AN EXPERIMENTAL STUDY ON THE TYPE AND FREQUENCY 0F REFERENCE A5 USED BY AN UNFAMILEAR SOURCE EN A MESAGE AND E75 EFFECT UPON FERCEW’ED (REGINU'FY ANID ATTITUDE CHANGE Thesis few féw Deg-Ne oi: pit. D. fiiCKlGéH STHE UR‘EVERZSITY Terry H. Ostermeier 1965 THESIS 0-169 26% UN “0 LIBRARY Itil‘lllllllllllllllllllllllllll|ll||1||NlllllllNlllWlHllL/ “Chis“ .S'm 31293 10383 4341 Unficrsxty This is to certify that the thesis entitled An Experimental Study on the Type and Frequency of Reference as Used by an Unfamiliar Source in a Message and Its Effect on Perceived Credibility and Attitude Change presented by Terry H. Ostermeier has been accepted towards fulfillment of the requirements for Bh.D_. degree in_§D_e_e.§_h LL ‘1 MM (// {43:45 6/ yéjor professor / Date £181 6, 1966 ABSTRACT AN EXPERIMENTAL STUDY ON THE TYPE AND FREQUENCY OF REFERENCE AS USED BY AN UNFAMILIAR SOURCE IN A MESSAGE AND ITS EFFECT UPON PERCEIVED CREDIBILITY AND ATTITUDE CHANGE by Terry H. Ostermeier The purpose of this study was to analyze the image of a communicator as derived from the use of reference by the communicator during the presentation of his mes— sage. Two types of reference, self and prestige, and four frequencies-—no, low, moderate, and high--were ma— nipulated. The effect of reference was measured in terms of perceived source credibility and attitude change toward the proposal. Reference was defined as the method by which a source of a message provides information about himself during the presentation of a message for the purpose of enhancing his image with the receivers of the message. Self—Reference was defined as a reference in which the source reveals his first—hand experience with the topic. Prestige—Reference was defined as a reference in which the source reveals a personal association with others who have had first-hand experience with the topic. The use of nine references was considered as a high frequency, six references as moderate, three references as low, and \ e F b .2 L. .2 r. C .I : ..... , . .1. T. f .2 r .. rs a. :a f. by .v C ‘ a AC .2 .. LL .fiq Q» r.. e T. .4 e t 3 c. S T. E l .1 E : . C ... .C e C. t :7. l f a e S a S 7 E C a... L. .c n r... C d t E m C -e u .1 ... Cu .2 «e L e t e e e c; m.“ .l p .. l r - e v4 v.1. V PC Q» C a t r S E Ru 5 C .7. .. . e n... u... .1 e C. .3 R. Q e E T. T. up. .i u s. C .e e :L A v n u. r. Q» :L e .aL D. MW .1 ‘Q \Iouttnr. i . Terry H. Ostermeier zero references as no reference. Thus, the messages used in the study were labeled (1) High Self—Reference, (2) Moderate Self-Reference, (3) Low Self—Reference, (4) High Prestige-Reference, (5) Moderate Prestige—Reference, (6) Low Prestige-Reference, and (7) No—Reference. The messages advocated the adoption of the honor system for examinations in college. The problem—solu— tion pattern of organization was followed. The six ex— perimental messages contained the same content as the No—Reference control message with one exception. Both the content and the source of each reference were in— cluded in the experimental messages, while only the con- tent but not the source of each reference was presented in the control message. All seven messages were recorded on tape by the same individual. The only introduction given to the source of the message was that he was a recent college graduate. Two hundred and fifty subjects from the beginning speech course at the State University of New York at Buffalo were randomly assigned as groups to the differ- ent treatments. A semantic differential was employed to measure perceived credibility in an after-only ex— perimental design. A Likert-type attitude questionnaire was used to measure attitude change in a pre—post exper- imental design. The data were analyzed by means of analysis of variance, linear regression analysis, and t- test procedures. I. .l Terry H. Ostermeier The major findings in this study were: 1. Results supported the position that self-refer- ence by a source did generate increased credibility for an unfamiliar source of a message and did result in significant changes in attitude on the part of receivers of the message. The source was perceived as more compet- ent and trustworthy but not more dynamic when he used self-reference. The findings indicated that the greater the frequency of self—reference, the greater the per— ceived credibility and attitude change for an unfamiliar source. 2. Results did not support the position that prestige-reference by a source generated increased credibility for an unfamiliar source or that it resulted in significant attitude change. Competence, dynamism, and attitude change were not significantly affected by increases in prestige-reference. An increase in per— ceived trustworthiness was revealed. 3. Results indicated that type and frequency of reference did not interact in affecting trustworthiness, dynamism, or attitude change. The findings revealed that type and frequency did interact in affecting per— ceived competence. 4. The findings showed that the type of reference did affect the perceived dynamism of the source but did Terry H. Ostermeier not appear to affect the source's perceived competence or trustworthiness or the attitude change by the receivers of the message. An unfamiliar source who used prestige— reference was perceived as less dynamic than a similar source who used self—reference. AN EXPERIMENTAL STUDY ON THE TYPE AND FREQUENCY OF REFERENCE AS USED BY AN UNFAMILIAR SOURCE IN A MESSAGE AND ITS EFFECT UPON PERCEIVED CREDIBILITY AND ATTITUDE CHANGE BY Terry Hs‘Ostermeier A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Speech 1966 to C m4 not“ q H u. .43? :e ‘ Cly . i ACKNOWLEDGMENTS The writer wishes to express his appreciation to Dr. Murray Hewgill for his guidance, encouragement, and constructive criticism in directing this research and to Dr. Kenneth Hance, Dr. Gordon Thomas, and Dr. Clyde Dow for their suggestions and constructive crit- icisms. The writer is grateful for the assistance and cooperation of several individuals at the State Univer— sity of New York at Buffalo. Special recognition is given to: Dr. Ernest Thompson, Jr., Assistant Professor of Speech, who recorded the messages and offered con- structive comments; Mr. Thomas McPhail, Graduate Teach— ing Assistant, for his assistance in the administration of the experiment; Dr. S. David Farr, Director, Educ— ational Research Center, for his suggestions concern— ing the analysis of the data; and Mr. Roger Campbell of the Computer Center for his services in running the statistical analysis of the data. The writer also wishes to extend his thanks to the faculty of the Speech 126 course for willingly relinquishing class time to conduct the experiment and to the AV Communication Center for its COOperation in providing the staff mem- ber and studio facilities to record the messages. ii Special thanks are expressed to my wife, Arlene, for her continuous encouragement, assistance, and help- ful comments in this endeavor. iii ACKNOWL LIST OF TABLE EDGMENTS . . . TABLES. . . . LIST OF FIGURES . . . LIST OF CHAPTER I. II. III, IV. APPENDICES. . INTRODUCTION . OF CONTENTS Relevant Experimental Studies. Scope of This Study: . Definitions. Summary. . . THEORETICAL BACKGROUND Rhetorical Theory. Psychological Theory Summary. . . Major Hypotheses Null Hypotheses. PROCEDURE. . . The Subjects The Messages Measurement. Administration Methods of Analysis. Summary. . . RESULTS . O . Perceived Source Competence. Perceived Source Trustworthiness Perceived Source Dynamism Attitude Change. Summary. . . iv Page ii vi xi xii 12 l4 l7 17 22 35 39 4O 41 41 45 55 6O 64 65 69 69 77 85 89 97 CHAPTER V. DISCUSSION. Final Summary APPENDICES . BIBLIOGRAPHY Page 100 115 123 221 TABLE LIST OF TABLES Page Dimensions, scales, and factor loadings for each scale used in the measuring of source credibility taken from factor analytic research by Berlo and Lemert. . . 56 Results of factor analysis indicating the factor loading, mean, and standard devia- tion for each of the twelve statements comprising the Likert-type attitude ques— tionnaire used to measure attitude change. 59 Factorial analysis of variance for com- petence scores toward the concept "The Speaker You Have Just Heard" concerning two types, self and prestige, and four frequencies--no, low, moderate, high-—of reference with an N of 25 for each of the eight experimental groups. . . . . . . . . 71 Mean ratings for the sums of responses to four seven—interval scales measuring perceived source competence as revealed in a message using either self—reference or prestige—reference with an N of 25 for each experimental group . . . . . . . . . 73 Simple randomized analysis of variance for competence scores toward the concept "The Speaker You Have Just Heard" using self-reference in four frequencies-—no, low, moderate, high——with an N of 25 for each of the four experimental groups . . . 75 Simple randomized analysis of variance for competence scores toward the concept "The Speaker You Have Just Heard" using prestige-reference in four frequencies-- no, low, moderate, high--with an N of 25 for each of the four experimental groups . 75 Linear regression analysis for competence scores toward the concept "The Speaker You Have Just Heard" using self-reference in four frequencies--no, low, moderate, high—— with an N of 25 for each of the four ex— perimental groups. . . . . . . . . . . . . 76 vi TABLE 8. 10- ll. 12. 13. 14. Page Factorial analysis of variance for trust— worthiness scores toward the concept "The Speaker You Have Just Heard" concerning two types, self and prestige, and four fre- quencies--no, low, moderate, high--of reference with an N of 25 for each of the eight experimental groups. . . . . . . . . Mean ratings for the sums of responses to four seven—interval scales measuring per- ceived source trustworthiness as re- vealed in a message using either self- reference or prestige-reference with an N of 25 for each experimental group. . . . . Simple randomized analysis of variance for trustworthiness scores toward the con- cept "The Speaker You Have Just Heard" using self—reference in four frequencies—- no, low, moderate, high--with an N of 25 for each of the four experimental groups . Simple randomized analysis of variance for trustworthiness scores toward the concept "The Speaker You Have Just Heard" using prestige-reference in four fre- quencies——no, low, moderate, high--with an N of 25 for each of the four exper- imental groups. . . . . . . . . . . . . .. Linear regression analysis for trustworth- iness scores toward the concept "The Speaker You Have Just Heard" using self-reference in four frequencies--no, low, moderate, high-—with an N of 25 for each of the four experimental groups . . . . . . . . . . .. Linear regression analysis for trustworth— iness scores toward the concept "The Speaker You Have Just Heard" using prestige-refer- ence in four frequenCies—-no, low, moderate, high-—with an N of 25 for each of the four experimental groups . . . . . . . . . . .. Factorial analysis of variance for dynamism scores toward the concept "The Speaker You Have Just Heard" concerning two types, self and prestige, and four frequencies—— no, low, moderate, high-—of reference with an N of 25 for each of the eight exper- imental groups. . . . . . . . . . . . . . . vii 78 80 83 83 84 86 TABLE 15, 16. 17. 18. 19. 2O 0 21. Page Mean ratings for the sums of responses to four seven-interval scales measuring perceived source dynamism as revealed in a message using either self-reference or prestige-reference with an N of 25 for each experimental group . . . . . . . Factorial analysis of variance for atti- tude scores toward the concept "Honor System for Examinations in College" con- cerning two types, self and prestige, and four frequencies—~no, low, moderate, high—- of reference with an N of 25 for each of the eight experimental groups . . . . . . Mean ratings for the sums of responses to twelve seven—interval scales measuring attitude change for subjects in exper— imental and control groups as revealed in a message using either self—reference or prestige—reference with an N of 25 for each group. . . . . . . . . . . . . . Simple randomized analysis of variance for attitude scores on the concept "Honor System for Examinations in College" us— ing self-reference in a message in four frequencies-—no, low, moderate, high—- with an N of 25 for each of the four experimental groups . . . . . . . . . . . Simple randomized analysis of variance for attitude scores on the concept "Honor System for Examinations in College" us- ing prestige-reference in a message in four frequencies-—no, low, moderate, high-- with an N of 25 for each of the four experimental groups . . . . . . . . . . . Linear regression analysis for attitude— difference scores on the concept, "Honor System for Examinations in College," us- ing self—reference in a message in four frequencies--no, low, moderate, high—- with an N of 25 for each of the four experimental groups . . . . . . . . . . . A comparison of the mean ratings for sums of responses to twelve seven—interval scales measuring attitude change in ex— perimental groups as revealed in a mes— sage using self—reference with an N of 25 for each group. . . . . . . . . . . . . . viii 86 9O 91 94 94 95 96 TABLE 22. 23. 24- 25- 26- 27. 28. 29- 30. 31. A comparison of the attitude—difference scores of subjects in the control groups and experimental groups on the concept "Honor System for Examinations in College" with an N of 25 for each group. . . . . . Data from three dimensions of perceived source credibility and attitude change for subjects who heard the No—Reference Message. . . . . . . . . . . . . . . .. . Data from three dimensions of perceived source credibility and attitude change for subjects who heard the Low Self- Reference Message. . . . . . . . . . .. . Data from three dimensions of perceived source credibility and attitude change for subjects who heard the Moderate Self—Reference Message. . . . . . . . . . Data from three dimensions of perceived source credibility and attitude change for subjects who heard the High Self— Reference Message . . . . . . . . . . . . Attitude change data for subjects who heard no message. . . . . . . . . . . . . Data from three dimensions of perceived source credibility and attitude change for subjects who heard the No—Reference Message. . . . . . . . . . . . . . . . . Data from three dimensions of perceived source credibility and attitude change for subjects who heard the Low Prestige- Reference Message . . . . . . . . . . . . Data from three dimensions of perceived source credibility and attitude change for subjects who heard the Moderate Prestige—Reference Message. . . . . . . . Data from three dimensions of perceived source credibility and attitude change for subjects who heard the High Prestige- Reference Message . . . . . . . . . . . . ix Page 97 207 208 209 210 211 212 213 214 214 TABLE 32. 33. Attitude change data for subjects who heard no message. Results of factor analysis for each of the original forty—four attitude state- ments. . . . Page 216 217 FIGURE LIST OF FIGURES Page Differences in mean ratings of perceived source competence for subjects in the four experimental conditions. . . . . . . 74 Differences in mean ratings of perceived source trustworthiness for subjects in the four experimental conditions. . . . . . . 81 Differences in mean ratings of perceived source dynamism for subjects in the four experimental conditions. . . . .. . . . . 88 Differences in mean attitude-difference scores for subjects in the four experimental conditions. . . . . . . . . . . . . . . . 93 xi Tu e o my. I I Ti E I I P». T‘ APPENDIX I. II. III, A. G. LIST OF APPENDICES ATTITUDE QUESTIONNAIRE FOR SELECTION OF THE TOPIC OF THE MESSAGE. . . . . PRE—TEST ATTITUDE QUESTIONNAIRE FOR EXPERIMENTAL AND CONTROL GROUPS: POST-TEST ATTITUDE QUESTIONNAIRE FOR CONTROL GROUPS. . . . . . .. . . . . POST—TEST ATTITUDE QUESTIONNAIRE FOR EXPERIMENTAL GROUPS . . . . o. . . . SEMANTIC DIFFERENTIAL ADMINISTERED TO EXPERIMENTAL GROUPS TO MEASURE PERCEIVED SOURCE CREDIBILITY . . . . THE INTRODUCTION GIVEN TO THE SOURCE OF THE MESSAGE . . . . . . . . . . . ADMINISTRATIVE PROCEDURE FOLLOWED WITH THE EXPERIMENTAL GROUPS . . . . ADMINISTRATIVE PROCEDURE FOLLOWED WITH THE CONTROL GROUPS. . . . . . . NO—REFERENCE MESSAGE o o o o o o o 0 LOW SELF—REFERENCE MESSAGE . . . . . MODERATE SELF—REFERENCE MESSAGE. . . HIGH SELF-REFERENCE MESSAGE. . . . . LOW PRESTIGE—REFERENCE MESSAGE . . . MODERATE PRESTIGE—REFERENCE MESSAGE. HIGH PRESTIGE-REFERENCE MESSAGE. . . Tables 23-330 0 o o o o o e o o o o o o o xii Page 124 125 130 135 137 138 139 141 150 159 168 178 187 196 206 CHAPTER I INTRODUCTION The importance of the relationship between source credibility and attitude change is obvious upon an analy- sis of the literature. Most of the available experimental evidence indicates that the credibility of a communicator is definitely related to the degree of effectiveness of the message. In particular, such research seems to show a relationship between the prestige or reputation of the communicator and attitude change by the receivers of the message. Relevant Experimental Studies Experimental research on the broadly defined con- cept of source credibility is quite extensive. A large segment of the research, however, is not directly rele- vant to the purpose of the present study. Several credi— bility studies have related a source to a proposition with no message involved.l Most of these studies have 1Such as Claude E. Arnett, Helen H. Davidson, and Hallett N. Lewis, "Prestige as a Factor in Attitude Changes," Sociology and Social Research, XVI (September—October 1931), 49-55; Mitchell Saadi and Paul Farnsworth, "The Degrees of Acceptance of Dogmatic Statements and Preferences for Their Supposed Makers," Journal of Abnormal and Social Psychology XXIX (July—September 19345, l43—lSO;Irving Lorge with Carl C. Curtiss, "Prestige, Suggestion and Attitudes," Journal of Social Psychology, VII (November 1936), 386-402; Helen B. Lewis, "Studies in the Principles of Judgments and At- tidues, IV: The Operation of 'Prestige Suggestion'," 1 2 dealt with variables quite different from those prOposed in this study. The communicator-message-audience relation- ship was not the basis for analysis in these studies. Another group of studies tends to be more closely related to this experimental study. Haiman presented the same tape—recorded message to three groups of subjects while attributing it to three different sources--the Sur- geon General of the United States, the Secretary of the Communist Party,and a university sophomore. Results in- dicated that the high credibility of a source is signifi— cantly effective in changing attitudes and does show significant differences in ratings of competence when compared with a source having low credibilityfl' Strother employed approximately the same techniques and found results consistent with Haiman's.2 Paulson utilized one Journal of Social Psychology, XIV (August 1941), 229- 256; and Malcom Moos and Bertram Koslin, "Prestige Sug— gestion and Political Leadership," Public Opinion Quart- erly, XVI (Spring 1952), 77—93. lFranklyn Haiman, "An Experimental Study of the Effects of Ethos in Public Speaking," (unpublished Ph.D. dissertation, Northwestern University, 1948). 2Edward Strother, "An Experimental Study of Ethos as Related to the Introduction in the Persuasive Speak- ing Situation" (unpublished Ph.D. dissertation, North- western University, 1951). 3 tape—recorded message which was attributed to a political science professor and a student. Male members of the audience showed more attitude change for the message when it was attributed to the professor than when the l The preceding same message was ascribed to the student. studies were concerned with differences in credibility which contrasted quite sharply. Hovland and Mandell in a study attempted a less obvious effort to modify the communicators credibility. The results of their study indicated that more individuals changed their opinions in the direction of the source's position in the impartial version than in the suspicion-arousing version, but not at an adequate level of significance. The audience did rate the impartial source as the more honest of the two.2 The studies previously cited centered on manipul— ating the credibility of the communicator prior to the delivery of the message. Another group of studies was concerned with the influence of source credibility both prior to, and during, the presentation of the message. For example, Kersten presented one tape—recorded message with two introductions of the communicator being the lStanley Paulson, "Experimental Study of Spoken Communication: The Effects of Prestige of the Speaker and Acknowledgement of Opposing Arguments on Audience Retention and Shiftcf Opinion" (unpublished Ph.D. dis— sertation, University of Minnesota, 1952). 2Carl I. Hovland and Walter Mandell, "An Exper- imental Comparison of Conclusion Drawing by the Communic— ator and the Audience," Journal of Abnormal and Social Psychology, XLVII (July 1952), 581-588. 4 manipulated variable. One introduction used techniques to enhance the source's credibility, while the other introduction did not. The favorable introduction re- sulted in significantly greater attitude changej' Pross employed essentially the same format and found no significant differences.2 Andersen and Clevenger state that an explanation for the conflicting results may be that ”neither Kersten nor Pross actually measured differences in ethos; they assumed that different in- troductions would affect that variable."3 The attitude scales in these studies measured attitude change toward the message but not the credibility of the source of the message. Bryson employed similar procedures in determining the effectiveness in persuasive spaking of the traditional speech of advocacy and the denotative speech under varying conditions of source credibility and subject of the speech. Sources were selected to represent high, medium, and low levels of prestige. JBarbara Kersten, "An Experimental Study to Deter- mine the Effect of a Speechcf Introduction Upon the Persuasive Speech That Followed" (unpublished thesis, South Dakota State University, 1958). a\\ Edward Pross, "A Critical Analysis of Certain Aspects of Ethical Proof" (unpublished Ph.D. dissertation, State University of Iowa, 1942). iienneth Andersen and Theodore Clevenger, Jr., "A Summary of Experimental Research in Ethos," Speech Monographs, XXX (June 1963), 70. .«c . L. :3 r” «3 .3 I. r,“ C. AA r 1.. F I . , . r. l . t. c w. _ . . “L .C (C «r . ..: .1‘. Av A... .r“ s. Ck A. LL 6 . . e r. ~M .«q 2‘ E .7. , , fi - IPA. «Fm .V1‘ . =9.» Pu .au A: “ ms ...1. r“ U AH» V; .m« L any a». w“ At) 7 ‘ 7 A .h» .- A ‘1. C. €711 r. _ . Hi I 1‘ C.» 5 Findings indicated that the credibility of the speaker, type of speech, and subject of the speech are not signific— ant in themselves} Andersen constructed three introductions de- signed to establish varying levels of prestige and authoritativeness for speakers dealing with the farm problem. His conclusions were these: (1) Students perceived significant differences between a college student and a Professor of Ag— riculture or a Farm Extension Agent on two scales: (a) the evaluathme and the dynamism dimensions of a semantic differential designed to measure ethos; (b) authoritativeness as estimated by a Likert—type scale . . . . (4) There was no proof that the variations in ethoszand authoritativeness affected persuasive— ness. . . . Andersen used two tape-recorded speeches, bothcf whom were attributed to three sources described in tape-recorded introductions. The principal results were these: (1) Despite great manuscript variations which speech experts predicted would produce different levels of ethos, the only significant differences between the two speeches were those measured on the dy- namism scale . . . . (3) The variations in ethos did not cause a significant difference in per- suasiveness. Using introductions to messages manifesting differ— ent levels of credibility, Hewgill and Miller found the high credibility source was rated significantly more 1Kenneth D. Bryson, "An Experimental Studycf the Effectiveness of the 'Denotative' Speech in Persuasion" (unpublished Ph.D. dissertation, Northwestern Univer— sity, 1952). Kenneth Andersen, "An Experimental Study of the Interaction of Artistic and Non-Artistic Ethos in Per- suasion" (unpublished Ph.D. dissertation, University of Wisconsin, 1961), as reported in Andersen and Clevenger, Jr., p. 70. 3 Ibid., p. 74. 6 competent, and on the dimensions of trustworthiness and dynamism all differences were in the expected direction. More importantly, it was indicated that a high fear message, when presented by a high credibility source, results in more favorable attitudes toward the message. Data for the low credibility source revealed no signific- ant differences on attitude.l Projecting further into the message itself, add- itional studies have varied the use of quotations by authorities in the message. Gilkinson, Paulson, and Sikkink presented two versions of the same message. One version used quotations by authorities, while the second did not. Results revealed no significant differences, with a slight trend toward a greater shift in attitude in the message employing quotations by authorities. In another study, Sikkink similarly used quotations, but no significant difference was found in attitude change.3 While it would seem that the use of authority 1 Murray A. Hewgill and Gerald R. Miller, "Source Credibility and Response to Fear—Arousing Communications,” Speech Monographs, XXXII (June 1965), 95—101. 