable draft system should be perceived as more justifiable and acceptable by the subjects. Participation data from the study indicate that the manipulation was successful. Probably the most straightforward test of perceived justification is a locomotion criterion; i.e., the number of subjects in each condition who chose to leave rather than to participate in the study. Table A contains a summary of the number of subjects in each condition who chose not to write arguments. It can be seen that seven of 51 subjects assigned to the two High Justification conditions chose not to engage in counterattitudinal communication. By contrast P" 23 Table 4. Number of Subjects in Each Condition Who Chose Not to Write Arguments. Condition Belief Belief Belief Belief Congruent Congruent Discrepant Discrepant Hi Just. Low Just. Hi Just. Low Just. Initial Es 25 25 26 29 Es not writing 1 12** 6 10** Es writing 2a 13 20* 19 *Additionally, one §_in this cell who stayed left after two minutes sa ing he could.not think of anything to write. **X = 9.35; df. = 3; p. <:.OOl. 22 of 5M subjects assigned to the Low Justification conditions declined to participate. Analysis of these differences yielded a significant X2 of 9.35, indicating that a significantly greater number of Low Justification subjects chose not to write arguments. The withdrawal rate also suggests that subjects did perceive they had a choice about participating in the experimental task. This fact is important, for such perceived choice is a necessary condition for an optimal test of dissonance theory predictions. 21+ Hypothesis 1: Self-Persuasion Following Simple Commitment: Table 5 contains the mean pretest attitude scores, mean attitude scores following simple commitment, and mean post-encoding attitude scores for subjects in the four conditions. The Belief-Congruent: High Justification and Belief-Congruent: Low Justification conditions were included primarily to obtain message stimuli for Phase 2 and to provide a baseline for assessing possible sensitization effects from pretest to experiment. Since the mean attitude scores for these two groups remained stable across the three times of measurement, it would seem that attitudes were not modified.by factors extraneous to the experiment. Of particular interest are the pretest-simple commitment attitude change scores for subjects in the Belief-Discrepant: High Justification and Belief-Discrepant: Low Justification groups. As Table 5 indicates, both conditions report significantly more favorable attitudes toward a lottery draft system following simple commitment to write‘ counter- attitudinal arguments favoring such a system. However, this effect is much more pronounced in the Belief-Discrepant: Low Justification condition, the group in which maximum dissonance was expected. A subsequent £_test comparing the mean pretest-simple commitment attitude change scores for the two groups revealed that significantly more self-persuasion occurred in the Belief-Discrepant: Low 25 Table 5. Attitude Scores and Amount of Change at Three Measurement Times. Message Encoding Simple Tl-T2 Post- T -T3 T 'T3 Condition Pretest Commitment Change Encoding CEange Ciange Belief-Congruent 25.38 23.n3 -1.95 2u.1u +.71 -1.2u High Justifica- tion 11:21 Belief-Congruent 25.75 2u.u5 -1.29 2u.61 +.15 -1.1u Low Justifica- tion n=13 Belief-Discrepant s.uo 10.33 +u.93* 10.60 +.27 +5.20% High Justification n=15 Belief-Discrepant 5.52 17.21 +11.69* 17.26 +.05 +11.7u* Low Justification n=19 #3 test p = <:.OS one tailed Justification condition (t = 2.71). Thus, the data support Hypothesis 1: Under conditions of Simple commitment, individuals provided with relatively low justification for engaging in belief-discrepant communication will demonstrate significantly greater attitude change in the direction of the discrepant position than will individuals provided with relatively high justification. 