2Howard Gilkinson, Stanley Paulson, and Donald Sikkink, "Effects of Order and Authority in an Argu— mentative Speech,” Quarterly Journal of Speech, XL (April 1954), 183-192. 3 Donald Sikkink, "An Experimental Study of the Effects on the Listener of Anti—Climax Order and Authority in an Argumentative Speech,"fflNNflKflfil Speech Journal, XXII (Winter 1956), 73—78. ""-V' ‘IG Du >r‘ u...» n "" pp '4: fl- 7 does not alter the credibility of a communicator, it must be emphasized that source credibility was not the criti- cal variable which was analyzed in these studies. It must be stressed that once again the audience was not asked to assess the credibility of the source of the message. Thus, it would seem that any conclusions on the concept of credibility from these studies would be premature. The studies discussed up to this point have dealt with the influence of source credibility prior to the actual delivery of the message. The following studies are concerned with the effects of credibility during the presentation of a message. That changes in ethos result from hearing speeches seems clear from a study of the effect of a campaign speech by Thomas E. Dewey. Com- paring ratings obtained before a speech with those recorded immediately afterwards, Thompson found that students raised their estimation of Dewey as a public speaker but did not change their opinions significantly concerning the soundness of his ideas and his acceptability as a cand- idate. 1 Ludlum constructed a speech in which he in- corporated several elements designed to increase the credibility of the source. His techniques include the acknowledgment of opposing arguments, '1eading thoughts rather than forcing,‘ show— ing alleged facts to be consistent with known lWayne Thompson, "A Study of the Attitude of Col- lege Students Toward Thomas E. Dewey Before and After Hearing Him Speak," Speech Monographs, XVI (August 1949), 125-134, as reported in Andersen and Clevenger, Jr., p. 70. 8 facts, showing material to be recent, and man- ifesting a 'high degree of credibility' by means of self-praising statements. Comparing the persuasiveness of this speech with that of a 'straight argumentative' address, he found the latter to be more effective. 1 The results of the Ludlum study must be considered with some reservations on the implications for the concept of source credibility. Since he did not measure received ethos, the ef- fect of the variables in the nonargumentative speech is unknown. Moreover, since all of the variables were incorporated in a single speech, it is impossible to isolate the effect of any one of them. If some of the techniques produced positive effects and others acted negatively, the effects may have counterbalanced one another. 2 Since the measurement of source credibility is of vital concern in experimental research, it becomes im— portant to analyze the contribution of studies on the measurement of this concept. Literature has progressed to the point of defining source credibility in terms of dimensions. Hovland, among others, has referred to the dimensions of perceived truthworthiness and perceived expertness. 3 Berlo and Lemert have prOposed the lThomas Ludlum, "A Study of Techniques for In— fluencing the Credibility of a Communicator" (unpublished Ph.D. dissertation, Ohio State University, 1956), as re- ported in Andersen and Clevenger, Jr., pp. 71-72. Andersen and Clevenger, Jr., p. 72. Carl I. Hovland, Irving L. Janis, and Harold H. Kelley, Communication and Persuasion (New Haven: Yale University Press, 1953), pp. 19—36. 9 dimensions of perceived competence and perceived dynam- ism. These authors have indicated through exploratory research that these dimensions, in addition to per- ceived trustworthiness, might be used as an index of source credibility.l The study by Andersen supports the point that differences in credibility can be measured by a dynamism scale of a semantic differential.2 (The study by Hewgill and Miller has demonstrated that dif— ferences in source credibility can be derived from competence, trustworthiness, and dynamism scales of a semantic differential. Sc0pe of This Study While choosing to employ the term ethos in their summary of experimental research in that area, Andersen and Clevenger point out that the number of quantitative studies using the term ethos in their titles is small. Studies using such terms as credibility and prestige involve a greater quantity and variety of research. David K. Berlo and James B. Lemert, ”An Em- pirical Test of a General Construct of Credibility," (unpublished paper presented at the Speech Associa- tion of America Convention, New York, December 29, 1961) Andersen. 3 Hewgill and Miller. 10 Studies concerning this concept have been conducted in written, non—verbal, and oral communication., Such studies have investigated the influence of credibility on the ef— fect of a communication, techniques for generating or changing credibility, and the measurement of credibility.1 This study was limited to an analysis of the com- municator—message—receiver relationship. It was a study in oral communication. Furthermore, it was narrowed in its primary purpose to an investigation into a technique for generating or changing source credibility. The ef— fect of source credibility on the acceptance of a message and the measurement of credibility were of related in— terest. One of the primary ways in which credibility has been studied is through the use of two sets of sources, one with high credibility and the other with low credibil— ity, to determine the difference in persuasive effects. Such experimental studies have manipulated source credi— bility prior to the presentation of the message. Research thus far has largely left unexamined the numerous message elements introduced by the communicator during the pre— sentation of his message. These message elements which are said by speech experts to establish source credibility 1 Andersen and Clevenger, Jr., p. 59. 11 include, for example, reference by the communicator to previous relationships with his audience, references to previous personal experience with the subject of the message, the use of recent information, and the use of testimony by authorities. Previous experimental studies have analyzed several message elements in one unified message. This procedure prevented the isolation of the effects of any one message element. Likewise, the ef- fect of the frequency with which message elements are employed in a message remains unexplored. Thus, the ef— fect of message elements on the perceived credibility of the communicator and upon attitude change has remained basically unclear. It has been noted that one of the means which a source employs to convey his own credibility to an aud- ience is through reference to himself and to his exper- ience with the tOpic of the message. This study dealt specifically with the relationship of reference by the communicator during the presentation of his message and its effect on source credibility and attitude change. The specific purpose of this study was to analyze the effects of both type and frequency of reference by a communicator. Two types of reference, self—reference and prestige—reference, and four frequencies of refer— ence—— no, low, moderate, and high-- were chosen for analysis. *' 1“- 12 Definitions Certain concepts need to be defined as a basis for understanding the development of the study. Source Credibility Reference Self-Reference Prestige-Reference Competence The receiver's image of a source of a message at a given time. Dimensions which compose credi- bility are commonly referred to as competence or expertness, trustworthiness, and dynamism. A type of information provided by the source of a message about himself during the pre- sentation of his message for the purpose of creating or en— hancing his image with the re- ceivers of the message. 1m . A type of information provided by the source of a message about himself which takes the form of revealing and stress— ing his first—hand experience with the topic of the message. A type of information provided by the source of a message about himself which takes the form of revealing and stress- ing his personal association with others who have had first- hand experience with the topic of the message. One of the dimensions of credi— bility as derived from the descriptive word scales on a semantic differential. Such descriptive words include, for example, whether the source is experienced—inexperienced, ex- pert-ignorant, important-unim— portant, and informed-uninformed. The score on the dimension of competence as measured on a semantic differential. 13 Trustworthiness One of the dimensions of credi- bility as derived from the descriptive word scales on a semantic differential. Such descriptive words include, for example, whether the source is just—unjust, honest-dishonest, reasonable-unreasonable, and Openminded—closedminded. The score on the dimension of trust— worthiness as measured on a semantic differential. L l Dynamism One of the dimensions of credi— bility as derived from the descriptive word scales on a semantic differential. Such . descriptive words include, for 1} example, whether the source is ‘g aggressive-meek, bold-timid, extroverted-introverted, and active—passive. The score on the dimension of dynamism as measured on a semantic differ- ential. Attitude Change The degree of listener opinion change on the tOpic of the mes- sage as measured by a Likert— type attitude questionnaire. Such terms as source, communicator, speaker; message, com— munication, speech; and receiver, listener, audience are to be found throughout the account of this study. This practice should not be interpreted as an indication that tjmatermsnecessarily have identical meanings. For pur— poses of clarity, these terms were referred to inter- changeably in this study. Qualifications have been noted wherever differences among these terms are critical. 14 Summary Prior research seems to indicate that the credi— bility of a communicator is related to the degree of effectiveness of the message. In particular, there seems to be a relationship between the prestige or reputation of the communicator and attitude change by receivers of the message. Sources with high credibility influence greater attitude change than sources with low credibility. The procedure of increasing the amount of support for an assertion in the form of quotations by authorities tends to increase the perceived credibility of the communicator. It is less certain whether or not this procedure, in turn, influences greater attitude change. Studies show that speeches of introduction influence the listener's conception of the source of a message. Turning to concepts specifically related to the scope of this study, it can be seen that studies explor- ing the techniques employed by the communicator during the presentation of his message to gain increased credi- bility have been limited. Several message elements which are said to induce credibility have been incorporated in one message. This procedure has made it difficult to isolate the effect of any one of the message elements. In addition, the credibility of the source often has not been measured in these studies. It must further be concluded that thefrequency with which a communicator 15 attempts to generate or increase his own credibility during the message has remained unexplored. Certain dimensions of credibility have been ad— vanced in studies concerned with the measurement of source credibility. Experimental studies have demon- strated the ability of such dimensions as perceived com- petence, trustworthiness, and dynamism in the form of a L. semantic differential to measure differences in source credibility. i This study was limited to an analysis of the image I of a source of a message as derived from the use of reference by the communicator during the presentation of his message. Two types of reference, self and prest— ige, and four frequencies of reference——no, low, moderate, and high——were manipulated. The effect of themanipulation of these variables was measured in terms of changes in perceived source credibility and attitude. In addition to the preceding limitations, definitive restrictions were presented regarding the key concepts under study. The next chapter contains an examination of rhetorical and psychological theories related to the concept of source credibility and a statement of the hy- potheses to be investigated in this study. Chapter III is an account of the procedure of the study which includes 16 a discussion of the subjects, messages, methods of measure- ment, administration, and methods of analysis. Chapter IV contains a description of the results of this study while Chapter V consistscf a discussion of the results and a final summary. An appendix and a bibliography fol— low Chapter V. E. In: CHAPTER II THEORETICAL BACKGROUND Rhetorical Theory References to the concept of credibility appear in writings even prior to the time of Aristotle; how— ever, emphasis on the concept does not seem to be very substantial until the writings of Aristotle. Aristotle discusses rhetoric in terms of two kinds of proofs or persuasions, artistic and non-artistic. The former are the forms of persuasion applied through the communicator's own efforts while the latter are not. "Of the means of persuasion supplied by the speech itself, there are three kinds. The first kind reside in the character [ethos] of the speaker; . . . . the most potent of all the means to persuasion."1 Two points seem worthy of emphasis. First of all, Aristotle places a definite de-emphasis on what might be termed the reputation of the speaker. Succeeding rhetoricians, as will be seen, tend to place more emphasis on this aspect of source credibility. Secondly, the importance placed on the effect of a communicator's credibility as a means of persuasion by Aristotle is definitely clear. Aristotle lAristotle, The Rhetoric of Aristotle, trans. Lane Cooper (New York: Appleton-Century—Crofts, Inc., 1932), pp. 8—90 17 :_ 7"‘1- - r n" 18 considers source credibility to be the most vital of the means of persuasion. Book Two of The Rhetoric provides an insight into Aristotle's views concerning the composition of credibil- ity. As for the speakers themselves, the sources of our trust in them are three, for apart from the arguments [in a speech] there are three things that gain our belief, namely intelligence, character, and good will.. . .It necessarily follows that the speaker who is thought to have all these qual- itiesfintelligence, character, and good will] has the confidence of his hearers. l Aristotle believed the means by which the source of a mes- sage gains these qualities are to be found in the choices he makes in the presentation of his message. "The speaker's ethos arises from choices evident in the invention, ar- rangement, style, and delivery Of the speech."2 This latter statement summarizes Aristotle's conception of the development Of sources of credibility. Cicero did not stress credibility to the same de— gree as did Aristotle. In his treatment of credibility, Cicero tended to discuss the concept more in terms of the reputation of the communicator. "Now feelings are lIbido, pp. 91-92. William Sattler, "Conceptions of Ethos in Rhe— toric" (unpublished Ph.D. dissertation, Northwestern University, 1941), p. 92. 1‘.‘ 19 won over by a man's merits, achievements or reputable life, qualifications easier to embellish, if only they . l are real, than to fabricate where non-eXistent." Quintilian generally held the Aristotelian view of credibility but again with the exception that he tended to stress the role of reputation. My aim, then, is the education of the perfect orator. The first essential for such a one is that he should be a good man, and consequently we demand of him not merely the possession of exceptional gifts of speech, but all the excel— lence of character as well. 2 It has been said that, following the time of Quin— tilian, credibility as a concept in the process of com— munication was treated with less importance. The ages from Quintilian to the Twentieth Century were ones in which ethos lost a place in the models of the rhetorical process; then, despite individual variations and excepting writers in the homilect— ical tradition, regained a place in rhetoric as ex- tensive as that which it held in ancient rhetor— ical systems. 3 Turning to some of the representative writing of rhetorical theory in the twentieth century, sources such as texts on speech criticism and persuasion are available. 1 Marcus Tullius Cicero, De Oratore, trans. E. W. Sutton and Horace Rackham (Cambridge: Harvard University Press, 1948), II, 327. 2 Marcus Fabius Quintilian, The Institutio Oratoria of Quintilian, trans. H. E. Butler (Cambridge: Harvard University Press, reprinted 1963), I, Book I, 9-11. Andersen, p. 22. 2O Thonssen and Baird indicate that a source of a message may give credibility to his message in a variety of ways. In general, a speaker focuses attention upon the probity of his character if he (1) associates either himself or his message with what is virtuous and elevated; (2) bestows, with propriety, tem— pered praise upon himself, his client, and his cause; (3) links the opponent or the opponent's cause with what is not virtuous; (4) removes or minimizes unfavorable impressions of himself or his cause previously established by his opponent; (5) relies upon authority derived from his personal experience; and (6) creates the impression of being completely sincere in his undertaking. With certain qualifications varying with the circumstances, it may be said that a speaker helps to establish the impression of sagacity if he (1) uses what is popularly called common sense; (2) acts with tact and moderation; (3) displays a sense of good taste; (4) reveals a broad familiarity with the interests of the day; and (5) shows through the way in which he handles speech materials that he is possessed of intellectual integrity and wisdom. Finally, a speaker's good will generally is revealed through his ability (1) to capture the proper balance between too much and too little praise of his audience; (2) to identify himself properly with the hearers and their problems; (3) to proceed with candor and straightforwardness; (4) to offer necessary rebukes with tact and consideration; (5) to offset any personal rea— sons he may have for giving the speech; and (6) to reveal, without guile or exhibitionism, his personable qualities as a messenger of the truth. 1 One notes in the preceding discussion of credibility an emphasis on the elements within the presentation of the 1Lester Thonssen and A. Craig Baird, Speech Criticism (New York: The Ronald Press Company, 1948), p. 387. 21 message that affect source credibility as well as the im— pact of the source's reputation. It is significant for the purposes of this study that reference to personal experience by the source of a message is indicated as one of the means to achieve credibility. Writers of persuasion texts reveal another insight into contemporary rhetorical views on source credibility. Oliver speaks of the communicator's personality, and examines credibility in terms of what the source of a mes— sage represents: his qualities, his attitudes, and his relationship to the audience.1 Brembeck and Howell hold that this concept may be broadly defined as character. In the context of persuasive speech it includes two ele— ments; the reputation or prestige enjoyed by the speaker with respect to his particular audience and subject at the moment he begins to speak, and the increasing or dimin- ishing of that prestige as a result of what he says and does during the speech.2 Abernathy notes that ethical appeal stems from three areas: the speaker's reputation, his personality, and his identification with theaudience. 1Robert T. Oliver, The Psycholggy of Persuasive S eech (2d ed. rev.; New York: David McKay Company, Inc., 19575,‘ pp. 69-79. 2 . . Winston L. Brembeck and William S. Howell, Persuasion - A Means of Social Control (New York: Prentice— Hall, Inc., 1952), p. 244. 22 Thus, the persuasive speaker's ethos is brought into the picture. Even before we have seen the man we know of his reputation. If it is one of integrity, courage, and competency, we tend to regard him with respect. If we know also that he is a careful and a prudent man, a just per— son, and normally consistent in his ideas, he will begin with a solid reputation . . . . While he speaks we further enlarge the picture. If he is both self—confident and modest, en— thusiastic, and friendly to the point of genial— ity, he has measurably helped his cause. 1 In addition to stressing the role of reputation, Abernathy emphasizes the influence of message elements on the source's credibility. As he speaks we also observe his efforts at audience identification. If he is tactful and courteous, objective, and with a sense of humor we tend to approve his attitude. If he refers to common experiences, we find ourselves closely bound to him . . . . Thus die speaker uses his self to further his message. The preceding discussion once again points out the import— ance placed by contemporary rhetorical theory on elements within the presentation of a message as sources of credi- bility. Psychological Theory Theory resulting from studies other than the area of rhetoric is more difficult to isolate. 1 Elton Abernathy, The Advocate - A Manual of Persuasion (New York: David McKay Company, Inc., 1964), p. 187. 2 Ibid. 'I‘ 23 It is important to remember that these stud— ies, which arise from such fields as psychology, speech, sociology, and education, are quite diverse in origin, that many of the experimenters did not use rhetorical terminology, and that many of them did not perceive a relationship be- tween their studies and ethos. l A selective analysis can be attempted as long as a recog- nized relationship to source credibility is inherent in the theory being considered. The concept of prestige or reputation has been stressed in the field of psychology. Saadi and Farnsworth discuss the degree of acceptance of dogmatic statements made by well—liked or disliked persons. According to these authors, readiness to accept dogmatic statements appears to some extent to vary with the likeableness of the person to whom the statements are attributed.2 Kulp notes the relationship of prestige to the premanency of changes in attiplde due to a single—experience situation. He indicates that different authorities have different degrees of prestige, that the same authorities have dif— ferent degrees of prestige under similar conditions, that there are different degrees of prestige under varying conditions, and that the prestige factor is not constant for all individuals in all situations. ..o-—~r~.. 1Andersen and Clevenger, Jr., p. 59. 2Saadi and Farnsworth. 3 Daniel H. Kulp, II, "Prestige; as Measured by Single-Experience Changes and Their Permanency,” Journal of Educational Researcp, XXVII (May 1934), 663-672. 'I}~_ :~ 24 According to Sherif, stereotype or prestige is a crucial element of the communication situation. He asks the question: Are the words of a person with whom we hold as a prestige factor viewed in the whole context of our relationship to that person? Sherif replies that stereo- type or prestige plays a considerable role in people's judgments. Attitudes toward the authors of statements serve as reference points in judging the actual state— ments.1 Among others, Haiman, Strother, Paulson, Bernberg, and Cole through experimentd_studies have tended to sup— port the position that the prestige or reputation of a source is a vital factor in changing opinions.2 Hovland, Janis, and Kelley in their psychological studies of opinion change discuss source credibility in a slightly different manner when they speak of the ex- pertness and trustworthiness of a source. lMuzafer Sherif, "An Experimental Study of Stereo— types," Journal of Abnormal and Social Psychology, XXIX (January-March 1935), 371-375. Haiman; Strother; Paulson; Raymond E. Bernberg, "Prestige Suggestion in Art as Communication," Journal of Social Psychology, XXXVIII (August 1953), 23—30; and David Cole,'“Rational Argument' and 'Prestige—Sug— gestion' as Factors Influencing Judgment,” Sociometry XVII (November 1954), 350-354. (n ()L I I’I’( ) 25 When acceptance is sought by using arguments in support of the advocated view, the perceived ex— pertness and trustworthiness of the communicator may determine the credence given them. An individual's tendency to accept a con— clusion advocated by a given communicator will depend in part upon how well informed and intel— ligent he believes the communicator to be. How— ever, a recipient may believe that a communicator is capable of transmitting valid statements, but still be inclined to reject the communication if he suspects the communicator is motivated to make nonvalid assertions. It seems necessary, there- fore, to make a distinction between 1) the ex- tent to which a communicator is perceived to be a source of valid assertions (his "expertness") and 2) the degree of confidence in the communicator's intent to communicate Uneassertions he considers most valid (his "trustworthiness"). In any given case, the weight given a communicator's assertions by his audience will depend upon both of these factors, and the resultant value can be referred to as the "credibility" of the communicator. 1 A look at the conclusions of several studies pre- sented in the Hovland volume presents an insight into the concept of credibility as viewed in the framework of psychological theory. Ewingf2 Hovland, Lumsdaine, and 3 4 Sheffield, Hovland and Weiss, Hovland and Mandell? lHovland, Janis, and Kelley, pp. 20—21. 2 . . Thomas N. Ewing, "A Study of Certain Factors In— volved in Changes of Opinion," Journal of Social Psychology, XVI (August 1942), 63-88. Carl I. Hovland, Arthur A. Lumsdaine, and Fred D. Sheffield, Experiments on Mass Communication: Vol. III of Studies in Social Psychology in World War II (Prince- ton: Princeton University Press, 1949). Carl I. Hovland and Walter Weiss, "The Influence of Source Credibility on Communication Effectiveness," Public Opinion Quarterly, XV (Winter 1951), 635-650. 5 Hovland and Mandell. 26 and Kelman and Hovlandl have investigated the factors of expertness and trustworthiness by means of experimental studies. In summary, the research evidence indicates that the reactions to a communication are signific— antly affected by cues as to the communicator's intentions, expertness, and trustworthiness. The very same presentation tends to be judged more favorably when made by a communicator of high credibilitythan one of low credibility. . . . #. From the results, it is not possible to disen— tangle the effects of the two main components of credibility--truthworthiness and expertness-—but it appears that both are important variables.2 Recent research by such individuals as Kerrick,3 L q Brehm and Lipsher,4 and Powell“ tends to substantiate the relationship of the credibility of the communicator and audience response. Herbert C. Kelman and Carl 1. Hovland, "'Reinstate- ment' 01 the Communicator in Delayed Measurement of Opinion Change," Journal of Abnormal and Social Psychology, XLVIII (July 1953), 327-335. 2 Hovland, Janis, and Kelley, p. 35. 3 Jean Kerrick, "The Effect of Relevant and Non- Relevant Sources on Attitude Change," Journal of Social Psychology,XLVII (February 1958), 15-20. Jack W. Brehm and David Lipsher, "Communicator— Communicatee Discrepancy and Perceived Communicator Trustworthiness," Journal of Personality, XXVII (September 1959), 352-361. Frederic A. Powell, "Open and Closed Mindedness and the Ability to Differentiate Source and Message," Journal of Abnormal and Social Psychology, LXV (July l962),61—64. 27 A further elaboration on the relationship between source credibility and attitude change may be found within the framework of balance theory. The term balance theory is used here to include several theories which have been advanced using such terms as dissonance, congruity, and balance. All such theories have at their heart the premise that the individual normally strives for consist- ency within himself, and than [sic] when incon— sistency or imbalance or incongruity or disson- ance occurs, tensions are produced which motivate the individual not only to resolve or reduce the dissonant state, but also to avoid those cognit— ions which are likely to increase dissonance. The object, of course, is to thereby achieve con— sonance, or a harmonious and balanced cognitive state. The amount of dissonance associated with a given cognition is a function of both the num- ber and/or importance to the individual of the competing c0gnitions; by the same token, the individual may reduce dissonance by decreasing the number and/or importance of the relevant cog— nitive elements consonant with his existing or committed behavior. Furthermore, the greater the dissonance, the greater the pressures to re— duce the dissonance. The theory of cognitive dissonance is one of the so—called ”equilibrium" theories which have ap— peared from time to time to explain and predict behavior.. .. Leon Festinger (1957), the origi- nator of the term, refers to "dissonance" as the existing of nonfitting, or inconsistent, re— lations among cognitions. By "cognition" is meant any knowledge, Opinion, or belief about the environment, one's self, or one's behavior. Cog- nitive dissonance, therefore, results when two elements of cognition are in obverse relation— ship to each other. That is, if X implies Y, and I believe in both X and the obverse of Y, then I - experience a motivational tension called dissonance.C 1'Robert S. Goyer, "Cognitive Dissonance and Com— munication Theory," Central States Speech Journal, XV (May 1964), 91—92. 2 Ibid., p. 91. 28 Osgood and Tannenbaum have chosen to use the term "congruity" to explain the interaction effects of credib— ility as manifested in the source and the message. The general congruity principle may be stated as fol- lows: Whenever two signs are related by an asser— tion, the mediating reaction characteristic of each shifts toward congruence with that characteristic of the other, the magnitude of the shift being inversely proportional to intensities of the in- teracting reactions. This "shift," obviously, may be in intensity, direction or both. Thus, if sign A elicits an intensely favorable evaluative component, and sign B only a slightly unfavorable evaluative component, the compromise will be such that the reaction to A will be only slightly modified, but that to B relatively more modified.1 The congruity theory differs from other consistency theories in that it measures both the direction and intensity of the elements composing the situation. Attitudes can be specified as favorable (+), neu— tral (O), and unfavorable (—). Assertions can be specified as positive or associative (+) or negative or dissociative (-). They may also, of course, include evaluative loading (e.g. when X denounces Y, we have both a dissociative as— sertion and a negative evaluation of Y). When at— titudes toward both objects of judgments are polar, the nature of the assertion determines congruence or incongruence.2 Osgood and Tannenbaum assign scale values to the intensity of the elements, for example, —3, -2, or +1. 1Charles E. Osgood, George J. Suci, and Percy H. Tannenbaum, The Measurement of Meaning (Urbana: University of Illinois Press, 1957), pp. 200—201. ’2 ‘ Charles E. Osgood and Percy H. Tannenbaum, "The Principle of Congruity in the Prediction of Attitude Change," Psychological Review, LXII (January 1955), 44. 29 Tannenbaum tested the congruity principle in a written communication situation. The accuracy in the pre— diction of direction of shift for source and concept in terms of an assertion was highly significant. 1 Berlo and Gulley employed similar procedures in an oral com— munication situation and found significant changes in attitude consistent with the prediction of the congruity principle. 2 Burdick and Burnes obtained a galvanic measure of skin resistance from subjects in balanced and imbalanced conditions. Their findings, in general, support the premise that imbalance states between source and concept cause greater tension and affect at— titude change.3 'Bettinghaus studied five variables believed responsible for attitude change in an oral communication situation. The following statements in— dicate the conclusions of the study: 1. The experimental subjects did tend to shift their attitudes toward the Speaker and the speech topic to more congruous positions. 2. The shift toward congruity in the oral communication situation seems to be determined lPercy H. Tannenbaum, "Attitude Toward Source and Concept as Factors in Attitude Change Through Communica— tion" (unpublished Ph.D. dissertation, University of Illinois, 1953). 2David Berlo and Halbert Gulley, "Some Deter— minants of the Effect of Oral Communication in Produc- ing Attitude Change and Learning," Speech Monographs, XXIV (March 1957), 10-20. 3Harry Burdick and Alan Burnes, "A Test of 'Strain Toward Symmetry' Theories," Journal of Abnormal and Social Psychology, LVII (November 1958), 367—370. GIG U '1, C~ : 1|. ... C . . u L— . . .8 r . .ru it . c .7... T. .ML K .L .3 .C , A A: wu . . 1 a ; L. C. a. a. a... .. . .Q . NC. . u to .5 no .r1 vi 9. . . QC 6 .v ”1; .. x. «C J. .wJ a F; C» we. Q Q.“ r . In} .r.. .1!“ .nL , A v {M rL Cu .3 .ti‘. ~r v «xu. .. m .UL r11. .c 5- . HI). In... 30 more by the listener's attitude toward the speaker than by the listener's attitude to— ward the speech topic. Heider, using the term balance theory, offers the basis of an approach which is particularly related to this present study. Heider assumes that a state of balance exists if entities which belong together are all positive or all negative. If two of the entities are of a dif— ferent sign, a state of imbalance results. Heider des— cribes a state of balance in terms of sentiment relation— ships and unit formations. Sentiment relationship is the term he uses for attitude. A unit formation can exist between such entities as a source and an attitude object. If the relationship is positive, it is called a unit. If the relationship is negative, no unit exists. For example, a labor union leader speaking in favor of the union movement would be viewed as a positive unit relationship. On the other hand, a labor union leader speaking against the union movement would be considered as a negative relationship. Balance theory is explained in the manner in which the different relations among the entities interact. These relations are classified according to such con- cepts as similarity, interaction, proximity, familiarity, and ownership. The concept of familiarity is of par- ticular relevance to this study. Balance theory 1 . . . Erwin Bettinghaus, Jr., "The Operation of Congru- ity in the Oral Communication Situation," Speech Monographs XXVIII (August 1961), 142. 31 hypothesizes that imbalance results in a situation in which one of the entities is unfamiliar. The negative effects of unfamiliarity may be as— sumed to occur as a result of at least two fact- ors. First, an unfamiliar situation is full of possibilities that may be sufficiently threaten— ing to an insecure person to turn him against it. An unfamiliar situation is cognitively un— structured; that is, the sequences of steps neces- sary to reach an object is not clearly known. . .- In addition, there is a second factor which has little to do with insecurity or danger. It is a more purely intellectual and aesthetic com— ponent of the resistance to the unfamiliar. The strange is experienced as not fitting the struc— ture of the matrix of the life space, as not fit— ting one's expectations. The adaptation or change in expectations which is required by meeting the unfamiliar demands energy. For example, if source X is unfamiliar to receiver Y and X asserts something about attitude object Z, an imbalanced situation would exist. The preceding conception dealance theory fits in with the rationale for this study. 2 It should be evident that a message typically tells the receiver some— thing not only about the subjectcf the message but about the sourcecf the message as well. The response a receiver Inakes upon hearing the presentation of a message actually lFritz Heider, The Psychologyyof Interpersonal Ibelations (New York: John Wiley & Sons, Inc., 1958), 2I am briefly describing the communication situa- titni as found in Heider, The Psychology of Interpersonal Ihelations and Fritz Heider, "Attitudes and Cognitive Efiaganization," Journal of Psychology, XXI (January 1946), 107—112. 32 involves at least two responses, one toward the subject and one toward the source. The typical communication situation includes a source who is attempting to change a preconceived attitude of a receiver toward the subject of a message. The preconceived attitude of the receiver is based on some previous experience with the subject. The message causes the receiver to make a response to- ward the subject of the message and the source of the mes— sage. This response is in addition to the previous re— sponse the receiver had toward the subject of the message. .i :‘T Balance theory suggests that a receiver holds both a direct attitude, one currently held by the receiver, and a mediated attitude, one derived from a source other than the receiver. In the communication situation, the source of the message is the means of mediation. The source presents a mediated attitude toward the subject of the message. His purpose is to have the receiver superimpose it over his own direct attitude, thus, altering or changing his direct attitude. A necessary condition for a receiver to accept the mediated attitude and, therefore, chang- ing his direct attitude in his familiarity with the source of the message. It would seem that a receiver's knowledge of the source of a message may be of varying kinds and degrees. For example, it is conceivable the receiver may find him- self in a communication situation where no information is available about the source of a message. In this 33 situation, it is possible the receiver may begin to re- ceive the message with no information about the communic- ator. In another communication situation, the receiver may be presented with information concerning the source which could be termed negative in nature. This negative information could take the form of showing that the source has limited qualifications on the subject of the message. For example, the source may be described as a high school student who has looked up information on the subject of the message for a classroom assignment. Negative in— formation could take the form of placing the source in a derogatory light. In this case, the source may be revealed as a member of the Communist Party. In both of these instances, the receiver is presented with nega- tive information about the communicator. In still a third communication situation, a receiver may be pre- sented with so-called positive information about the com— municator. The source may be revealed to be a college professor with extensive experience with the subject of the message. In this case, positive information about the communicator is made available to the receiver of the message. Balance theory has indicated that cognitive im- balance is created in situations where negative inform- ation is presented about the sourcecfi a message. Reject— ion Of the proposal and lowering the perceived credibility 34 of the communicator have been suggested as the most ef- fective means to remove or reduce the cognitive imbalance. Studies concerned with the effects of prior reputation on perceived credibility and attitude change generally lend support to this conclusion. Balance theory further indicates that imbalance is created in situations where no information is revealed about the communicator. The source of the message, in other words, is unfamiliar to the receiver. Once again, rejection of the proposal and lowering the perceived credibility of the source are suggested as the most ef- fective means to remove or reduce the imbalance state that has been created. On the other hand, accepting the proposal and increasing the perceived credibility of the communicator, would be more likely in situations where positive information concerning the source of the message is made available to the receiver. Furthermore, it would seem that as the positive information about the source increases, the attitude change and perceived credibility should increase. The preceding generaliza- tions form the basis for the rationale from which the hypotheses prOposed in this study are formulated. The hypotheses are specifically delineated at the end of this chapter. 1Such as Hewgill and Miller; Andersen. 35 Summary Writers of rhetorical theory have been aware of the concept of credibility as an important part in the communication process dating in time before Aristotle. Aristotle was among the first to stress the importance of source credibility and to analyze its component parts. Cicero and Quintilian basically held to the Aristotelian viewpoint but did not appear to stress the overall con— cept as much as Aristotle. In addition, they tended to perceive credibility as essentially reputation. The place of source credibility in the model of com— munication seemed to be de-emphasized from the time of Quintilian until the twentieth century. In the twentieth century, credibility as seen in texts in speech criticism and persuasion, is discussed more extensively once again. Elements both prior to and during the presentation of a message are stressed as contributing to the credibility of a source and his message. The writings of Aristotle, Thonssen and Baird, Brembeck and Howell, and Abernathy are of particular relevance concerning the theoretical framework for this study. While they do present their viewpoints within a Inore general framework, all of these authors emphasize the role of elements within a message in affecting the credibility of a source and the persuasiveness of his Inessage. They do indicate that one of the ways in which ,a source can achieve or increase his credibility is to Inaveal his own personal experience with the subject of 'fi.’ . - ..‘- - A w}! ‘. g... -7 --i I ”Ar]~€n '- vaonv ~ ' y... K." . “,‘.~r"" w. v0 f 5..." " . hfirr“ ,v..\.e [A s ‘. ‘ye -r- ...e V Fir. VbA“ FAQ- ..\.‘ .fiu EU a C 36 the message. With this general approach, rhetorical theory offers a rationale for the basis of this study. The fields of psychology, sociology, speech, and education have contributed to an understanding of the concept of source credibility within a framework of psychological theory. The lack of rhetorical terminol— ogy or a theoretical framework directly related to the concept of source credibility has tended to limit the relevant theory. Two aspects of credibility have been stressed in the literature. It has been generally held that the prestige or reputation of a source of a message influences the receiver's reception of the message. Secondly, the manner in which a receiver perceives the expertness or trustworthiness of a source seems to affect reactions not only to the communication but to the source as well. The contribution of psychological theory on these two general aspects of credibility tends to support what lias been advanced for sane time in rhetorical theory. More recently the concept of source credibility lias been analyzed within the framework of so-called con— :sistency theories. These theories, developed under such rumnes as balance, congruity, and dissonance, are all lmased on the concept of equilibrium. They are based (Hi the assumption that man normally strives to be con— :aistent in his behavior. When inconsistency occurs, truisions are produced which encourages man to remove, .49 “hQ E. 37 reduce, or avoid what is causing the inconsistency. Some aspects of the source—concept relationship pre- dicted by consistency theory have been subjected to ex- perimental research. Results ofstudies by Tannenbaum, Berlo and Gulley, Burdick and Burnes, and Bettinghaus generally substantiate the predictive value of con— sistency theory in the source—concept communication situation. Heider uses the term balance theory to explain the relationships among the entities in a communication situation. The relationship of such entities as source and concept may be classified according to several con- cepts, however, the concept of familiarity is particul— arly relevant to this study. Balance theory hypothes- izes that imbalance results in a situation in which one of the entities is unfamiliar. An unbalanced state would exist if a source, unfamiliar to the receiver, made an assertion on the tOpic of a message. One of the primary means of reducing or removing imbalance would be to make the source familiar to the receiver. A second point of importance in balance theory is the stress it places on the fact that the message itself tells the receiver something about the source as well as something about the subject of the message. A typical communication situation could include a situation in which the audience has no prior knowledge 38 about the communicator. In this case, there would be no speech of introduction of the source by another person in which the qualifications and background of the source would be stated. In the same communication situation, the source may reveal positive information about him- self during the presentation of his message. This method could be described as the use of reference by a communicator and, of course, could take place in dif— ferent frequencies. Given a communication situation in which receivers of a message have no information about the communicator prior to the presentation of the message, the following is hypothesized within the frame- work of balance theory. In such a communication situa— tion, if the receivers are presented with positive in- formation about the source of the message, one would expect the receivers to show a higher estimation of the perceived credibility of the source and greater at— titude change on the subject of the message, than in a similar communication situation in which the receivers are presented with no information about the source. Furthermore, given similar communication situations in which receivers of a message are presented with increas- ing amounts of positive information about the communica- tor during the message, one would expect the receivers to show an increasingly higher estimation of the .n.‘ 1.3"' ‘ 39 perceived credibility of the source and an increasing degree of attitude change on the subject of the message. This rationale leads to the following hypotheses proposed in this study. 1) 2) 3) 4) 5) 6) Major Hypotheses Increasing the frequency of self—reference by a source of a message will effect: a) greater perceived competence for that source. b) greater perceived trustworthiness for that source. c) greater perceived dynamism for that source. Increasing the frequency of prestige-reference by a source of a message will effect: a) greater perceived competence for that source. b) greater perceived trustworthiness for that source. c) greater perceived dynamism for that source. The increase in perceived competence for a source of a message is directly proportional to the increase in the use of: a) self-reference by the source, that is, the trend is linear. b) prestige—reference by the source, that is, the trend is linear. The increase in perceived trustworthiness for a source of a message is directly proportional to the increase in the use of: a) self-reference by the source, that is, the trend is linear. b) prestige—reference by the source, that is, the trend is linear. The increase in perceived dynamism for a source of a message is directly proportional to the increase in the use of: a) self—reference by the source,that is, the trend is linear. b) prestige—reference by the source,that is, the trend is linear. Increasing the frequency of self—reference by a source of a message will effect greater attitude change. 1) 2) 3) 4) 40 Increasing the frequency of prestige—reference by a source of a message will effect greater attitude change. Increases in attitude change are directly proportional to the a) b) increases in the use of: self-reference by the source, that is, the trend is linear. prestige—reference by the source, that is, the trend is linear. Null Hypotheses Frequency and type of reference do not interact in affecting: a) b) c) d) the perceived competence of the source of a message. the perceived trustworthiness of the source of a message. the perceived dynamism of the source of a mes- sage. attitude change. Frequency of reference by a source of a message does not affect: a) b) c) d) the source's perceived competence. the source's perceived trustworthiness. the source's perceived dynamism. attitude change. Type of reference by a source of a message does not affect: a) b) c) d) Single the source's perceived competence. the source's perceived trustworthiness. the source's perceived dynamism. attitude change. messages do not result in significant changes in attitude. '1.- v. CHAPTER III PROCEDURE An abbreviated pilot study was conducted employ— ing the procedures proposed for this present study. On the basis of an observation of the administration of the pilot study, slight modifications were made in instruc— tions to the subjects and in the composition of the mes- sages. The study was conducted as planned since no major changes seemed necessary. The Subjects Subjects were selected from Speech 126, the begin- ning speech course at the State University of New York at Buffalo. The catalog description of the course reads as follows: Speech 126,126R EFFECTIVE SPEECH (3) Prerequisite: English 101 or its equivalent. Beginning speech course to improve the student's effectiveness in oral communica— tion. 1 First semester freshmen were ineligible for the course, as can be seen in the above course description. The 1State University of New York at Buffalo Bulletin— University College, 1965/1966, p. 135. 