26 Hypothesis 2: Self-Persuasion Following Message Encoding: Of particular interest in testing Hypothesis 2 are the simple commitment-post-encoding attitude change scores for subjects in the Belief-Discrepant: High Justification and Belief-Discrepant: Low Justification conditions (Table 5). Hypothesis 2 predicts a facilitation effect in self-persuasion for subjects in the Belief- Discrepant: High Justification grow; i.e., it rests on the incentive theory assumption that encoding.belief-discrepant arguments under conditions of high justification creates reinforcing contingencies that enhance self-persuasion. Examination of Table 5 reveals no support for this hypothesis. In both Belief-Discrepant groups, there was virtually no simple commitment - post-encoding attitude change. Possible reasons for failure to confirm Hypothesis 2 are discussed in Chapter u. Overall Self-Persuasion: Taken together, Hypotheses l and 2 militated against an hypothesis concerning which of the two Belief-Discrepant conditions would report greatest self-persuasion from pretest through post- encoding, for it was impossible to predict whether the expected dissonance effect following simple commitment or the expected incentive effect following actual encoding would be of greater magnitude. Even so, it is interesting to note the overall magnitude of self- persuasion for the two groups. As Table 5 indicates, subjects in both 27 in Belief-Discrepant: High Justification and Belief—Discrepant: Low Justification conditions report significant attitude change from pretest through actual encoding of the counterattitudinal arguments. This effect is significantly greater, however, for subjects in the Low Justification grow (3 = 2.28). Moreover, it should be emphasized that the locus of this effect is to be found entirely in pretest- Simple commitment change. Change following actual encoding is negligible for both grows. Phase 2 Table 6 contains the pretest-posttest attitude scores and the attitude change measures for Phase 2 subjects who read messages prepared in the four Phase 1 conditions. In order to determine if the messages were persuasive, 3 tests for correlated measures were calculated for pretest-posttest change in each condition. Since all differences were significant, it can be concluded that each of the four messages was persuasive. 28 Table 6. Pretest-Posttest Attitude Measures for Phase 2 Ss in the Four Message Encoding Conditions* Time Condition Pretest Posttest Change Belief-Congruent 12.57 lu.63 +2.06** High Justification 2519 Belief-Congruent 15.07 17.92 +2.85** Low Justification £313 Belief-Discrepant 1u.1u 16.67 +2.su** High Justification 3315 Belief-Discrepant 13.93 16.68 +2.75** Low Justification 2316 *u = maximally unfavorable attitude; 28 = maximally favorable attitude. ** t test p = 41.05 one tailed Hypothesis 3: Relative Persuasive Efficacy of Belief- Discrepant Messages Prepared under Conditions of High and Low Justification: Examination of Table 6 reveals negligible differences in the per- suasive efficacy of the four sets of messages. Since the cell sizes are unequal, a two-factor analysis of variance was first carried out in order to determine if any of the effects were of significant magnitude to warrant an approximation analysis for unequal nfs. Since the initial analysis yielded F ratios ofhd =p=(.05 3 test one tailed CHAPTER IV DISCUSSION By far the most striking effect observed in the present study is the substantial self-persuasion that occurred among Phase 1 subjects in the Belief-Discrepant: Low Justification condition fbllowing simple commitment to engage in counterattitudinal communication. The amount of attitude change (11.69 units) is of a magnitude not typically reported in prior dissonance research dealing with self-persuasion and provides strong support for Hypothesis 1, as well as for Rosenberg's (1966) suggestion that dissonance effects should hold for situations involving only simple commitment. It is probable that the extreme magnitude of self-persuasion among Belief-Discrepant: Low Justification subjects results from the creation of Optimal conditions