41 42 student composition of the course, therefore, was es- sentially sophomores. Students in this classification normally comprise over 60 percent of the enrollment in the course. The remainder of the enrollment was com— posed of equal numbers of seniors, juniors, and second semester freshmen. The course was multi-sectioned with several instructors in charge. Each instructor was responsible for instruction and recitation for his own sections of students; however, a common syllabus was followed by all sections. Like many studies which have dealt with human subjects, ideal sampling procedures were modified ac- cording to the availability of subjects.1 It was not possible to assign randomly each subject in Speech 126 to a particular treatment group. Students never as— sembled as one group during the semester since there was no mass lecture. It was possible to assign randomly groups of students to the different treatments. During the first semester of the 1965—1966 school year, there were eighteen sections of Speech 126 to draw upon for this study. Twelve of these sections met during a Monday-Wednesday—Friday sequence, while six sections were scheduled for a Tuesday—Thursday-Saturday or Tues- day-Thursday sequence. The experimental session was run 1E. F. Lindquist, Design and Analysis of Experi— ments in Psychology and Education (Boston: Houghton Mifflin Company, 1953), p. 73. 43 on back—to—back days. The fact that there was one day difference in the administration of the total experiment would not appear to have biased any group since no ex- perimental or control group contained only Monday or only Tuesday subjects. Sections of Speech 126 students were randomly assigned to form ten groups. Two sections each were assigned to each of the following treatments: High Self-Reference, Moderate Self-Reference, Low Self- Reference, High Prestige—Reference, Moderate Prestige- Reference, and Low Prestige-Reference. Since the en- rollment was higher in two sections, three sections of subjects were assigned to the No—Reference treatment. Three additional sections were assigned to the No Mes— sage treatment. Subjects in the No-Reference sections were then individually randomly assigned to form two groups. In the same manner, subjects in the No Message sections were assigned to two groups. A total of ten groups of subjects was involved in the study. There was no reason to suspect differences among the groups due to the above procedure. In spite of the fact that the course is multi—sectioned, the names of the instructors for the sections were not known by the students prior to or during registration. The student first learns the identity of his instructor on the first day of class. Given this occurrence plus the fact that 44 enrollment was decided by the student and his advisor alone, it would appear that a student's selection of a particular section indicated nothing more than scheduling convenience. The conclusion of no suspected initial dif- ferences among the sections was further substantiated by the results of the pre-test on attitude. The ten groups were found to have similar attitude mean scores of 44.0, 44.3, 44.6, 45.3, 45.5,45.8, 45.8, 46.4, 46.7, and 47.0 based on a scale of twelve to eighty—four. Two hundred and ninety—one subjects in the eight experimental groups and the two control groups particip- ated in the pre—test on attitude. A copy of the pre—test attitude questionnaire may be found in Appendix IB. Ten days later, immediately following the presentation of the particular message for each experimental group, the post—test attitude questionnaire was administered. Subjects in the two control groups completed the same attitude questionnaire without being exposed to a mes- sage. The smallest N was twenty—five in one group; therefore, subjects were eliminated randomly from each group until an N of twenty-five was achieved. In other words, 250 subjects in the ten groups comprised the N for the study. 45 The Messages It was decided that the tOpic of the message should be of current interest in order to be salient to the subjects hearing the message. This procedure would seem to place the subjects in a situation in which they would be more likely to respond to the topic rather than merely to the experimental setting. A second criterion for the selection of a tOpic was that opinion on it should not be one-sided. The rationale for this decision was that selectionof a tOpic on which Opinion was evenly divided would more likely avoid a marked ceiling effect in audience response. Such an occurrence would have made interpretation of the responses less meaningful. A list of possible topics for a message was pre- pared. The topic, "Honor System for Examinations in College," was selected from this list. An attitude questionnaire was administered to subjects similar to those who were to be involved in the study to test whether the tOpic was one-sided. The attitude ques- tionnaire consisted of the honor system topic plus four other tOpics originally considered for the message. Each tOpic was phrased in the form of a statement fol— lowed by a Likert—type attitude scale. A COpy of the attitude questionnaire is found in Appendix IA. The questionnaire was administered to fifty—five subjects in four sections of Speech 126, none of whom was r- -o- q ..‘.,4 U 5. or e.» C 4.. -..\.. “‘“Lr‘ ~ K \‘upx.-g - . 46 involved in the present study. Results of the ques- tionnaire showed the following breakdown Of responses on the tOpic of the honor system: agree,38 percent; neutral 16 percent; and disagree, 46 percent. With the topic Of the message determined,the next step was researching the literature for material on the tOpic. Books, periodicals, pamphlets, and news- papers were examined. Notations were made of specific arguments on both sides of the issue as well as the means Of supporting the arguments, namely, factual in- formation and Opinion. Any material which could be termed "reference by the source of the message to some relationship he has had with the tOpic" was recorded. In addition, any indications Of possible patterns of organization for the message were considered. Based on a survey of the literature written on the tOpic of the honor system, the following criteria for the composition Of the control message were devised. The problem—solution pattern of organization was selected. The survey had indicated the problem area was "college cheating." The solution phase was emphasized in the literature,while less was written on the problem phase since it was accepted to sane degree by both proponents and Opponents Of the honor system. A ratio Of approx— imately one-third of the message pertaining to the problem phase and about two—thirds to the solution 47 phase seemed consistent with the results of the analy— sis Of the literature. It was decided to have the prob— lem phase consist Of an introduction and definition Of the topic plus a limitation of the problem. The solution phase consisted Of: (1) setting up criteria for a solution to the problem, (2) comparing the two basic solutions, the honor system and the faculty—proctor system, in terms of the criteria, (3) showing how the honor system solution meets the problem, (4) illustrat- ing the advantages Of the prOposed solution, (5) consid- ering the primary objections to the proposed solution, and (6) concluding with a short summary and an appeal for the adOption of the prOposed solution. Statistics, examples, and Opinion were used in support Of the central idea in the content of the mes— sage. Testimony attributed tO well—known experts was not employed because few experts known to the audience were cited in recent sources of information. The guideline for the length of the control mes— sage was set at the ten to fifteen minute range since most credibility studies have used this time interval. The experimental messages were composed of the same content as the control message plus self—reference or prestige—reference in different frequencies. It must be emphasized that the reference itself, not the con- tent Of the reference, was manipulated in the study. 48 The control message contained the content of the refer- ence but nO mention of the source of the reference. Thus, the experimental messages were slightly longer than the control message. The stimuli chosen were messages stressing the ability of the honor system to cope with the problem Of cheating on examinations in college. Since both the type and frequency of reference by the source of a mes— sage were manipulated, a total Of seven messages was involved. The messages were labeled: (1) High Self- Reference, (2) Moderate Self-Reference, (3) Low Self— Reference, (4) High Prestige—Reference, (5) Moderate Prestige—Reference, (6) Low Prestige—Reference, and (7) NO—Reference—-the control message. Self-Reference was defined as a reference in which the communicator reveals and stresses his first-hand experience with the subject during the presentation of the message. Prest— ige-Reference was defined as a reference in which the communicator reveals and stresses a personal associa— tion with others who have had first—hand experience with the subject during the presentation of the message. NO-Reference was defined as a situation in which the communicator makes no reference or mention Of any con— nection he may have had with the subject during the presentation of the message. In other words, the first type was concerned with references the communicator 49 makes to himself regarding his own experiences with the topic. The second type pertained tO references the com— municator makes to individuals he personally knows re- garding their experience with the subject of the message. Both of these types Of references may be classified as data provided by the communicator about himself, a practice said to generate or increase the ‘ credibility of a source. 1 An investigation Of past relevant experimental theory did not reveal any specific rationale for the frequency with which a communicator employs reference in a message. It was decided to investigate a sample of persuasive messages in an attempt tO discover a rationale. Thirty persusasive messages found in fllflf ning Orations — 1961 were examined in an effort to set up criteria. This avenue appeared to offer a basis for comparison since these messages were persuasive in nature, approximately ten to thirteen minutes in length, delivered by college students, and presented to college audiences. An analysis of the messages showed that at least thirteen contained examples of reference by the communicator. Of these, seven messages contained one to three references, three messages had five to seven references, and three messages contained nine or ten references. Most messages employing one or two refer- ences located the references in the beginning one—third of the message. As the frequency Of references 50 increased, they appeared to be located throughout the entire message. It must be noted that there were ex— ceptions to the above two generalizations. The general— izations on frequency and location of reference in a message seemed logical. It would seem that, if a communicator planned to make only a limited number of references during the presentation of his message, he would make them early in his message to establish credibility with his audience at the onset. On the other hand, if the communicator intended to make several references to his own first—hand experience with the tOpic, it would appear likely he would place them throughout his message to remind the audience contin- ually Of his credibility. The preceding sample of persuasive messages re- vealed, for the most part, a frequency range for refer— ences of zero to ten. An analysis Of the placement of the references indicated that the references were spread throughout the entire message when more than one or two were employed. In addition to the frequency and placement of references, a third criterion was neces- sary. The increasing frequency of references was placed at equal intervals for purposes of control and rigor in analysis. A frequency design of zero, three, six, and nine was selected. This design included the minimum- maximum frequency range, the placement Of references, 51 and the element Of an equal interval. For stylistic purposes, the zero frequency was labeled nO—reference; the three frequency, low reference; the six frequency, moderate reference; and the nine frequency, high refer- ence. Since the study dealt with two types of reference, the seven messages were labeled: NO-Reference, Low Self— Reference, Moderate Self—Reference, High Self-Refer- ence, High Self—Reference, Low Prestige-Reference, Mod— erate Prestige—Reference, and High Prestige—Reference. As was stated previously, all seven messages con— tained the same content with the exception of differ— ences in type and frequency of reference. The Low Self- Reference message contained three references by the communicator tO first—hand experiences with the honor system. The three self—references increased the length of the message by about twenty seconds over the NO—Ref— erence control message. The Moderate Self—Reference message, in addition to the same three self—references found in the Low Self—Reference message, contained three more self—references for a total of six. The Moderate Self-Reference message was approximately one minute longer than the control message. The High Self—Refer- ence message was composed of the same six self—refer- ences as the Moderate Self—Reference message plus three additional ones for a total of nine. The High Self— 52 Reference message was approximately one minute and fif- teen seconds longer than the control message. The same procedure was followed for the three frequencies Of prestige—reference. The approximate length Of the three messages beyond the control message was as follows: Low Prestige-Reference, forty seconds; Moderate Prest— ige-Reference, one minute; and HighPrestige—Reference, one minute and forty—five seconds. Thus, the greatest difference in length among the seven messages was one minute and forty—five seconds. The NO—Reference control message contained the material basic to all seven messages plus the content Of the references but no mention of the source of the references. In other words, the NO— Reference message included the experiences with the tOpic Of the message found in the other six messages. The references were not attributed to the communicator nor to individuals personally known to the communicator nor, in fact, to anyone. The experiences were written into the NO-Reference message in what might be called a neutral form. They were phrased as experiences that probably have taken place under the honor system some— where at sometime. It must be emphasized that the references used at each comparable frequency in the self—reference and prestige—reference messages were as similar as possible with the exception Of the source of the reference. Manuscripts Of the seven messages 53 with the references underlined are found in Appendix II. The may ber the following examples taken directly from the messages help to clarify the preceding discussion. The num- in each of the headings indicates the order in which reference occurred in the text of the appropriate message. Self—Reference l As a former undergraduate college student, I per— sonally studied under both the honor system and the faculty-proctor system. During this speech I will draw upon some Of my first-hand exper— iences which I believe are pertinent to the point I am making inthe speech. Prestige-Reference 1 A personal friend Of my family's, Professor William Canning Of Stanford University, has taught for twenty years at five different colleges, under both the honor system and the faculty—proctor system. During this speech, I will draw upon some Of his first—hand experiences as well as tge first—hand experiences Of others I have per- sonally known. NO-Reference l It is likely there are several individuals who have experienced under both the honor and faculty— proctor systems that will be presented in this speech. Self-Reference 6 Thinking back on my own experiences under the honor system, I am reminded of how the faculty, in the beginning, didn't think the students would do much better than faculty supervision; but now, after four years, the faculty holds the opinion that, though matters are not per- fect, they have improved greatly. FEW-4A." 7' 54 Prestige—Reference 6 A former professor of mine, Dean Carl H. Fine, who is now at the University of Michigan, told me that the faculty, in the beginning,didn't think the students would do much better than the faculty supervision; but, after four years, the faculty holds the Opinion that, though mat— ters are not perfect, they have improved greatly. NO—Reference 6 It is possible the faculty at first may doubt whether the students can do much better than themselves; but, after awhile they will likely say that, though there are problems, the situa— tion has improved. Self—Reference 8 However, I hope my first—hand experiences with both the proctor system and the honor system, to which I have referred during this speech, have shown you that the honor system not only pgp but does work. Prestige-Reference 8 However, I hope the first-hand experiences with the proctor system and honor system related during thhsspeech by individuals I have person— ally known, have shown that the honor system not only gap but does work. No-Reference 8 It can be seen, however, that the honor system not only can but does work. The only introduction given to the communicator for all seven messages was that he was a recent college graduate. A COpy of the introduction is found in Ap- pendix IE. Introductory information about the source was deliberately kept at a minimum. This procedure was consistent with the specific purpose Of the study, which was to measure the influence Of credibility due to Inaterial injected during the presentation Of the message. 55 The label "recent college graduate" was chosen because previous studies have shown such individuals do not have high initial credibility. Thus, any increase of credi- bility could be related to the manipulation of reference by the communicator during the presentation of the mes- sage. All seven messages were recorded on tape by the A“ same communicator, a professor of speech who had exper— ience recording messages. The messages were recorded in a studio under the direction of a staff member of the AV Communication Center at the State University of New York at Buffalo. An attempt was made to keep all phases of the recordings constant. An analysis Of the finished tapes by this researcher revealed, in his Opinion, no apparent recording differences. Measurement The Semantic Differential Four seven—interval scales for each Of the three dimensions Of source credibility were employed.1 The concept, "The Speaker You Have Just Heard," was measured cmu a semantic differential consisting Of descriptive word scales. Each subject's score was Obtained by summing Berlo and Lemert. 56 across the four scales used to measure each Of three dimensions of perceived credibility. For example, the perceived competence of the communicator was the sum- mation Of each subject's scores from the scales on the dimension Of competence. The same procedure was fol— lowed for perceived trustworthiness and dynamism. Table 1 contains the dimensions and scales Of credi— bility and the factor loadings for each scale. Table l.——Dimensions, scales, and factor loadings for each scale used in the measuring of source credibility taken from factor analytic research by Berlo and Lemert Dimension Scale Factor Loading Competence Experienced—inexperienced 9O Expert-ignorant 9O Important—unimportant 87 Informed—uninformed 8O Trustworthiness Just—unjust 82 Honest—dishonest 75 Reasonable—unreasonable 74 Openminded—closedminded 73 Dynamism Aggressive-meek 73 Bold—timid 72 Extroverted—introverted 64 Active-passive 61 1 Ibid. 57 The Attitude Scale In the process of researching the tOpic of the message, a list Of unfavorable and favorable statements on the honor system was obtained with the ultimate goal Of forming a Likert—type attitude questionnaire. The original list contained a total Of seventy—three state- ments. Each Of the statements was compared to the criteria set forth by Wang.1 The list was reduced tO forty—four attitude items using this procedure. Of thisknumber, twenty—two statement favored the honor system while twenty—two Opposed the honor system. The forty—four attitude statements were administered in the form of a Likert—type questionnaire to a group of subjects similar to those who were involved in the study. The attitude statements were reproduced in random order and distributed to students in seven sect— ions of Speech 126. The subjects were asked to respond to each attitude statement on a five-interval scale, which consisted Of strongly agree, agree, neutral, dis- agree, and strongly disagree. The attitude question— naire was completed by 110 subjects which was an ac— ceptable N. The results Of the attitude questionnaire 1 Charles K. A. Wang, "Suggested Criteria for Writ— ing Attitude Statements," Journal Of Social Psychology, III (February 1932), 367—373. 2 William J. Goode and Paul K. Hatt, Methods in Social Research (New York: McGraw-Hill Book Company, Inc., 1952), p. 274. an. ‘r-fi‘. . RY‘ .. Vfi‘v s.“ A- ‘» 58 were subjected to factor analysis. The results of the analysis are given in Appendix III, Table 33. The selection of the statements for the final at— titude questionnaire was based on the following set of criteria: (1) a high factor loading on the first factor, (2) the "cleanness" Of the item on the first factor, (3) a mean and a standard deviation which would allow for shifts in attitude, and (4) the relevance of the at— titude statement to the material covered in the context of the message. A set Of twelve statements was derived utilizing this procedure. Table 2 contains the twelve statements with the factor loading, mean, and standard deviation of each. Each subject's attitude toward the tOpic of the message was the summation Of his responses on the twelve statements. The difference between a subject's score on the pre-test and the post—test was used as an index of attitude change. A seven—interval Likert—type scale consisting of strongly agree, agree, slightly agree, neutral or don't know, slightly disagree, disagree, and strongly disagree was employed for the questionnaire. The seven—interval scale was chosen because it would allow a finer indication of the degree of attitude change. In addition, it would be similar in structure to the seven—interval semantic differential which was used to measure perceived credibility. This similarity would appear to aid in the reduction of any confusion in ' qr. in 59 the subject's responses to the two scales. While a five- interval scale has been most common, a smaller and a 1 larger number of scales have been used by researchers. Table 2.--Resu1ts of factor analysis indicating the factor loading, mean, and standard deviation for each Of the twelve statements comprising the Likert—type attitude questionnaire used to measure attitude change Factor Standard Statement Loading Mean Deviation The honor system fails to deal with the factors which encour— age cheating. .66 2.74 0.90 The honor system is based On un- sound educational principles. .61 2.58 1.03 The faculty-proctor system works satisfactorily in preventing cheating. .64 2.27 0.75 The honor system encourages in students a sense Of individual responsibility. .58 2.55 0.88 Competition for grades in col— lege requires the use Of the proctor system. .71 3.00 1.02 The honor system helps the dis- honest student and penalizes the honest student. .67 2.61 0.87 The honorsystem makes a signifi— cant contribution to'the college education Of the student. .68 2.73 1.14 The honor system is essential for a strong university educational program. .75 2.67 0.99 The honor system is another un- warranted attempt to teach mor— ality in the university. .65 2.90 1.09 The honor system is fine in theory but unworkable in practice.78 2.74 1.02 The honor system does reduce the amount of cheating. .52 2.51 0.98 The honor system, as a solution to cheating, is superior to all others. .78 2.26 0.96 lClaire Selltiz et al.,Research Methods in Social Relations (2d ed. rev.; New York: Henry Holt and Company, Inc., 1960), p. 366. 60 Administration Two hundred and fifty subjects were randomly as— signed as groups Of subjects to eight experimental and two control groups. The attitude questionnaire was administered to the subjects to determine their pre- message attitude toward the tOpic ten days before the session involving the presentation Of the message. Each of the instructors in charge Of sections of Speech 126 was briefed the previous week on the general nature Of the study and the attitude questionnaire. Each group Of subjects was informed only that the rationale for the questionnaire was to determine student reaction toward a controversial tOpic for a speech. They were told that this was being done for the purpose of setting up a scale to measure attitudes toward the honor sys- tem for examinations in college. In other words, in— structors were asked to pass out the questionnaires, read page one of the instructions to the subjects, and collect the questionnaires afterthey had been com— pleted. The instructors were cautioned not to deviate from this procedure, especially not tO say anything additional to the subjects. Appendix IB, contains a copy of the pre—test attitude questionnaire. A graduate student in speech and this researcher administered the second part Of the study. In this way, an attempt was made to hold conditions constant in the administration Of the key part of the study. A 61 session was held to discuss the general nature Of the attitude questionnaire, the semantic differential, and the procedure for the presentation of the messages. Mimeographed COpies Of all Of the materials were made available to allow the two administrators to become familiar with them. The ten treatment groups were divided into two sets in an attempt to control bias caused by the choice of administrator. Set one contained High Self-Reference, Moderate Prestige-Reference, Low Self—Reference, one No—Reference group, and one NO Mes- sage control group. Set two was composed Of High Prestige—Reference, Moderate Self-Reference, Low Prest— ige—Reference, one No—Reference group, and one NO Mes— sage control group. The sets were assigned randomly tO the administrators, with chance determining which administrator received which set. In the same manner, no administrator received all Of the self-reference nor all of the prestige-reference groups. It was felt this procedure negated the influence, if any existed, of the administrator on the results Of the experiment. The step—by-step procedure in the administration of the experiment was mimeographed and copies given in advance to the administrators. An evaluation session by the two administrators upon completion Of the ex— periment uncovered no apparent differences in adminis- tration which would seemingly influence the results. 62 Appendix IF, contains a copy Of the procedure followed with the experimental groups, while Appendix IG, shows a copy of the procedure used with the control groups. Ten days after completion of the pre—test attitude questionnaire each group Of subjects was asked to co- operate in an attempt to investigate one of the aspects of effective communication. The subjects were not told the specific purpose of the study. They were in— formed that they would be asked to listen to a speech. It was stated that at the conclusion of the speech they would be asked to indicate their reactions by means of a questionnaire. It was emphasized that they would not be tested on the content of the speech, nor would their reactions in any way affect their grade in Speech 126. The only information presented to the subjects about the communicator prior to hearing the message was that he was a recent college undergraduate who had graduated from a university. A copy of the introductory material read to the subjects about the communicator is found in Appendix IE. Each of the eight experimental groups listened to one Of seven verions of the tape—recorded message. It must be noted that two groups heard the No-Reference rmessage since this treatment was used as a control in an after—only design to determine perceived source credibility. A five-page booklet was passed out to 63 each of the subjects immediately following the presenta— tion Of the message. The contents of the booklet were divided into two parts. Part I, pages two and three, was the identical seven—interval Likert—type attitude questionnaire the subjects had completed in the pre— test. The attitude questionnaire, consisting Of twelve statements, was preceded by one page Of instructions. Appendix IC, contains a copy Of the post—test attitude questionnaire administered to the subjects in the eight experimental groups. The summation of each subject's responses to the twelve statements became the post- message index of his attitude toward the topic of the message. Part II, page five, was the semantic dif- ferential consisting Of descriptive word scales on the dimensions of perceived competence, trustworthiness, and dynamism. The semantic differential scale was preceded by one page of instructions,which contained an explanation of the use Of the semantic differential and requested the subject to react to "The Speaker You Have Just Heard." A copy of the semantic differential administered to the eight experimental groups can be found in Appen— dix ID. A score for each subject, a summation of his responses to each Of the four seven-interval scales, was computed on each of the three dimensions of credi— bility. While a pre—post design was employed to deter- mine attitude change, an after-only design was used to measure differences in perceived source credibility. 64 The subjects in the two control groups, without listening tO a message, responded to the identical at- titude questionnaire that they had completed in the pre- test. The attitude questionnaire used in both the pre- test and post-test for the control groups is found in Appendix IB. Methods of Analysis Bartlett's test for homogeneity of variance was applied tO the data from each of the three dimensions Of perceived source credibility for the eight exper— imental groups.1 The eight experimental groups were analyzed by means of factorial analysis of variance procedures.2 In addition, the test for homogeneity of variance was applied separately to the data for per— ceived credibility from the four experimental conditions of self—reference and the four experimental conditions Of prestige—reference. For further analysis, each of the preceding groups Of four treatments was analyzed by means Of simple randomized analysis of variance and linear regression analysis procedures.3 Finally, the critical difference technique was applied to those analyses which yielded significant F ratios. lLindquist, pp. 87-88. 2 Ibid., pp. 207-216. 3Ibid., pp. 47—66;343-344. 4Ibid., pp. 90—96. 65 The data from the twelve statements comprising the attitude—difference scores toward the concept, "Honor System for Examinations in College," for the eight ex- perimental groups were tested for homogeneity of variance. The eight groups were analyzed with factorial analysis of variance procedures. The test for homogeneity Of variance was applied to the data from the four exper— imental conditions of self—reference and the four con— ditions of prestige—reference. Each of the two groups Of four treatments was analyzed by means Of simple randomized analysis Of variance and linear regression analysis procedures. Each experimental treatment was compared with the appropriate control group by a t-test for the difference in means. Summary The procedures prOposed for this study were tested in an abbreviated pilot study. Slight modifications were made in instructions to the subjects and in the composition Of the messages. Two hundred and fifty subjects selected from the beginning speech course at the State University of New York at Buffalo were randomly assigned by groups tO the ten treatments. The tOpic, "Honor System for Examinations in College," was selected for the message. A set of Criteria plus the ractions of a group of fifty—five 66 subjects formed the basis for determining the choice of tOpic. The tOpic was researched not only for con— tent materials but also for ideas on the organization of the message and examples Of reference by the com— municator. The problem-solution pattern of organiza— tion was selected. Since both type and frequency of reference were manipulated, the following messages were devised and labeled: (1) NO-Reference, (2) High Self—Reference, (3) Moderate Self-Reference, (4) Low Self—Reference, (5) High Prestige—Reference, (6) Moderate Prestige- Reference, and (7) Low Prestige-Reference. The six experimental messages contained the same content as the control message plus the use of reference in in- creasing frequencies. The experimental messages con— tained both the content and source of the reference while the control message contained only the content of the reference. Upon an analysis Of a sample of persuasive messages, a frequency design Of zero, three, six, and nine was chosen for the references. The greatest difference in length Of time was one minute and forty—five seconds between the NO—Reference and High Prestige—Reference messages. The communicator was introduced as a recent college graduate. All of the messages were tape-recorded in a studio by the same individual. An analysis of the finished tapes by this researcher showed no apparent recording dif— ferences. ‘7‘; 1:. ,6 IN’. ~n‘ C» 67 A semantic differential consisting of four-seven- interval scales for each of three dimensions Of source credibility was selected to measure perceived credibil— ity. Each subject's score was Obtained by summing across the four scales to measure each Of the three dimensions. A seven—interval Likert—type attitude question- naire composed of twelve statements, six favorable and six unfavorable toward the tOpic of the message, was used to measure attitude change. Each subject's at- titude toward the topic was the summation of his responses to the twelve statements. The difference between a subject's score on the pre—test and the post- test was employed as an index of attitude change. The attitude questionnaire was administered to the subjects prior to the session involving the present- ation Of the message to determine their pre—message attitude toward the proposal. Ten days later each of the eight experimental groups heard one Of seven versions of the tape-recorded message. The NO-Reference message was heard by two of the experimental groups since this treatment was used as a control in an after— Only design to determine perceived credibility. Sub— jects in each of the eight groups completed both the semantic differential and the attitude questionnaire immediately following the presentation of the messages. 68 The subjects in the two control groups completed the attitude questionnaire without hearing a message. Data from each Of the three dimensions Of per— ceived source credibility and from the twelve state— ments comprising the attitude-difference scores for the eight experimental groups were subjected to a test for homogeneity of variance, factorial and simple random- ized analysis Of variance procedures, and linear re- gression analysis. The critical difference technique was applied to those analyses on perceived credibility which yielded significant §_ratios. Each experimental group was compared with the apprOpriate control group by means of a t-test for the difference in means for the attitude-difference score data. CHAPTER IV RESULTS The .05 level Of significance was originally set up for all statistical tests; however, a higher level of significance was used in some cases. The rationale for this procedure is presented in paragraph two under the sub—heading "Perceived Source Competence." Analysis Of the data yielded the following results. Perceived Source Competence As was indicated previously, responses were de- rived for three dimensions Of perceived source credi- bility. Analysis Of the data for perceived source competence revealed the following results. Application Of Bartlett's test for homogeneity of variance to the data from the eight experimental treatments, consisting of the two types of reference with four frequencies of each, revealed an observed chi—square value Of 17.83. Since the Observed value was greater than the value of 14.07 at the .05 level Of significance the hypothesis of equal variances was rejected. It must be noted, "however, the heter- ogeneity of variance must be quite extreme to be of 69 70 any serious consequence."l Therefore, "when the heter— ogeneity is 'marked' but not 'extreme,' allowance may be made for the fact by setting a higher 'apparent' level of significance for the tests of treatment effects that would otherwise be employed."2 Since an examination of the data showed slight but not extreme heterogeneity, the preceding procedure was employed. Following this procedure, a .01 level Of significance was employed in— stead Of the .05 level in analyzing the data from the eight treatment groups for perceived source competence. Factorial analysis of variance procedures was utilized to analyze the mean ratings Of perceived com— petence for subjects in the eight experimental conditions. Table 3 contains a summary of that analysis. The Type X Frequency interaction yielded a significant §_of 3.88 based on a .01 level Of significance. The test of the Frequency effect was significant with an §_Of 8.87. The test Of the Type effect was not significant. lLindquist, p. 86. 21bid. 71 Table 3.——Factorial analysis Of variance for competence scores toward the concept "The Speaker You Have Just Heard" concerning two types, self and prestige, and four frequencies--no, low, moderate, high——Of reference with an N Of 25 for each of the eight experimental groups Source SS DF MS F Type 21.12 1 21.12 3.02 Frequency 186.25 3 62.08 8.87* Type X Frequency 81.54 3 27.18 3.88* Error 1343.44 192 7.00 Total 1632.35 199 *Differences significant at the 1 percent level. Since the simple effects of the four frequencies for the two types of reference were an important con— sideration in this study, separate analyses were con— ducted on that data. Application of the test for homo- geneity of variance to the data from the four treatment conditions for self—reference resulted in an Observed value Of 1.49. Since the Observed value was less than the chi-square value of 7.82 at the .05 level of Significance, the hypothesis of equal variances among self-reference treatment groups was accepted. The Value for the four treatments comprising prestige-refer- ence was 10.28. Thus, the hypothesis Of equal variances among prestige—reference treatment groups was rejected. 72 Following the procedure outlined earlier, a .01 level Of significance was used for the analysis Of the four treatment conditions on prestige—reference. Table 4 contains the mean ratings for perceived source competence for subjects in the four experimental groups exposed to different frequencies Of self—reference and the four groups exposed to different frequencies Of prestige-reference. For the former conditions, mean ratings ranged from a high of 10.88 for subjects in the NO-Reference condition to a low Of 7.20 for subjects in the Moderate Self—Reference condition. For the lat— ter conditions, mean ratings ranged from a high of 10.16 in the NO—Reference condition to a low of 9.08 in the High Prestige-Reference condition. Mean ratings of subjects exposed to the four frequencies of self-reference and prestige-reference were plotted by means Of a graph. Figure 1 contains a representation Of the differences among the means. It must be noted that a lower rating indicates greater perceived competence. As can be seen, subjects ex— posed tO the High Self-Reference message rated the source 3.56 scale units lower on the competence dimen- Sion of credibility than did subjects exposed to the NO—Reference message. Subjects presented with the High Prestige—Reference message, however, rated the source Only 1.08 scale units lower on perceived competence than 73 did the subjects presented with the NO-Reference message. Thus, while the differences for the four frequencies Of both self—reference and prestige-reference were in the theorized direction, the differences were in the theor- ized direction, the differences were greater for the frequencies of self—reference. Table 4.—-Mean ratings for the sums of responses to four seven-interval scales measuring perceived source com— petence as revealed in a message using either self-refer— ence or prestige-reference with an N of 25 for each experimental group —— l4 4— ‘— l Type of Reference Frequency of Reference No Low Moderate High Self 10.88 9.76 7.20* 7.32* Prestige 10.16 9.24 9.28 9.08 *Differences significant at the 5 percent level are as follows: Listeners rated the Moderate Self- Reference presentation significantly higher on perceived competence than they did the NO—Reference or Low Self— Reference presentations. The listeners rated the High Self-Reference presentation significantly higher than they did the NO-Reference or Low Self—Reference pre- sentations. No significant differences were revealed between the Moderate Self—Reference and High Self—Ref- erence presentations. No significant differences were found among the Prestige-Reference presentations at the 1 percent level. Results Of simple randomized analysis Of variance procedures for the four frequencies of self—reference and a similar analysis for prestige-reference were con- sistent with the preceding Observation. Tables 5 and 6 74 l4— 13- 12- ll- 10- 9- 8— 7- 6— 5- f I ‘ I I NO Low Moderate High Frequency Of Reference Type of Reference NO ____Se1f Low I —-— Prestige Moderate High Figure l.——Differences in mean ratings of perceived source competence for subjects in the four experimental conditions 75 contain summaries Of those analyses. Table 5 shows that differences among the frequencies of self—reference yielded a significant F_of 15.85. Table 6 indicates that the test Of differences among the frequencies of prestige—reference was not significant. Table 5.——Simple randomized analysis of variance for competence scores toward the concept "The Speaker You Have Just Heard" using self-reference in four frequenc— ies——no, low, moderate, high-—with an N of 25 for each Of the four experimental groups Source SS DF MS F Between 249.95 3 83.32 15.85* Within 504.64 96 5.26 Total 754.59 99 *Differences significant at the 5 percent level. Table 6.—-Simple randomized analysis of variance for com— petence scores toward the concept "The Speaker You Have Just Heard" using prestige-reference in four frequencies—- no, low, moderate, high-—with an N of 25 for each of the four experimental groups Source SS DF MS F Between 17.84 3 5.95 0.68 Within 838.80 96 8.74 Total ,- 856.64 99 76 Linear regression analysis was employed to examine separately the nature Of the relationships among the frequencies of self-reference. Table 7 contains a sum- mary Of that analysis. As was noted previously, simple randomized analysis of variance procedures revealed a significant F ratio which indicated that a relation- ship did exist among the four self-reference conditions. The test for a curvilinear relationship resulted in a non—significant F_ratio. On the other hand, the test for a linear relationship among the four self-refer- ence conditions yielded a significant §_Of 41.68. Since the simple randomized analysis Of variance proced— ures had found no significant differences among the four prestige-reference conditions, there was no need to test for curvilinear or linear relationships. Table 7.—-Linear regression analysis for competence scores toward the concept "The Speaker You Have Just Heard" us— ing self-reference in four frequencies-—no, low, moderate, high-—with an N of 25 for each of the four experimental groups Source SS DF MS F Linear Regression 219.12 1 219.12 41.68* Dev. Linear Re- gression 30.83 2 15.41 2.93 Within 504.64 96 5.26 Total 754.59 99 *Differences significant at the 5 percent level. 77 Finally, the critical difference technique was applied to the mean ratings for the four self—reference conditions. Table 4 contains a summary Of that analy- sis. Results were basically as theorized for the self— reference presentations. Subjects rated the Moderate Self-Reference and High Self-Reference presentations significantly greater on perceived source competence than did the subjects receiving the NO-Reference and Low Self—Reference presentations. All other differences for the self—reference conditions were not significant. The critical difference technique was not applied to the mean ratings from the four conditions of prestige- reference since a significant F ratio had not been re— vealed using simple randomized analysis Of variance procedures. Perceived Source Trustworthiness Application Of the test for homogeneity of var— iance to the data for the eight experimental conditions resulted in a significant chi—square value of 17.40. An examination of the data revealed slight but not ex— treme heterogeneity. Therefore, as with the analysis of the eight experimental conditions for perceived com— petence, a .01 level of significance was used instead Of the .05 level. ~\v «xv .\~ _. . .s“ C r; 78 Table 8 contains a summary Of the factorial analy- sis Of variance procedures which were used to analyze the mean ratings on perceived source trustworthiness for subjects in the eight experimental conditions. The Type X Frequency interaction as well as the test for the Type effect resulted in non-significant F_ratios. The test for the Frequency effect yielded a significant F of 16.53. Table 8.——Factorial analysis Of variance for trustworth- iness scores toward the concept "The Speaker You Have Just Heard" concerning two types, self and prestige, and four frequencies-—no, low, moderate, high-—of reference with an N of 25 for each of the eight experimental groups e35: Source SS DF MS F Type 4.20 l 4.20 0.45 Frequency 459.45 3 153.15 16.53* Type X Frequency 37.58 3 12.52 1.35 Error 1778.72 192 9.26 Total 2279.95 199 *Differences significant at the 1 percent level. The test for homogeneity of variance was applied to the data from the four treatment conditions of self- reference. Results Of the test yielded an Observed value 79 of 6.55. The observed chi-square value for the four treatment conditions Of prestige-reference was 5.63. Thus, the hypotheses of equal variances were accepted for both types of reference on the perceived trust- worthiness dimension Of source credibility. Table 9 contains the mean ratings for perceived trustworthiness for subjects in the four experimental groups presented with increasing frequencies Of self— reference as well as for the subjects in the four groups for prestige-reference. For the former conditions, the mean ratings ranged from a high Of 11.64 for subjects in the NO-Reference condition to a low Of 7.04 for sub- jects in the High Self—Reference condition. For the latter conditions, mean ratings ranged from a high of 11.24 for subjects in the NO—Reference condition to a low of 7.72 in the High Prestige—Reference condition. Mean ratings were plotted for subjects exposed to the four self-reference presentations and the four prestige—reference presentations. Figure 2 contains a graphic representation of the differences among the means. Once again it must be pointed out that a lower rating represents greater perceived source trustworth— iness. It can be seen that subjects exposed to the High Self—Reference message rated the source 4.60 scale units lower than did the subjects exposed to the No— REference message. In a similar manner, subjects pre- Sented with the High Prestige—Reference message rated 80 the source 3.52 scale units lower than did subjects presented with the No—Reference message. These overall mean differences are consistent with the theorized position taken in this study. Table 9.-—Mean ratings for the sums of responses to four seven—interval scales measuring perceived source trust— worthiness as revealed in a message using either self— reference or prestige—reference with an N Of 25 for each experimental group Type Of Reference Frequency of Reference No Low Moderate High Self 11.64 10.56 8.12* 7.04* Prestige 11.24 9.92 9.64 7.72* *Differences significant at the 5 percent level are as follows: Listeners rated the Moderate Self-Refer- ence presentation significantly higher on perceived trustworthiness than they did the NO—Reference or Low Self-Reference presentations. Listeners rated the High Self—Reference presentation significantly higher than they did the NO-Reference or Low Self-Reference present- ations. NO significant differences were revealed be- tween the Moderate Self—Reference and High Self-Refer— eence presentations. Listeners rated the High Prestige- IQeference presentation significantly higher on perceived 'trustworthiness than they did the other three prestige— ITeference presentations. NO other significant differ- eances were found on prestige—reference. Results Of the simple randomized analysis of var- iance procedures for the four self-reference conditions and a similar analysis for the four prestige-reference my. U 14— 13- 12— 11- 10- 81 NO Low fi— I‘F Moderate High Frequency of Reference NO Low Moderate High Type Of Reference Self ——— Prestige Figure 2.--Differences in mean ratings Of perceived source trustworthiness for subjects in the four experimental conditions riff 0‘ ("AU- 1... .3,“ I5‘. ...I.*. will .. ~ "C .0 we Wfl 4t. 7. . mu 2. £1. .71 NIJA § .114 \ | .9 . :ul. «WU $1» .f- 0 ‘r. \ .LQ ...I.Il« .: .Jul ml .. 82 conditions were consistent with the preceding observa- tions. Tables 10 and 11 contain summaries Of those analyses. Table 10 shows differences among the self- reference presentations yielded a significant F_Of 15.87 while Table 11 indicates that differences among the prestige-reference presentations resulted in a sign- ificant F_Of 4.62. Application of the test for linear regression analysis to the four conditions Of self—reference dis— closed the following results which are summarized in Table 12. The test for a curvilinear relationship showed a non—significant F ratio,whi1e the test for a linear relationship yielded a significant EDOf 46.30. Concern- ing the four prestige-reference presentations, the re- sults are summarized in Table 13. The test for a curvi— linear relationship revealed a non-significant §_ratio, while the test for a linear relationship resulted in a significant F_of 12.87. Table 10.——Simple randomized analysis of variance for trustworthiness scores toward the concept "The Speaker You Have Just Heard" using self-reference in four fre— quencies--no, low, moderate, high--with an N of 25 for each Of the four experimental groups ‘ Source SS DF MS F Between 338.92 3 112.97 15.87* Within 683.52 96 7.12 Total 1022.44 99 *Differences significant at the 5 percent level. 