for dissonance arousal. As indicated above, both the withdrawal rate and the verbal remarks of the subjects indicate that they perceived the circumstances as barely minimal to justify engaging in belief-discrepant behavior. In addition, the high withdrawal rate suggests that subjects felt they had considerable freedom to choose whether or not they wished to participate in writing counterattitudinal arguments. Since both minimal justification and a high level Of perceived choice are necessary theoretical antecedents 32 33 for maximum dissonance arousal, and since magnitude of dissonance and amount of self-persuasion are positively related, the success of the present manipulations may have enhanced the robustness of the obtained, predicted effect. The hypothesis which stated that high justification will facilitate selfepersuasion following actual encoding of belief- discrepant arguments was not supported by the present findings. Phase 1 subjects in the Belief-Discrepant: High Justification condition demonstrated.practically no attitude change during the simple commit- ment through post-encoding period. There are several possible explanations for this lack of confirmation. One defensible inter- pretation lies in the possibility that, unlike the low justification situation, the high justification manipulation was not entirely successful. Rather than achieving high justifying conditions in any absolute sense, it is possible that subjects perceived the situation as providing moderate, or even moderately low justification for engaging in counterattitudinal communication. Several considerations lend credence to this interpretation. First, six of 26 initial subjects in the Belief-Discrepant: High Justification condition chose not to write beliefediscrepant arguments, a circumstance suggesting that justification was not perceived as extremely high. Moreover, Belief-Discrepant: High Justification subjects demonstrated a significant and sizeable amount of self- persuasion following simple commitment to write belief—discrepant messages. Since incentive theory holds that the motivational dynamics 34 of self-persuasion are rooted in the actual encoding process, this change fellowing simple commitment indicates that Belief-Discrepant: High Justification subjects may have been experiencing relatively high dissonance -- dissonance that most likely would result from perceptions of relatively low justification. In short, rather than high and low, the levels of justification psychologically established in this study may actually have been moderately low and extremely low. Future research could examine this possibility by making the high justification manipulation more powerful. For example, in addition to linking the task with a favorable sponsor, monetary payment could be extended for engaging in counterattitudinal communication. There is a second reason why message encoding of discrepant arguments under conditions of high justification may not have resulted in subsequent self-persuasion. Incentive theory holds that increases in justification lead to more biased positive scanning of former negative arguments, and that the reinforcement derived from such biased scanning in turn leads to greater self-persuasion. It may well be, then, that a necessary condition for confirming incentive theory predictions is the use of subjects who are cognitively familiar with a number of relevant arguments about the particular issue. In the present study, it appeared as if most subjects, while reporting strong affective reactions to a lottery draft system, were not acquainted with many arguments for or against such a system. For the most part, the written essays emphasized one or two very general arguments relating to the universal "fairness" of a lottery system. I 35 Thus, the circumstances surrounding encoding of belief—discrepant messages may have militated against incentive based self-persuasion. While admittedly Speculative, this possibility could be tested by using groups of subjects who differ in their initial familiarity with the message tapic. Neither Hypothesis 3 nor 4 is supported.by the Phase 2 findings of the study. Contrary to theoretic expectations, belief-discrepant messages prepared under conditions of high justification were no more persuasive than counterattitudinal communications written under low justification conditions. Two plausible explanations for this lack of differential persuasive efficacy have already been discussed: first, it is possible that high justification conditions were never adequately established in the study: or second, it may be that subjects were so unfamiliar with arguments about a lottery draft system that the necessary biased scanning of discrepant arguments could not occur. In fact, failure to confirm the earlier hypothesis concerning the facilitative effect of high justification on self-persuasion (Hypothesis 2) would almost seem to dictate a subsequent lack of confirmation for Hypothesis 3. In addition, a potential procedural shortcoming in the Phase 2 data collection may have militated against confirmation of Hypothesis 3. Due to pressures of time, the experimenter was able to allow only two days between the pretest and the subsequent message exposure and posttest. Although every effort was made to disassociate the two events 36 (e.g., creating the illusion of two different experiments and using different formats for pretest and posttest questionnaires) a number of subjects indicated verbally that they believed the two steps were related. Given such sensitization possibilities, it is difficult to ascertain whether the Phase 2 subjects were re8ponding to the message stimuli or to the demand characteristics of the experimental situation. Failure to confirm the hypothesis that belief-congruent messages will be more persuasive than belief-discrepant comunications (Hypothesis 4) is readily explainable. It will be recalled that the rationale for this hypothesis was rooted in the belief that Phase 1 subjects would be familiar with a greater number of belief-congruent than belief-discrepant arguments. As indicated earlier, this was not the case; rather, the subjects seemed to be familiar with only a limited.number of arguments on either side of the lottery draft system issue. This lack of differential familiarity is further reflected.by the fact that ratings of such qualitative message dimensions as amount and quality of information did not differ significantly for belief- congruent and belief-discrepant messages. An adequate test of Hypothesis u must therefore await a replication using subjects who are more familiar with the arguments associated.with the experimental issue. Thus, of the four hypotheses tested in the present study, only one is clearly supported. Even so, this single positive result is an interesting one which provides strong support for the dissonance 37 position that justification and self-persuasion are negatively related, at least at the level of simple commitment. The findings indicate that if an individual_agrees to take a position at odds with his prior beliefs, and if minimal justification is offered for taking the position, considerable attitudinal modification, or self-persuasion, will occur. This fact would appear to have numerous implications for the communication strategist who seeks to use involvement as a means for achieving attitudinal and behavioral support. BIBLIOGRAPHY _ Aronson, E. "The Psychology of Insufficient Justification: An Analysis of Some Conflicting Data," in Feldman, 8. (Ed.) Cognitive Consistency. New York: Academic Press, 1966, pp. 115-136. Berger, C. R. "The Effects of Influence, Feedback and N Influence on the Relationship Between Incentive Magnitude and Attitude Change." Unpublished Dissertation proposal at Michigan State University, April 3, 1968. Bostrom, R.; Vlandis, J. and Rosenbaum, M. "Grades as Reinforcing Contingencies and Attitude Change." Journal of Educational Psychology, 1961, 52, 112-115. Carlsmith, J. M., Collins, B. E., and