83 Table 11.-—Simple randomized analysis Of variance for trustworthiness scores toward the concept "The Speaker You Have Just Heard" using prestige—reference in four frequencies——no, low, moderate, high—-with an N of 25 for each Of the four experimental groups Source SS DF MS F Between 158.11 3 52.70 4.62* Within 1095.20 96 11.41 Total 1253.31 99 *Differences significant at the 5 percent level. Table l2.-—Linear regression analysis for trustworthiness scores toward the concept "The Speaker You Have Just Heard" using self-reference in four frequencies—-no, low, moder— ate, high--with an N Of 25 for each of the four experimetal groups Source SS DF MS F Linear Regression 329.67 1 329.67 46.30* Dev. Linear Re- gression 9.25 2 4.62 0.65 Within 683.52 96 7.12 Total 1022.44 99 *Differences significant at the 5 percent level. 84 Table l3.--Linear regression analysis for trustworthiness scores toward the concept "The Speaker You Have Just Heard" using prestige-reference in four frequencies-~no, low, moderate, high-—with an N Of 25 for each Of the four ex— perimental groups Source SS DF MS F Linear Regression 146.88 1 146.88 12.87* Dev. Linear Re- gression 11.23 2 5.61 0.49 Within 1095.20 96 11.41 Total 1253.31 99 *Differences significant at the 5 percent level. Since significant §_ratios were Obtained in separ— ate analyses both for the four self-reference conditions and the four prestige-reference conditions, the critical difference technique was applied to the mean ratings on perceived trustworthiness. Table 9 contains a summary of that analysis. Subjects rated the Moderate Self-Refer— ence and High Self-Reference presentations significantly greater on perceived source trustworthiness than did the subjects hearing the NO—Referenceend Low Self-Reference presentations. For prestige-reference, subjects rated the High Prestige—Reference presentation significantly more trustworthy than the other three prestige-reference presentations. All other differences among self—refer— ence and prestige—reference presentations were not significant. 85 Perceived Source Dynamism Application of the test for homogeneity Of variance to the data from the experimental conditions resulted in a non—significant chi—square value of 7.78. Thus, the hypothesis Of equal variances was accepted. Table 14 contains a summary of the factorial analy- sis Of variance procedures on the mean ratings of per— ceived dynamism for subjects in the eight experimental conditions. The Type X Frequency interaction as well as the test of the Frequency effect revealed non—signific- ant results. The test of the Type effect, however, yielded a significant §_Of 5.08. The test for homogeneity Of variance applied to the four self—reference conditions yielded a non-signific— ant Observed value of 0.72, while the observed value for the four prestige—reference conditions was non-significant at 3.54. The hypotheses Of equal variances were accepted for both types of reference on the perceived dynamism dimension of source credibility. Table 15 contains the mean ratings for perceived dynamism for subjects in the four experimental grOUps exposed to the increasing frequencies of self-reference as well as for the four groups exposed to increasing frequencies of prestige—reference. 86 Table l4.—-Factoria1 analysis of variance for dynamism scores toward the concept "The Speaker You Have Just Heard" concerning two types, self and prestige, and four frequencies—-no, low, moderate, high——Of reference with an N Of 25 for each of the eight experimental groups Source SS DF MS F Type 53.04 1 53.04 5.08* Frequency 37.09 3 12.36 1.18 Type X Frequency 37.06 3 12.35 1.18 Error 2004.96 192 10.44 Total 2132.15 199 *Differences significant at the 5 percent level. Table 15.-—Mean ratings for the sums Of responses to four seven—interval scales measuring perceived source dynamism ‘as revealed in a message using either self—reference or prestige—reference with an N of 25 for each experimental group Type Of H Reference Frequency of Reference NO Low Moderate High Self 10.24 9.84 9.56 9.04 Prestige 9.88 11.84 10.84 10.24 87 For the former presentations, the mean ratings ranged from a high of 10.24 for subjects in the NO-Refer- ence condition to a low of 9.04 for subjects in the High Self—Reference condition. For the latter presentations, the mean ratings ranged from a high Of 11.84 in the Low Prestige-Reference condition to a low of 9.88 in the NO— Reference condition. Plotted mean ratings for subjects presented with messages representing the four frequencies of self— reference and the four frequencies Of prestige—reference are shown in Figure 3. A lower rating indicates greater perceived dynamism. Subjects exposed to the High Self— Reference presentation rated the source 1.20 scale units lower than did the subjects exposed to the NO—Reference presentation. The plotted mean ratings for the fre- quencies Of prestige—reference tended to follow the Op- posite pattern. Subjects presented with the High Prest— ige-Reference message rated the source 0.36 scales units higher than did subjects presented with the No—Reference rnessage. In addition, the Low Prestige—Reference mes— ssage was rated 1.96 scale units higher on perceived dy- riamism than the NO—Reference message. Simple randomized analysis of variance procedures vware not applied to the data since results Of the factor— ixal analysis of variance procedures revealed a non— 88 l4— 13- 12- ,\ z”, \”‘ 4 ll- ’/ \~\ I \ ‘-~~—,~ -__________‘ 9- 8- 7- 6- 5- v I a 1 No Low Moderate High Frequency of Reference Type of Reference No Self Low --- Prestige Moderate High Figure 3.——Differences in mean ratings of perceived source dynamism for subjects in the four experimental conditions 89 significant §_ratio among the eight experimental condit— ions and an inspection Of the plotted mean ratings on dynamism indicated minor differences. Linear regression analysis was not conducted for the same reason. The critical difference technique was not applied to the data because no significant relationship was indicated for the Frequency effect. Attitude Change_ Responses were derived from a seven—interval Likert- type attitude questionnaire consisting Of twelve statements, six favorable and six unfavorable, toward the concept "The honor system should be adOpted for examinations in col— lege." A pre—and post—test design was employed with eight experimental groups and two control groups. Analy— sis Of the data on attitude change revealed the following results. Application Of the test for homogeneity Of var— iance to the data from the eight experimental conditions yielded a non-significant Observed chi—square value of 11.09 at the .05 level Of significance. The hypothesis of equal variances was accepted. . Results Of the factorial analysis of variance Elrocedures employed on the mean ratings Of the attitude— ciifference scores for subjects in the eight experimental cxanditions are found in Table 16. The Type X Frequency 90 interaction and the test Of the Type effect were not significant. The test Of the Frequency effect was significant,yielding an §_Of 2.97. Table 16.—-Factorial analysis of variance for attitude scores toward the concept "Honor System for Examinations in College" concerning two types, self and prestige, and four frequencies-—no, low, moderate, high--of reference with an N Of 25 for each of the eight experimental groups Source SS DF MS F Type 162.00 1 162.00 2.03 Frequency 711.58 3 237.19 2.97‘ Type X Frequency 312.72 3 104.24 1.31 Error 15318.88 192 79.78 Total 16505.18 199 *Differences significant at the 5 percent level. The test for homogeneity Of variance applied to the four treatment conditions of self—reference resulted in a non-significant value of 4.65. The observed value for the four treatment conditions of prestige—reference was not significant at 4.05. The hypotheses of equal variances were accepted for both types of reference on attitude change. Table 17 contains the mean ratings derived from the attitude-difference scores for subjects in the four experimental conditions exposed to increasing frequencies 91 of self-reference and the four conditions exposed to increasing frequencies of prestige-reference. The mean ratings for subjects in the two control conditions are included. Both a minus sign and a higher mean rating indicate attitude change in the direction advocated in the message. For the self-reference presentations, the mean ratings ranged from a high of —l2.12 for subjects exposed to the Moderate Self-Reference message to a low Of .36 for subjects in the control condition who heard no message. For prestige—reference, the mean ratings ranged from a high Of —9.48 for subjects presented with a High Prestige—Reference message to a low of 1.20 for subjects in the control condition who heard no message. Table l7.—-Mean ratings for the sums of responses to twelve seven—interval scales measuring attitude change for sub— jects in experimental and control groups as revealed in a message using either self—reference or prestige—reference with an N Of 25 for each group Type Of Reference Frequency of Reference Control NO Low Moderate High Self .36 -6.04 -5.76 -l2.12 -10.84 Prestige 1.20 -5.88 —6.08 -6.12 —9.48 Both a minus sign and a higher mean rating for an experimental group than for a control group indicate the message was effective. 92 Figure 4 contains a representation of the differences among the mean ratings for subjects in the eight experimen- tal conditions. A higher mean rating indicates a greater shift in attitude on the part of the subjects. The great— est difference can be seen between the following conditions. Subjects exposed to the Moderate Self—Reference present- ation showed an attitude change of 6.36 scale units more than did subjects exposed to the Low Self—Reference pre- sentation. Subjects presented with the High Prestige- Reference message revealed an attitude change of 3.60 scale units more than subjects presented with the No-Refer— ence message. It can be seen that, while the differences were in the theorized direction for both self-reference and prestige-reference conditions, the differences were greater among the self—reference conditions. Results of simple analysis Of variance procedures for the four frequency conditions of self—reference and a similar analysis for the four conditions of prest— ige—reference were consistent with the preceding Obser— vations. Tables 18 and 19 contain summaries Of those analyses. Table 18 shows differences among the four experimental conditions of self-reference disclosed a significant F_of 2.89. Table 19 indicates that dif— ferences among the four experimental conditions Of prest— ige—reference were not significant. 14-; 13— » 12- 11- 10- 93 r Moderate High Frequency of Reference NO Low Moderate High Type Of Reference Self -—— Prestige Figure 4.-—Differences in mean attitude—difference scores for subjects in the four experimental conditions 94 Table 18.--Simple randomized analysis of variance for at— titude scores on the concept "Honor System for Examinations in College" using self—reference in a message in four fre- quencies—-no, low, moderate, high--with an N of 25 for each of the four experimental groups Source SS DF MS -. F Between 799.87 3 266.62 2.89* Within 8865.52 96 92.35 Total 9665.39 99 *Differences significant at the 5 percent level. Table 19.——Simple randomized analysis of variance for at— titude scores on the concept "Honor System for Examinations in College" using prestige—reference in a message in four frequencies——no, low, moderate, high—-with an N Of 25 for each Of the four experimental groups Source SS DF MS F Between 224.43 3 74.81 1.11 Within 6453.36 96 67.22 Total 6677.79 99 Linear regression analysis was employed to examine the nature of the relationship among the four self-refer- ence conditions. Table 20 contains a summary of that analysis. The test for a curvilinear relationship resulted 95 in a non—significant E ratio. The test for a linear relationship among the four conditions of self-reference yielded a significant §_Of 5.83. Since the simple ran— domized analysis Of variance procedures indicated no significant relationship among the four prestige-refer— ence conditions, tests for curvilinear and linear re- lationships were not conducted. Table 20.—-Linear regression analysis for attitude-dif- ference scores on the concept, "Honor System for Examina— tions in College," using self—reference in a message in four frequencies--no, low, moderate, high—-with an N of 25 for each Of the four experimental groups Source SS DF MS F Linear Regression 538.72 1 538.72 5.83* Dev. Linear Re— gression 261.15 2 130.57 1.41 Within 8865.52 96 92.35 Total 9665.39 99 *Differences significant at the 5 percent level. The critical difference technique was applied to the mean ratings for the four self-reference presentations since these conditions had yielded a significant F ratio. Table 21 contains a summary of that analysis. Attitude change was significantly greater for subjects presented 96 ‘with the Moderate Self—Reference message than for sub- jects hearing the No—Reference and Low Self-Reference presentations. All other differences were not significant. Table 21.-—A Comparison of the mean ratings for sums of responses to twelve seven-interval scales measuring at— titude change in experimental groups as revealed in a message using self-reference with an N of 25 for each group Type of Reference Frequency Of Reference NO Low Moderate High Self —6.04 -5.76 -12.l2* —10.84 *Differences significant at the 5 percent level eare as follows: Attitude change was significantly greater :Eor subjects exposed to the Moderate Self-Reference pre- :sentation than for subjects hearing the NO—Reference and Ixaw Self—Reference presentations. All other differences \uere non-significant. Table 22 contains a summary of a comparison of the rnean ratings for each of the experimental groups with the apuiropriate control group. A t-test for the comparison Of Ineans was employed. In each instance, the experimental gIXJup showed a significantly more favorable attitude txMNard the honor system for examinations in college than diti the control group- 97 'Table 22.-—A comparison of the attitude—difference scores of subjects in the control groups and experimental groups on the concept, "Honor System for Examinations in College," with an N of 25 for each group No Low Moderate High Reference Reference Reference Reference Control Self—Reference .05* .05* .05* .05* Control Prestige-Refer- ence .05* .05* .05* .05* ‘In each instance the group designated by the column heading showed significantly greater attitude change tO— ward the honor system for examinations in college than did the group designated by the heading for the row. Summary A summary of the results Of this study can best be estated in terms of the results Of each of the major and Iaull hypotheses. Major Hypotheses ‘ 1.) Increasing the frequency Of self-reference by a source of a message will effect: a) greater perceived competence for that source. (This study lends support to this hypothesis.) b) greater perceived trustworthiness for that source. (This study lends support to this hy- pothesis.) c) greater perceived dynamism for that source. (This study provides no support for this hy- pothesis) 2) Increasing the frequency of prestige-reference by a SOUrce Of a message will effect: a) greater perceived competence for that source. (This study provides no support for this hy- pothesis.) 13) 4) S) 6) 7) 8) 98 b) greater perceived trustworthiness for that source. (This study lends support to this hypothesis.) c) greater perceived dynamism for that source. (This study provides no support for this hypothesis.) The increase in perceived competence for a source of a message is directly prOportional to the increase in the use of: a) self—reference by the source, that is, the trend is linear. (This study lends support to this hypothesis.) b) prestige-reference by the source, that is, the trend is linear. (This study provides no sup- port for this hypothesis.) The increase in perceived trustworthiness for a source of a message is directly prOportional to the increase in the use of: a) self-reference by the source, that is, the trend is linear. (This study lends support to this hypothesis.) b) prestige-reference by the source, that is, the trend is linear. (This study lends sup— port to this hypothesis.) The increase in perceived dynamism for a source of a message is directly proportional to the increase in the use of: a) self-reference by the source, that is, the trend is linear. (This study provides no sup— port for this hypothesis.) b) prestige-reference by the source, that is, the trend is linear. (This study provides no sup— port for this hypothesis.) Increasing the frequency Of self-reference by a source Of a message will effect greater attitude change. (This study lends support to this hypothesis.) Increasing the frequency Of prestige-reference by a source of a message will effect greater attitude change. (This study provides no support for this hypothesis.) Increases in attitude change are directly pro- portional to the increases in the use of: a) self-reference by the source, that is, the trend is linear. (This study lends support to this hypothesis.) b) prestige—reference by the source, that is, the trend is linear. (This study provides no support for this hypothesis.) 22) :3) 4) 99 Null Hypotheses Frequency and type Of reference do not interact in affecting: a) b) c) d) the perceived competence of the source of a message. (Rejected at .01) the perceived trustworthiness of the source of a message. (Fail to reject at .01) the perceived dynamism of the source of a mes— sage. (Fail to reject at .05) attitude change. (Fail to reject at .05) Frequency of reference by a source of a message does not affect: a) b) c) d) the source's perceived competence. (Rejected at .01) the source's perceived trustworthiness. (Re- jected at .01) the source's perceived dynamism. (Fail to re- ject at .05) attitude change. (Rejected at .05) Type of reference by a source of a message does not affect: a) b) c) d) Simple the source's perceived competence.(Fai1 to re- ject at .01) the source's perceived trustworthiness. (Fail to reject at .01) the source's perceived dynamism. (Rejected at .05) attitude change. (Fail to reject at .05) messages do not result in significant changes in attitude. (Rejected at .05) The purpose of this chapter has been to present true results of the study. A discussion Of these results frillows in Chapter V. CHAPTER V DISCUSSION The basic assumption underlying this study was that cognitive imbalance is created in a communication situation when the source Of a message is unfamiliar to the receivers of the message. It was suggested that rejecting the proposal and/or lowering the perceived credibility of the source would be the most effective means for the receiver to remove or reduce this imbal— anced state. When positive information is revealed about the source during the presentation oflythe message, the most likely outcome would be for the receiver to accept the proposal and/or to increase the perceived credibility of the communicator. In other words, the source is made familiar to the receivers of the message. It was further suggested that as the positive information about the source increases, the attitude change on the part of the receivers of the message and/Or the perceived credibility Of the source will increase. The result would be that a source of a High RSifterence message would be perceived as more competent, trUStworthy, and dynamic than a source of a NO-Reference message. In addition, a source of a High Reference mes- sage would be perceived as more competent, trustworthy, 100 Lu «C C u .11. 101 and dynamic than a sourcecr a Moderate Reference mes— sage, who in turn, would be rated higher on those same dimensions of credibility than a source of a Low Ref- tlier instances, psychological theory has gone beyond VHfiEit has been presented in rhetoric on the concept of SCnJrce credibility. Heider's balance theory, one Of tfliez consistency theories, seemed to be of particular reElevance for the rationale for this study. Balance thfiaory hypothesizes that imbalance may result when one ‘3f the entities in a communication situation is unfamil- iiaJT. Thus, if the source is unfamiliar to the receiver arui the source asserts a proposal on the tOpic of a mes- Seuge, an imbalanced state will exist. A second prin- <3iE>1e advanced in balance theory is that the message ncrt only reveals something about the topic but also reVeals something about the source. The concepts, (1) ‘urtfamiliarity Of source leads to an imbalanced state arhfl (2) message itself reveals something about the source 117 Of the message, taken together lead into the rationale for this study. It is evident that a typical communication situa- tion could include a situation in which the audience is unfamiliar with the source Of a message. Without a speech introducing the communicator to the audience, the communicator could conceivably proceed to deliver his message with the audience having no knowledge about him. In a similar communication situation, however, the source could reveal positive information about him- self as he delivers his message. This method could be described as the use of reference by the source. Given the assumption that an unfamiliar source does indeed lead to cognitive imbalance in the receiver and given the assumption that the use of reference by an unfamil— iar source increases the familiarity Of the source to the receiver, the following was hypothesized. If lis— teners to a message are presented with positive inform- ation about the communicator, they will exhibit greater perceived source credibility and greater attitude change toward the proposal than if no information is presented. Furthermore, if listeners are presented with positive information in increasing frequency, they will manifest an increasingly greater degree Of perceived source cred- ibility and attitude change. The specific hypotheses pT—‘Oposed in this study are delineated on pages 39-40. 118 The results of an abbreviated pilot study in- dicated no major changes were necessary; therefore, the study was conducted as originally planned. Two hundred and fifty subjects from the beginning speech course at the State University of New York at Buffalo were randomly assigned as groups to the ten treatments. The tOpic for the message was "Honor System for Exam— inations in College." Since both type and frequency of reference were manipulated, the messages composed were: (1) No—Reference, (2) High Self-Reference, (3) Moderate Self—Reference, (4) Low Self-Reference, (5) High Prestige—Reference, (6) Moderate Prestige—Refer— ence, and (7) Low Prestige—Reference. The six experi- mental messages contained both the content and source of each reference while the control message contained only the content of each reference. A frequency de- sign of zero, three, six, and nine was chosen for the increasing frequency of reference. All of the messages were recorded by the same individual on tape in a studio. The communicator was introduced to the audience as a recent college graduate. A semantic differential consisting of four seven- interval scales for each of three dimensions of source credibility measured perceived credibility. A seven- interval Likert—type attitude questionnaire composed of twelve statements on the tOpic of the honor system 119 measured attitude change. The attitude questionnaire was administered to the subjects to determine their pre-message attitude toward the proposal. Ten days later, after hearing the appropriate message, the at— titude questionnaire was again administered along with the semantic differential. The control groups completed the attitude questionnaire without listening to a mes- sage. The data were subjected to tests for homogeneity of variance, factorial and simple randomized analysis of variance procedures,and linear regression analysis. The critical difference technique was applied to those analyses on perceived credibility yielding significant E ratios. For the attitude-difference score data, each experimental group was compared with the appropriate con— trol group by means of a t—test. The specific findings of this study can best be stated in terms of the results of each of the null and major hypotheses which are found on pages 97—99. In a broad sense, the findings may be discussed in the following context. The theoretical position taken in this study was that cognitive imbalance would lead receivers exposed to a message presented by an unfam- iliar source to rate the source as less competent, trust- worthy, and dynamic than receivers exposed to a message delivered by an unfamiliar source made familiar through the use of reference in a message. Under similar 120 conditions, it was assumed that receivers who heard an unfamiliar source would exhibit less attitude change than receivers who heard a familiar source. It was further hypothesized that a linear relationship existed between an increase in the use of reference by an unfamiliar source and an increase in perceived credibility and at- titude change. The findings of this study generally lend sup- port to the theoretical position for self—reference but basically do not provide support for prestige—reference by a communicator. This researcher believes several contributions were made to further the understanding of the concept of source credibility as a result of this study. The findings did lend support to the position that self— reference by a source generates perceived credibility for an unfamiliar source of a message and results in significant changes in attitude on the part of receivers of that message. An unfamiliar source was perceived as more competent and trustworthy but not more dynamic when he used self—reference in a message. It was found that increases in the frequency of self—reference ap- peared to be followed by prOportional increases in perceived credibility and attitude change. This find- ing concerning self-reference differs from previous research which has concluded that neither source credi- bility nor attitude change was affected by message 121 elements designed to increase credibility. The finding on the frequency effect is of particular significance since previous research has been lacking in an analysis of the frequency effect of message elements. The findings did not provide support for the position that prestige—reference by a source generates credibility for an unfamiliar source or that its use results in significant changes in attitude. Competence, dynamism, and attitude change were not significantly af- fected by increases in prestige-reference by the com— municator. An increase in perceived trustworthiness was revealed. The findings on prestige—reference when compared with the findings on self-reference would in- dicate the presence of differential effects among the various message elements aimed at increasing credi- bility. These results would seemingly point to the need to study message elements separately and at various frequencies to determine their influence on source cred— ibility and attitude change. The findings in this study did not provide support for the position that type and frequency of reference interact in affecting perceived trustworthiness, dy— namism, or attitude change. The results did lend sup— port for the premise that type and frequency interact in affecting perceived competence. 