Helmreich, R.L. "Studies in Forced Compliance: The Effect of Pressure for Compliance on Attitude Change Produced by Face-to-Face Role Playing and Anonymous Essay Writing." Journal of Personality and Social Psychology, 1966, 1+, 1-13. Cohen, A. R. and Brehm, J. W. Explorations in Cognitive Dissonance. New York: John Wiley 8 Sons, 1962. Elms, A. C., and Janis, J. L. "Counter—norm Attitudes Induced by Consonant vs. Dissonant Conditions of Role Playing,” Journal of Experimental Research in Personality, 1965, 1, 50-60. Festinger, L. A Theory of Cognitive Dissonance. Stanford, Calif.: Stanford Univ. Press, 1957. Festinger, L. and Carlsmith, J. M. "Cognitive Consequences of Forced Compliance." Journal of Abnormal and Social Psychology, 1959, us, 92-99. Friedman, J. L. and Sears, D. 0. "Warning, Distraction and Resistance to Influence." Journal of Personality and Social Psychology, 1955, 1, 262-266. Janis, I. L. and Gilmore, J. B. "The Influence of Incentive Conditions on the Success of Role Playing in Modifying Attitudes." Journal of Personality and Social Psychology, 1965, 1, l7-27. 38 111%: 411' I'll. ‘ 39 Jones, E. E., and Aneshansel, J. "Learning and Utilization of Contravalent Material." Journal of Abnormal and Social Psychology, 1956, 53, 27-33. Levine, J. M. and Murphy, G. "The learning and Forgetting of Controversial Material." Journal of Abnormal and Social Psychology, 1993, 38, 507-517. Linder, D., Cooper, J., Jones, E. "Decision Freedom as a Determinant of the Role of Incentive Magnitude in Attitude Change." Journal of Personality and Social Psychology, 1967, 6, 2'45-25“. Rosenberg, M. J. "When Dissonance Fails: on Eliminating Evaluation Apprehension from Attitude Measurement." Journal of Personality and Social Psychology, 1965, l, 28-u2. Rosenberg, M. J. "Some Limits of Dissonance: Toward a Differentiated View of Counterattitudinal Performance" in Feldman, S. (Ed.) Cognitive Consistency, New York: Academic Press, 1966, pp. 137-172 0 Rosenthal, R. Experimenter Effects in Behavioral Research. New York: Appleton, Century, Crofts, 1966. Scott, W. A. "Attitude Change Through Reward of Verbal Behavior." Journal of Abnormal and Social Psychology, 1957, 55, 72-75. Scott, H. A. "Attitude Change by Re5ponse Reinforcement: Replication and Extension." Sociometry, 1959, 22, 328-335. APPENDIX A OPINION PROFILE QUESTIONNAIRE M0 Opinion Profile Name No. A On the following you are asked to make a number of opinion judgements. Record your first impression. Please mark the item as fast as you can. Do not stop to think it over. Do not go back to change a mark. Here is an item like those you will see on the following pages: U. S. withdrawal from Viet Nam. Wise : : : : : : Foolish Bad : : : : : : Good Your job is to place one checkmark in each seven-point scale directly above the blank that best shows how you feel about the item. If you feel it would be very wise for the U.S. to withdraw from Viet Nam, you might check the second blank from the left. If you feel it is neither wise nor foolish to withdraw from Viet Nam you might check the middle blank which means "neutral" or "neither" or "I'm not sure," All scales Will be marked in the same manner, but the checkmarks will go in different placeSCNtthe scales depending on how strongly you hold your opinions. Remember: Do not place checkmarks anywhere except immediately above the appropriate blanks. Work rapidly. Now turn the page and begin. -.-- no . ,- .1 .41 ‘5 -o- -w'“--._r ...—...... . - c 1 ‘ - ”1“. . “1'. $ _ r . 1 \ . 0" ‘ I. ‘ . r ‘ . ',r- o Making the sale of cigarettes illegal. Bad Valuable Important Foolish Legalization of Marijuana. Bad Valuable Important Foolish More strict gun control laws. Bad Valuable Important Foolish A lottery Bad,_ Valuable Important Foolish a o o O 0 W““_*““ draft system. Good Worthless 'Trivial Wise Good Worthless Trivial Wise Good Worthless Trivial Wise Good Worthless Trivial Wise 'fl 91.. m4.v;x-.m. ”— nor-eo- . r . .. 2 .. ~ ' n I 0 e - - e , a e n .