122 Results showed that the type of reference did af- fect the perceived dynamism of the source but did not affect the competence or trustworthiness of a source or the attitude change of listeners to the message. An unfamiliar source who used prestige—reference was per- ceived as less dynamic than a similar source who used self—reference. APPENDIX I A. ATTITUDE QUESTIONNAIRE FOR SELECTION OF THE TOPIC OF THE MESSAGE Questionnaire Indicate your Opinion on the following statements by circling the point on the scale which most closely ap— proximates your answer. Compulsory military service should be abolished. strongly neutral or strongly agree agree don't know disagree disagree 2 l O -l -2 The EurOpean educational system is superior to ours. strongly neutral or strongly agree agree don't know disagree disagree 2 1 O -1 -2 New York State should abolish compulsory auto— mobile liability insurance. strongly neutral or strongly agree agree don't know disagree disagree 2 l O —l —2 The honor system should be adOpted for exam— inations in college. strongly neutral or strongly agree agree don't know disagree disagree 2 l O —1 —2 Compulsory unionism should be made illegal. strongly neutral or strongly agree agree don't know disagree disagree 2 l O —l -2 124 B. PRE-TEST ATTITUDE QUESTIONNAIRE FOR EXPERIMENTAL AND CONTROL GROUPS; POST—TEST ATTITUDE QUESTIONNAIRE FOR CONTROL GROUPS We are interested in your reactions toward a controversial topic for a speech. Our purpose is to set up a scale to measure attitudes toward the honor system for examinations in college. This questionnaire has 23 bearing on your grade for Speech 126. You are asked to read statements phrased as fol- lows. Compulsory unionism should be made illegal. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree You are asked to check the response that most accurately describes your reaction to the statement. If you strongly agree with the statement as it stands, check the blank next to the words Strongly Agree; if you agree with the statement as it stands, check next to the word Agree; and so on, with regard to the other responses (Slightly Agree, Uncertain, Slightly Disagree, Disagree, Strongly Disagree). Do not spend a lot of time thinking about any one statement. We are interested in your immediate reaction. There are no right or wrong answers. Record your respon— ses to every statement. The following definitions are provided for a clearer understanding of the topic of the statements: Honor System — N2_faculty member is obliged to be present in the room while col- lege students are taking an ex- amination. N2 instructor serves as proctor for the students; the students are on their own. 125 126 Proctor System — A faculty member is present in the room while college students are taking an examination. An instructor serves as proctor during the exam. His presence is meant to discourage and/or detect any cheating. Complete the following preliminary information, then proceed immediately to the statements. Name Age Classification Sex (Fr.,Soph.,Jr.,Sr.,Special) fl_or E Thank you for your cooperation. 127 Check the response that most accurately describes your reaction to the statements. 1. The proctor system fails to deal with the factors which encourage cheating. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree strongly Disagree 2. The honor system is based on unsound educational principles. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 3. The faculty-proctor system works satisfactorily in preventing cheating. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 4. The honor system encourages in students a sense of individual responsibility. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 5. Competition for grades in college requires the use of the proctor system. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 128 6. The honor system helps the dishonest student and penalizes the honest student. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 7. The honor system makes a significant contribution to the college education of the student. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 8. The honor system is essential for a strong univer- sity educational program. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 9. The honor system is another unwarranted attempt to teach morality in the university. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 10. The honor system is fine in theory but unworkable in practice. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 129 11. The honor system does reduce the amount of cheating. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree ||||||| 12. The honor system, as a solution to cheating, is superior to all others. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree C. POST—TEST ATTITUDE QUESTIONNAIRE FOR EXPERIMENTAL GROUPS Preface You have just listened to a speaker deliver a speech on the honor system. We would now like your reactions to it. This five-page booklet consists of two parts. Part I calls for your reactions to state- ments about the tOpic of the speech while Part II requests your reactions to the speaker. This question- naire has pg effect on your grade for Speech 126. INSTRUCTIONS PART I On the next two pages, you will find statements pertaining to the topic of the speech. You are asked to read statements phrased as follows: The honor system does more harm than good to the university community. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree You are asked to check the response that most accurately describes your reaction to the statement. If you strongly agree with the statement as it stands, check the blank next to the words Strongly Agree; if you agree with the statement as it stands, check next to the word Agree; and so on, with regard to the other responses (Slightly Agree, Uncertain, Slightly Disagree, Disagree, Strongly Disagree). Do not spend a lot of time thinking about any one statement. We are interested in your immediate reaction. There are no right or wrong answers. Record your response to each and every statement. Please complete the following information, then proceed immediately to the statements. 130 131 Name Age Classification Sex TFr.,SOph.,Jr.,Sr.,Specia17’ M or F Thank you for your cooperation. 132 Check the response that most accurately describes your reaction to the statements. 1. The proctor system fails to deal with the factors which encourage cheating. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 2. The honor system is based on an unsound educational principles. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 3. The faculty—proctor system works satisfactorily in preventing cheating. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 4. The honor system encourages in studentsa sense of individual responsibility. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 133 5. Competition for grades in college requires the use of the proctor system. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 6. The honor system helps the dishonest student and penalizes the honest student. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 7. The honor system makes a significant contribution to the college education of the student. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 8. The honor system is essential for a strong univer- sity educational program. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 9. The honor system is another unwarranted attempt to teach morality in the university. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 134 10. The honor system is fine in theory but unworkable in practice. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 11. The honor system does reduce the amount of cheat— ing. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree 12. The honor system, as a solution to cheating, is superior to all others. Strongly Agree Agree Slightly Agree Uncertain Slightly Disagree Disagree Strongly Disagree D. SEMANTIC DIFFERENTIAL ADMINISTERED TO EXPERIMENTAL GROUPS TO MEASURE PERCEIVED SOURCE CREDIBILITY INSTRUCTIONS PART II At the tOp of the next page, you will find the heading "THE SPEAKER YOU HAVE JUST HEARD" and beneath it a list of descriptive words. These are the kinds of words you would likely use in "describing" a per- son, or in this specific case, a speaker. You are asked to rate "THE SPEAKER YOU HAVE JUST HEARD" on each of the twelve sets of words, each of the scales, in order. Here is an example showing how to use the scales. THE SPEAKER YOU HAVE JUST HEARD GOOD : : : : : : BAD -2 -3 3 2 1 O -l If you felt that the speaker was inygeneral extremely good,you would place a check mark in space number 3. If quite good (but not extremely good), in 2; if slightly good in 1; if neither good nor bad, in 0; if slightly bad, in —1; if ggite bad, in -2; and if extremely bad, in -3. The "O" or neutral space on the scale may also be used for "I don't know" or "I don't think this scale applies" answers. IMPORTANT: 1) Be sure to check each and every scale — Do Not Omit Any. 2) Place your check marks within the spaces; not on the lines separating the spaces. 3) Put one, and only one, check mark on each scale. Do not puzzle over individual sets of words. It is your first impression, your "immediate feelings," that we want. On the other hand, please do not be care— less. Remember, rate the speaker you have just heard. 135 136 THE SPEAKER YOU HAVE JUST HEARD AGGRESSIVE __:__:___:__ __ __ __ MEEK IGNORANT ___:____:___:__:____:____:__ EXPERT UNIMPORTANT : :_: : : : IMPORTANT INFORMED : : UNIFORMED EXTROVERTED : : : INTROVERTED HONEST : : DISHONEST CLOSED—MINDED OPEN—MINDED INEXPERIENCED ___:__:________:_____:____:__:___ EXPERIENCED REASONABLE ____:___:____:____:______:___:___ UNREASONABLE PASSIVE _____:__:__:____:__:___:___ ACTIVE BOLD ___:___:_____:__:___:___:____ TIMID JUST : z : : : : UNJUST E. THE INTRODUCTION GIVEN TO THE SOURCE OF THE MESSAGE PREFACE Individuals concerned with the problems of human com- munication are constantly attempting to learn more about this concept. Today, we would like your cooperation as we investigate one aspect of the communication situation. In a few moments, you will hear a tape recording of a speech. At the conclusion of the speech,you will be asked your reactions to it. A questionnaire will be passed out to you at that time. You will not be tested on the content of the speech nor will your reactions affect your grade in Speech 126. Howevery please listen carefully. A WORD ABOUT THE SPEAKER The SPEAKER is a recent college undergraduate having graduated from a university this past June. 137 F. ADMINISTRATIVE PROCEDURE FOLLOWED WITH THE EXPERIMENTAL GROUPS Taped Speech 1. Read aloud to the students the one—page sheet con— sisting of a preface to the study and an introduction of the speaker. 2. Play the tapeerecorded speech with the volume set at five. Sit in the back of the room while the speech is being delivered. 3. At the conclusion of the speech, hand out the five- page booklet to each subject. (Say aloud to the subjects) PLEASE FOLLOW ALONG WITH ME AS I READ THE PREFACE. NOW, PLEASE CONTINUE ON, READING THE INSTRUC- TIONS CAREFULLY ON THE REMAINDER OF THIS PAGE. AT THE BOTTOM OF THE PAGE, FILL IN THE INFORM— ATION THAT IS REQUESTED. THEN, PROCEED IM- MEDIATELY TO COMPLETE PAGES TWO AND THREE. AFTER YOU HAVE COMPLETED PAGE THREE, WAIT FOR FURTHER INSTRUCTIONS BEFORE YOU GO ON. YOU MAY BEGIN. After three to four minutes, check to see if everyone has finished. As soonemsthe subjects have finished, continue by — (Saying aloud to the subjects) TURN TO PAGE FOUR AND READ WITH ME THE INSTRUC- TIONS FOR PART II. Upon finishing the reading of the instructions, tell the subjects — YOU MAY BEGIN. After all subjects have finished but before they hand in their booklets — (Say aloud to the subjects) PLEASE CHECK TO SEE THAT YOU HAVE FILLED IN ALL INFORMATION THAT IS REQUESTED. 4. Collect the booklets. 138 G. ADMINISTRATIVE PROCEDURE FOLLOWED WITH THE CONTROL GROUPS No Speech 1. Before handing out the questionnaires — (Say aloud to the subjects) DUE TO THE NATURE OF THE PROJECT WE ARE CONDUCT- ING IN SPEECH 126, WE ARE ASKINGTHNJTO COMPLETE THE QUESTIONNAIRE ON THE HONOR SYSTEM. LET ME STRESS THAT IT IS VERY IMPORTANT THAT YOUR RESPONSES ACCURATELY DESCRIBE YOUR RE- ACTIONS TO THE STATEHENTS. PLEASE DO NOT BE CARELESS IN YOUR ANSWERS. READ EACH STATEMENT CAREFULLY AND ANSWER EACH ONE CONSCIENTIOUSLY. 2. Hand out the questionnaires to the subjects - (Saying aloud) YOU MAY BEGIN. 3. After all subjects have finished but before they hand in their questionnaires - (Say aloud to the subjects) PLEASE CHECK TO SEE THAT YOU HAVE FILLED IN ALL INFORMATION THAT IS REQUESTED. 4. Collect the questionnaires. 139 APPENDIX II A. NO-REFERENCE MESSAGEl Consider, if you will, the following statement from the College of William and Mary Quarterly. "While nothing is required of a college student but attention to his studies, nothing is forbidden but dueling which might be fatal to his life, and gambling and drunken- ness and tavern haunting, which must be pernicious to his health and murals." This unusualsflxtement was an attempt to formulate a system of student conduct, a code of honor, at that college in the late 1800's. It is generally held that the first attempt in American colleges to phrase a code of honor for students was inaugurated by Thomas Jefferson in the founding of the University of Virginia in 1825. Through the years codes of student conduct have taken on many forms and have been applied to colleges of all classifications. At Barnard, principles of student conduct have per- tained only to quizzes and written reports. At Beloit College in Wisconsin, a code has been set up for the student use of reserve books in the college library. All student activity fell under a system of honor at South Carolina; while, as we have seen, only moral conduct was included at William and Mary. On the other lApproximate message length - eleven minutes and five seconds. 141 142 hand, an honor code applied to all rules made by the student association at Wilson College in Pennsylvania. These colleges illustrate the unusual and many diverse attempts to set down principles of student conduct. They are not, however, typical of what we know as the honor system today. For an idea of the form of the honor system used most widely, place yourself in the following situation. This semester you, along with thousands of other college students, face examinations in your various course of study. If you were to walk into your next exam and find no proctor present, no instructor watch- ing you during the examination to discourage and note any attempts at cheating, you in effect would be under the honor system. The use of unproctored tests is the most prevelant form of the honor system and is the topic which I would like to discuss with you today. It is likely there are several individuals who have experienced under both the honor and faculty—proctor systems what will be presented in this speech.1 You might ask why a college should have the honor system for examinations. The fundamental pur- pose for any system of this nature is to attempt to curb cheating. Academic cheating, in one form or another, has been with us through the ages. In the Oriental Library at Princeton University, an interesting 1 . . . 'Material presented w1th no reference to the com- municator is underlined throughout the message. 143 artifact from ancient China, a cribbing shirt, is dis— played. I guess this only proves that exam pressure was evident even then and that short cuts were employed to meet this pressure. What about the extent of cheating in college to— day? Columbia University, in February 1965, released the results of what has been reported to be the most comprehensive study ever made among college students. Nearly half of the 5,000 students questioned, in strict confidence, at 99 colleges and universities admitted they had engaged in some form of cheating since enter— ing college. Nearly fifty percent cheat! We can all agree that is quite a staggering figure. Does this mean that today's college student is excessively im- moral? It has been said that is not a fair accusation. The same Columbia study just cited found that ninety percent of the students, including many who admitted to cheating, said they were opposed to the practice on moral grounds. This information would seem to cast doubt on the label of "excessively immoral" being placed on college students. The question then arises, if students are basically opposed to cheating, why do they do it. The environment at their_particular col- leges might be a key to understanding this problem. 144 Within the context of the environment and pur- pose of a university, let's examine the two systems aimed at eliminating the incidence of cheating on examinations-—the proctor system and the honor system. To seek, guard, and speak for the truth are considered the duties of any institution of higher learning. We would hope college seeks, through the exchange between students and faculty, not merely to convey knowledge and gain skills, but also to develop understandings, insights, and character. How do the proctor and honor systems fit in with this philosophy of a university? The theory behind the proctor system for exam- inations is that, with the instructor watching you, the threat of his detecting cheating on your part will<fiscourage you from cheating. How does the proctor system work in practice? Let's look at the examination situation in our attempt to answer this question. Don't we often think of an examination as a matching of wits between student and instructor? In effect, through the material on the test, the instructor is offering a challenge to the student and the student attempts to meet this challenge. This is a healthy situation. However, now add the proctor to the examination. Hasn't the area of challenge broadened? Since the instructor is present to detect cheating, isn't his physical pre- sence an invitation to students to attempt to "beat the 145 system?" Such a system does not help to maintain an environment where a free exchange between student and faculty can truly take place. The theory of the honor system is based on the premise that to attack the problem of cheating one must foster an atmosphere of learning in which cheating is impractical. In other words, the university should develop a setting in which it is expected that cheat- ing will not take place. The honor system simply means no instructor is obliged to be present during examinations. Undoubtedly it seems amazing to some people that an instructor could leave his class dur- ing an examination. It can be seen that with the fac— ulty-proctor absent, one of the motives for cheating is removed. There is no longer the same challenge to try to outsmart the instructor. Some of you may say that some students will still attempt to cheat. However, a key factor is working under the honor system which is not present under the proctor system. Students taking examinations with a proctor depend upon him to deal with cheating. Students may assume the attitude of dissociating themselves from the prob— lem. Let the instructor worry about it! It is his problem! The honor system places cheating in a dif- ferent light. Suddenly, the students as a group and 146 and as individuals have the responsibility placed on their shoulders. The student who may have cheated under the proctor system now has second thoughts. He is more reluctant, whether for personal reasons or a desire not to go against his peers. Gradually, it be— comes a personal obligation on the part of each student to see that fairness is maintained. The question inevitably arises: Does the honor system actually curb cheating? It is fine in theory but do students take it seriously. It is possible the faculty at first may doubt whether the students can do much better than themselves but after awhile they will likely say that, though there are problemsl the situa- tion has improved. There is considerable evidence from students and faculty alike that the honor system is ef— fective. We can see results at a specific university. A five-year study was conducted prior, during, and after the adoption of the honor system at Brigham Young University. It was discovered that rates of cheating in the classes were reduced by nearly two—thirds. While these findings do not necessarily prove a direct causal relationship between the honor system and a reduction of cheating, the liklihood is certainly pre- sent. Some of you may question the typicalness of such a specific instance. Yet, the study of 99 colleges by Columbia University cited earlier, noted that cheating 147 is most evident at schools where the faculty alone tries to control the problem. It occurs far less often at colleges with an honor system. Thus, it would seem reasonable to conclude that the honor system is ef— fective in achieving one of its primary purposes--c0ping with the problem of cheating. In addition to being a workable solution, the honor system has other distinct advantages. Its great- est benefit is that it is educational as well as regul— atory. Shouldn't a university help to create an en— vironment which will be most favorable to develOping responsible individuals? What about the proctor system for examinations? In the most favorable setting, it is a kind of supervision. Most of the time, it is un— fortunately an inspection system and, at its worst, it is a policing system. Is this the best our colleges and universities can do for us? The develOpment of responsibility and character should not stOp at an inspection or police stage. One does his best when he really believes somebody has faith in him. Its the indifference under the proctor system that makes cheat- ing so much easier. It should be possible for you, as college students,to assume the responsibility of maintaining fairness in examinations. Now it is important to realize that the honor system is not without criticism. The Air Force Academy 148 incident of 1965 is pointed to as an example of the failure of the honor system to curb cheating. But, do you realize that only four percent of all of the cadets were implicated in the cheating incident. Compare that figure with the nearly fifty percent rate cited at most colleges-—particularly those with no honor system. The honor system is not perfect but, according to this com— parison alone, it is certainly more effective than the proctor system. Another criticism is that, while the honor sys— tem may work in certain small private colleges, itis not practical when applied to larger state univer- sities. Yet, do you consider Dartmouth College, the University of North Carolina, the University of Texas, or the University of California small colleges? Of course not, and these universities have honor systems! Finally, the honor system is accused of being fine in theory but unworkable in practice. It can be seen, however, that the honor system not only can but does work. The honor system is no overnight paneca to elim- inating cheating but one basic fact must be emphasized. The honor system is effective in curbing cheating be— cause it establishes an atmosphere of trust and confidence. The individual responsibility and integrity of the stu— dent are given a chance. Even those skeptical at first 149 of the honor system have said it should be given a fair trial. Now, I say to you. Give it a fair trial. AdOpt the honor system! B. LOW SELF—REFERENCE MESSAGEl Consider, if you will, the following statement from the College of William and Mary Quarterly. "While nothing is required of a college student but attention to his studies, nothing is forbidden but dueling which might be fatal to his life, and gambling and drunken- nuss and tavern haunting, which must be pernicious to his health and morals." This unusual statement was an attempt to formulate a system of student conduct, a code of honor, at that college in the late 1800's. It is generally held that the first attempt in American colleges to phrase a code of honor for students was inaugurated by Thomas Jefferson in the founding of the University of Virginia in 1825. Through the years codes of student conduct have taken on many forms and have been applied to colleges of all classifications. At Barnard, principles of student conduct have pertained only to quizzes and written reports. At Beloit College in Wisconsin, a code has been set up for the student use of reserve books in the college library. All stu— dent activity fell under a system of honor at South Carolina while, as we have seen, only moral conduct was lApproximate message length - eleven minutes and twenty—five seconds. 