~_Q0’ha. . . . . ..., .. .... .. - ,. . . . . - . ' t . ..-. .-. - . .-. .... . ‘ l . -r I ~ h. ' l - I '. o O I ~ . . .--. ..-. - - --.. ~ . . A . v ’ ~ . a . . . ,-.. .-.. . . A . , ' “:1 ' I - . . . . , . . . ...-» _-....» _.. o» . - . I - . . . ' - -— - _....- V-_.~.. .. ' I ., ‘ , “1' . - . ,: . . - - ‘.( ., I I ‘,I .. .-n- .44. -. -.. .. . - . - - -‘-o'\or—~-'- '” ...—L ..- . . . . ' A y . ' n I... - . . . .-,.-. ... .. r .. . . .... . c ' . I. l ' . a o | . V. -..- . . . . . - . . . APPENDIX B OPINION POLL ON CURRENT TOPICS 145 Opinion Poll on Current Topics Name No. On the following you are asked to make51number of Opinion judgments. Record your first impression. Please mark the item as fast as you can. Do not stop to think it over. Do not go back to change a mark. Here is an item like those you will see on the following pages: U. S. Withdrawl from Vietnam. Wise : : : : : : : Foolish Your job is to place one checkmark in each seven-point scale directly above the blank that best shows how you feel about the item. If you feel it would be very wise for the U.S. to withdraw from Viet Nam, you night check the second blank from the left. If you feel it is neither wise nor foolish to withdraw from Viet Nam you might check the middle blank which means "neutral" or "neither" or "I'm not sure." All scales will be marked in the same manner, but the checkmarks will go in different places on the scales depending on how strongly you hold your opinions. Remember: Do not place checkmarks anywhere except immediately above the appropriate blanks. Work rapidly. Now turn the page and begin. Eliminating college football. Stupid : : : : : : Intelligent A lottery draft system. Good ____:____:____:____:____:____:____: Bad A legal voting age of 18. Unjust : : : : : : Just Eliminating college football. Immature : Mature A lottery draft system. Worthless ____ Valuable A legal voting age of 18. Careful : Careless Eliminating college football. Cruel ___r___: : : : Kind A lottery draft system. Trivial : : ____ Important A legal voting age of 18. Unsure Confident Eliminating college football. Admirable : : : : . : : Con- temptible A lottery draft system. Wise Foolish A legal voting age of 18. Responsible : : : ' : Irrespon- sible ...-.. --. . ---n v‘ __ a . A . '|v I Q ,- o .. , .. ‘ o u... __ _, n o a — . n. 9 a.“ l APPENDIX C OPINION PROFILE QUESTIONNAIRE 48 Qpinion Profile Name No. r On the following you are asked to make a number of opinion judgements. Record your first impression. Please mark the item as fast as you can. Do not stop to think it over. Do not go back to change a mark. Here is an item like those you will see on the following pages: U. S. withdrawal from Viet Nam. Wise : : : : : : Foolish Bad : : : : : : Good Your job is to place one checkmark in each seven-point scale directly above the blank that best shows how you feel about the item. If you feel it would be vegy wise for the U.S. to withdraw from Viet Nam, you might check the second blank from the left. If you feel it is neither wise nor foolish to withdraw from Viet Nam you might check the middle blank which means "neutral" or "neither" or "I'm not sure," All scales will be marked in the same manner, but the checkmarks will go in different placescntthe scales depending on how strongly you hold your opinions. Remember: Do not place checkmarks anywhere except immediately above the appropriate blanks. Work rapidly. Now turn the page and begin. - «o-- . .. .. . .- .. . .-..l l vA- . . I - - u , . .. I ' I . . . C ‘v- a v I~ . V' A ~ . .1 ‘ " mi'v'" . .n c r . e . , . . . ~ . . ~ . . . v. I 9,.1 ~ ' . . 1. '0 ' . . . - a; ,,I 3 . ‘ .I - . c ‘ u ' . z. . . ’_ . .,_, : ... '»' l' -, . . , . 1 '5 Aw. I . r l - ' .. ». ." a n I7 . . I ' ' . ' . ..l v . ' ' '\ . . I'L, " ‘ I, ‘ I . Making the sale of cigarettes illegal. Bad Valuable Important Foolish o O o 0 “*“fl Legalization of Marijuana. Bad Valuable Important Foolish More strict gun control laws. Bad Valuable Important Foolish A lottery draft system. Bad_ Valuable Important Foolish Good Worthless "Trivial Wise Good Worthless Trivial Wise Good Worthless Trivial Wise Good Worthless Trivial Wise APPENDIX D MESSAGE EVALUATION AND QUESTIONNAIRE 51 Now that you have read the arguments and underlined the main points you could find, we are interested in what you thought of the theme. Please make the following judgments about the essay you have just read. 1. In general, how would you rate the logic of the arguments presented? very good quite good fairly good not very good not good at all Hi! 2. In general, how would you rate the amount and quality of information in the essay? very good quite good fairly good not very good not good at all HH 3. In general, how would you rate the clearness with which the essay is written? very good quite good fairly good not very good not good at all HI! 4. In general, how would you rate the overall quality of the message? very good quite good fairly good not very good not good at all HIH . ‘~ , ... -. -. - ' r . - 1 . .. . . > - . e .. _ ...f..n - .2 ..i... I. . .z- .. ‘1 . . . . . . . ._ ' "'. ’ ' .- ~ -. . . .s u , ,, l‘ ‘ . l - r . . . ... - ..‘.. . .. ., ,3“ .. -- . .,,- .. . , . -' - -- -_ -. v . -~ . 'r . ..I-. f O - ‘ '--Dl - . . . .- . ‘ .. . - n. u . L - 5 . . ... c u- . .l 1_ . . v . y t , -.-,' '.l ‘... ., .- ... _ ‘.(1_ ..Y.- _‘ .. .1 ., . , . I .\ ' ,. . . . ‘ z .. -.. .' . . 1 A .. l . . . . ‘L. I ‘ ' v . .-. "l . .. .'. H ’ -‘. .,, .. ,...., -..- _. I a... '.. ls- .. ' l“. . . ., .. ..' _ _ .', ,t. - ‘ V. a a. 6‘7‘ .‘I ' ou-Omll' .. . . v, ‘ w ‘ 1 .1.-!. . ‘ --—v~-- ‘- ‘ I . , " b I ,. w - ,. -.- .. .t ‘ 1'. ... '- Fl! ‘nr 4 I ‘ ' .. I ’ ». ‘7 .. .. .. .- ... .. . . ' 1 , ... ,. . . . . . , -.. . , .. . . .J' ---.-.-~ o—-—- ’ I .» .. ' ..;,_ . _-.—...-~_.- v " 11‘ ; " - mu. -1 I ' u , . . ' I o - 1 . '( , 5. . . ' - 'r - - i" 1.!.-~' .' .. ‘... . ':-5. . . . 4. :-‘ "-. .. 1‘ ‘ V ‘ .l "_ . -~-‘~'l o. ‘ A ‘ e _.--.--....- r _» _ l . ‘- ‘--. .. .‘l' ‘, . .‘ , .. -... , . I A'- ' a O 1 . . . . . ... Opinion Poll on Current Topics Name No. On the following you are asked to makeainumber of opinion judgments. Record your first impressiOn.. Please mark the item as fast as you can. Do not stop to think it over. Do not go back to change a mark. Here is an item like those you will see on the following pages: U. S. Withdrawl from Vietnam. Wise : : : : : : : Foolish Your job is to place one checkmark in each seven—point scale directly above the blank that best shows how you feel about the item. If you feel it would be veyy wise for the U.S. to withdraw from Viet Nam, you might check the second blank from the left. If you feel it is neither wise nor foolish to withdraw from Viet Nam you might check the middle blank which means "neutral" or "neither" or "I'm not sure." All scales will be marked in the same manner, but the checkmarks will go in different places on the scales depending on how strongly you hold your Opinions. Remember: Do not place checkmarks anywhere except immediately above the appropriate blanks. Work rapidly. Now turn the page and begin. Eliminating college football. Stupid : : ° ° . : Intelligent A lottery draft system. Good : : : : : : : Bad A legal voting age of 18. Unjust : : : : : : : Just Eliminating college football. Immature : : : : : : : Mature A lottery draft system. Worthless : : §___: : : : Valuable A legal voting age of 18. Careful : : : : : : : Careless Eliminating college football. Cruel : : : : : : : Kind A lottery draft system. Trivial : : : :___j____y___: Important A legal voting age of 18. Unsure :4VT} : : : : : Confident Eliminating college football. Admirable : : : : : : : Con- temptible A lottery draft system. Wise : : : : : : : Foolish A legal voting age of 18. Responsible : : : : : : : Irrespon- sible ' fi‘AL'T-L. _- q... .- ‘ - I -. . . .>-.A. . ...--. . .lr. . .- -—~‘-u .~— ...”. ‘_-‘I—. .-.-_... .- l...... . . ._ .. .-.. .. . l‘ 4v : . - . I , _.. ...- . . .-.--. ,. .. . . MICHIGAN STATE UNIVERSITY LIBRARIES I I 3 III 3 1293 III I I IIIIII 45 5128 0 1