150 151 included at William and Mary. On the other hand, an honor code applied to all rules made by the student association at Wilson College in Pennsylvania. These colleges illustrate the unusual and many diverse at- tempts to set down principles of student conduct. They are not, however, typical of what we know as the honor system today. For an idea of the form of the honor sys- tem used most widely, place yourself in the following situation. This semester you, along with thousands of Other college students, face examinations in your various courses of study. If you were to walk into your next exam and find no proctor present, no instructor watch- ing you during the examination to discourage and note any attempts at cheating,you in effect would be under the honor system. The use of unproctored teats is the most prevalent form of the honor system and is the topic which I would like to discuss with you today. Ag a former undergraduate college studentl I personally studied under both the honor system and the faculty- proctor system. During this gpeech, I will draw upon some of my first-hand experiences which I believe are pertinent to theypoint I am making in the speech.1 lSelf—reference by the communicator is underlined throughout the message. 152 You might ask why a college should have the honor system for examinations. The fundamental purpose for anysfietem Of this nature is to attempt to curb cheating. Academic cheating, in one form or another,has been with us through the ages. In the Oriental Library at Prince— ton University, an interesting artifact from ancient China, a cribbing shirt, is displayed. I guess this only proves that exam pressure was evident even then and that short cuts were employed to meet this pressure. What about the extent Of cheating in college tO— day?-iCOlumbia University, in February 1965, released the results Of what has been reported tO be the most comprehensive study ever made among college students. Nearly half Of the 5,000 students questioned, in strict confidence, at 99 colleges and universities admitted they had engaged in some form of cheating since enter- ing college. Nearly fifty percent cheat! We can all agree that is quite a staggering figure. Does this mean that today's college student is excessively immoral? It has been said that is not a fair accusation. The same Columbia study just cited found that ninety percent Of the students, including many who admitted to cheat- ing, said they were Opposed to the practice on moral grounds. This information would seem to cast doubt on the label Of "excessively immoral" being placed on col- lege students. The question then arises, if students 153 are basically Opposed to cheating, why do they do it. The environment at their particular colleges might be a key to understanding this problem. Within the content Of the environment and pur— pose Of a university, let's examine the two systems aimed at eliminating the incidence Of cheating on ex— aminations——the proctor system and the honor system. To seek, guard, and speak for the truth are considered the duties of any institution Of higher learning. We would hope college seeks,through the exchange between students and faculty, not merely tO convey knowledge and gain skills, but also to develOp understandings, insights, and character. How do the proctor and honor systems fit in with this philosophy Of a university? The theory behind the proctor system for exam- inations is that, with the instructor watching you, the threat Of his detecting cheating on your part will discourage you from cheating. How does the proctOr system work in practice? Let's look at the examination situation in our attempt to answer this question. Don't we Often think of an examination as a matching Of wits between student and instructor? In effect, through the material on the test, the instructor is Offering a challenge to the student and the student attempts to meet this challenge. This is a healthy situation. However, now add the proctor to the exam— ination. Hasn't the area Of challenge broadened? 154 Since the instructor is present to detect cheating, isn't his physical presence an invitation to students to at— tempt tO "beat the system?" Such a system does not help tO maintain an environment where a free exchange between student and faculty can truly take place. The theory Of the honor system is based on the premise that to attack the problem Of cheating one must foster an atmosphere of learning in which cheating is impractical. In other words, the university should develop a setting in which it is expected that cheating will not take place. The honor system simply means no instructor is obliged to be present during examinations. I remember the day we had a visitor to the class I was attending who wished to see our library. Our class was taking its final exam, but our instructor said he would go with him. I recall the lOOk Of amazement on the visitor's face upon hearing that our instructor could leave his class during an examination. It can be seen that with the faculty—proctor absent, one Of the motives for cheating is removed. There is no longer the same challenge to try to outsmart the instructor. Some Of you may say that some students will still attempt to cheat. However, a key factor is working under the honor system which is not present under the proctor system. Students taking examinations with aproctor depend upon him to deal with cheating. Students may assume the 155 attitude of dissociating themselves from the problem. Let the instructor worry about it! It is his problem! The honor system places cheating in a different light. Suddenly, the students as a group and as individuals have the responsibility placed on their shoulders. The student who may have cheated under the proctor sys- tem now has second thoughts. He is more reluctant, whether for personal reasons or a desire not to gO against his peers. Gradually, it becomes a personal Obligation on the part of each student to see that fairness is maintained. The question inevitably arises: Does the honor system actually curb cheating? It is fine in theory but do students take it seriously. It is possible the faculty at first may doubt whether the students can do much better than themselves but after awhile they will likely say that,though there are problems, the situation has improved. There is considerable evidence from students and faculty alike that the honor system is effective. We can see results at a specific univer— sity. A five—year study was conducted prior, during, and after the adoption Of the honor system at Brigham Young University. It was discovered that rates Of cheating in the classes were reduced by nearly two— thirds. While these findings do not necessarily prove a direct causal relationship between the honor system 156 and a reduction of cheating, the liklihood is certainly present. Some of you may question the typicalness of such a specific instance. Yet, the study of 99 colleges by Columbia University cited earlier, noted that cheat— ing is most evident at schools where the faculty alone tries to control the problem. It occurs far less Often at colleges with an honor system. Thus, it would seem reasonable to conclude that the honor system is ef— fective in achieving one Of its primary purposes-~c0p— ing with the problem Of cheating. In addition to being a workable solution, the honor system has other distinct advantages. Its great— est benefit is that it is educational as well as reg- ulatory. Shouldn't a university help to create an environment which will be most favorable to developing responsible individuals? What about the proctor system for examinations? In the most favorable setting, it is a kind Of supervision. Most Of the time it is un— fortunately an inspection system and, at its worst, it is a policing system. Is this the best our colleges and universities can do for us? The development Of responsibility and character should not stop at an inspection or police stage. I remember feeling con— fident after taking mnyhOdes Scholar examinations thinkingyto myself: The honor system gave me an op— portunity tO do my best. We all do our best when we 157 really believe somebody has faith.in us. It's the in- difference under theyproctor system that makes ration— alization for cheating so much easier. It should be possible for you, as college students,to assume the responsibility Of maintaining fairness in examinations. Now it is important to realize that the honor system is nOt without criticism. The Air Force Academy incident Of 1965 is pointed to as an example Of the failure Of the honor system to curb cheating. But, do you realize that only four percent of all Of the cadets were implicated in the cheating incident. Compare that figure with the nearly fifty percent rate cited at most colleges-—particu1arly those with no honor system. The honor system is not perfect but, according to this comparison alone, it is certainly more effective than the proctor system. Another criticism is that, while the honor sys— tem may work in certain small private colleges, it is not practical when applied to larger state univer— sities. Yet, do you consider Dartmouth College, the University Of North Carolina, the University Of Texas, or the University Of California small colleges? Of course not, and these universities have honor systems! Finally, the honor system is accused Of being fine in theory but unworkable in practice. It can be seen, however, that the honor system not only can but does work. 158 The honor system is no overnight paneca to elim- inating cheating but one basic fact must be emphasized. The honor system is effective in curbing cheating be- cause it establishes an atmosphere Of trust and con— fidence. The individual responsibility and integrity Of the student are given a chance. Even those skeptical at first Of the honor system have said it should be given a fair trial. Now, I say to you. Give it a fair trial. AdOpt the honor system! c. MODERATE SELF—REFERENCE MESSAGEl Consider, if you will, the following statement from the College Of William and Mary Quarterly, "While nothing is required Of a college student but attention tO his studies, nothing is forbidden but dueling which might be fatal to his life, and gambling and drunken— FA ness and tavern haunting, which must pernicious to his health and morals." This unusual statement was an at- tempt tO formulate a system Of student conduct, a code 2---...— I of honor, at that college in the late 1800's. It is generally held that the first attempt in American colleges to phrase a code Of honor for students was inaugurated by Thomas Jefferson in the founding of the University of Virginia in 1825. Through the years codes of student conduct have taken on many forms and have been applied tO colleges of all classifications. At Barnard, principles of student conduct have per— tained only tO quizzes and written reports. At Beloit College in Wisconsin, a code has been set up for the student use Of reserve books in the college library. All student activity fell under a system Of honor at South Carolina while, as we have seen, only moral con- duct was included at William and Mary. On the other hand, an honor code applied to all rules made by the lApproximate message length — twelve minutes and five seconds. 159 160 student association at Wilson College in Pennsylvania. These colleges illustrate the unusual and many diverse attempts to set down principles of student conduct. They are not,however, typical Of what we know as the honor system today. For an idea Of the form Of the honor system used most widely, place yourself in the following situation. FR This semester you, along with thousands of other college students, face examinations in your various courses Of study. If you were to walk into your next in exam and find no proctor present, no instructor watch- ing you during the examination to discourage and note any attempts at cheating, you in effect would be under the honor system. The use Of unproctored tests is the most prevelant form Of the honor system and is the topic which I would like to discuss with you today. As a former undergraduate college student, I personally studied under both the honor system and the faculty: proctor sygtem. Duripg this speech, I will draw upon some of my first-hand experiences which I believe are pertinent to the point I am making in the speech.1 You might ask why a college should have the honor system for examinations. The fundamental purpose for any system Of this nature is to attempt to curb cheating. a.;.——‘( V, lSelf-reference by the communicator is underlined throughout the message. 161 Academic cheating, in one form or another, has been with us through the ages. In the Oriental Library at Princeton University, an interesting artifact from ancient China, a cribbing shirt, is displayed. I guess this only proves that exam presure was evident eventhen and that short cuts were employed to meet this pressure. What about the extent of cheating in college today? Columbia University, in February 1965, released the results Of what has been reported to be the most comprehensive study ever made among college students. Nearly half of the 5,000 students questioned, in strict confidence, at 99 colleges and universities admitted they had engaged in some form Of cheating since enter— ing college. Nearly fifty percent cheat! We can all agree that is quite a staggering figure. Does this mean that today's college student is excessively immoral? I was fortunate as a college student to have served on a faculty-student committee which studied cheating at New York Universities. From what I was able to learn from this valuable personal experience, I do not be- lieve that is a fair accusation. The same Columbia study just cited found that ninety percent Of the stu- dents, including many who admitted tO cheating, said they were Opposed to the practice on moral grounds. This information would seem to cast doubt on the label 162 Of "excessively immoral" being placed on college stu— dents. The question then arises, if students are basically Opposed to cheating, why do they do it. The environment at their particular colleges might be a key to understanding this problem. Within the context of the environment and pur- pose Of a university, let's examine the two systems aimed at eliminating the incidence Of cheating on ex- aminations--the proctor system and the honor system. TO seek, guard, and speak for the truth are considered the duties Of any institution Of higher learning. We would hope college seeks, through the exchange between Students and faculty, not merely tO convey knowledge and gain skills, but also to develop understandings, insights, and character. How do the proctor and honor systems fit in with this philOSOphy of a university? The theory behind the proctor system for ex- aminations is that, with the instructor watching you, the threat Of his detecting cheating on your part will discourage you from cheating. How does the proctor system work in practice? Let's look at the examination situation in our attempt to answer this question. Don't we Oftenthink Of an examination as a matching Of wits between student and instructor? In effect, through -the material on the test, the instructor is offering a challenge to the student and the student attempts to 163 meet this challenge. Thts is a healthy situation. How— ever, now add the proctor to the examination. Hasn't the area Of challenge broadened? Since the instructor is present to detect cheating, isn't his physical pres— ence an invitation to students to attempt to "beat the system?" Such a system does not help to maintain an environment where a free exchange between student and faculty can truly take place. The theory Of the honor system is based on the premise that to attack the problem Of cheating one -€ must foster an atmosphere Of learning in which cheat— ing is impractical. In other words, the university should develop a setting in which it is expected that cheating will not take place. The honor system simply means no instructor is obliged to be present during examinations. I remember the day we had a visitor tO the class I was attending who wished to see our library, Our class was taking its final examyybut our instructor said he would gO with him. I recall the lOOk Of amaze- ment on the visitor's face upon hearing that our in— structor could leave his class during an examination. It can be seen that with the faculty-proctor absent, one Of the motives for cheating is removed. There is no longer the same challenge to try to outsmart the instructor. Some of you may say that some students will still attempt to cheat. However, a key factor is 164 working under the honor system which is not present under the proctor system. Students taking examinations with a proctor depend upon him to deal with cheating. Students may assume the attitude Of dissociating themselves from the problem. Let the instructor worry about it! It is his problem! The honor system placed cheating in a P different light. Suddenly, the students as a group and as individuals have the responsibility placed on their shoulders. The student who may have cheated under the proctor system now has second thoughts. He is more H reluctant, whether for personal reasons or a desire not to go against his peers. Gradually, it becomes a personal Obligation on the part Of each student to see that fair— ness is maintained. The question inevitably arises: Does the honor system actually curb cheating? It is fine in theory but do students take it seriously. Thinking back on my own experiences under the honor system, I am reminded Of how the faculty, in the beginning; didn't think the students would do much better than faculty supervision; but now, after four years, the faculty holds the Opin- ion thaty though matters are not perfect, they have im- proved greatly. There is considerable evidence from students and faculty alike that the honor system is effective. We can see results at a specific university. A five-year study was conducted prior, during, and after the adOption Of the honor system at Brigham Young 165 University. It was discovered that rates of cheating in the classes were reduced by nearly two-thirds. While these findings do not necessarily prove a direct causal relationship between the honor system and a reduction of cheating, the liklihood is certainly present. Some of you may question the typicalness Of such a specific instance. Yet, thesxudy of 99 colleges by Columbia University cited earlier, noted that cheating is most evident at schools where the faculty alone tries to control the problem. It occurs far less Often at col- leges with an honor system. Thus, it would seem reason- able tO conclude that the honor system is effective in achieving one Of its primary purposes--coping with the problem Of cheating. In addition tO being a workable solution, the honor system has other distinct advantages. Its great— est benefit is that it is educational as well as reg- ulatory. Shouldn't a university help to create an environment which will be most favorable to developing responsible individuals? What about the proctor sys— tem for examinations? In the most favorable setting, it is a kind of supervision. Most Of the time it is unfortunately an inspection system and, at its worst, it is a policing system. Is this the best our colleges and universities can do for us? The development Of responsibility and character should not stop at an 166 inspection or police stage. I remember feelingyconfident after taking my Rhodes Scholar examinations thinking to myself: The honor system gave me an Opportunity to do my best. We all do our best when we really believe some- body has faith in us. It's the indifference under the proctor system that makes rationalization for cheating so much easier. It should be possible for you, as col— lege students, to assume the responsibility Of maintain— ing fairness in examinations. Now it is important to realize that the honor system is not without criticism. The Air Force Academy incident Of 1965 is pointed to as an example Of the failure Of the honor system to curb cheating. But, do you realize that only four percent of all Of the cadets were implicated in the cheating incident. Compare that figure with the nearly fifty percent rate cited at most colleges-—particularly those with no honor system. The honor system is not perfect but, according to this com- parison alone, it is certainly more effective than the proctor system. Another criticism is that, while the honor system may work in certain small private colleges, it is not practical when applied to larger state universities. Yet, do you consider Dartmouth College, the University of North Carolina, the University Of Texas, or the Un- iversity Of California small colleges? Of course not, and these universities have honor systems! 167 Finally, the honor system is accused of being fine in theory but unworkable in practice. However, I hope my first-hand experience with both the proctor, system and the honor system, to which I referred during this speech, have shown you that the honor system not only can but does work. The honor system is nO overnight paneca tO elim- inating cheating but one basic fact must be emphasized. The honor system is effective in curbing cheating be- cause it establishes an atmosphere Of trust and confid- ence. The individual responsibility and integrity Of the student are given a chance. Even those skeptical at first Of the honor system have said it should be given a fair trial. Now, I say tO you. Give it a fair trial. AdOpt the honor system! D. HIGH SELF—REFERENCE MESSAGEl Consider, if you will, the following statement from the College Of William and Mary Quarterly, "While nothing is required Of a college student but attention to his studies, nothing is forbidden but dueling which #1 might be fatal to his life, and gambling and drunken- ness and tavern haunting, which must be pernicious to his health and morals." This unusual statement was an .‘I-.' ‘a '1 ‘1 attempt to formulate a system Of student conduct, a code Of honor, at that college in the late 1800's. It is generally held that the first attempt in American colleges to phrase a code Of honor for students was inaugurated by Thomas Jefferson in the founding of the University Of Virginia in 1825. Through the years codes Of student conduct have taken on many forms and have been applied to colleges of all classifications. At Barnard, principles Of student conduct have per- tained only to quizzes and written reports. At Beloit College in Wisconsin, a code has been set up for the student'use Of reserve books in the college library. All student activity fell under a system Of honor at South Carolina while, as we have seen, only moral 1Approximate message length — twelve minutes and twenty-five seconds. 168 169 conduct was included at William and Mary. On the other hand, an honor code applied to all rules made by the student association at Wilson College in Pennsylvania. These colleges illustrate the unusual and many diverse attempts to set down principles Of student conduct. They are not, however, typical Of what we know as the a honor system today. For an idea Of the form of the honor system used most widely, place yourself in the following situation. This semester you, along with thousands Of other L college students, face examinations in your various courses Of study. If you were to walk into your next exam and find no proctor present, no instructor watch— ing you during the examination to discourage and note any attempts at cheating, you in effect would be under the honor system. The use Of unproctored tests is the most prevalent form Of the honor system and is the topic which I would like to discuss with you today. As a former undergraduate college student, I personally studied under both the honor system and the faculty- proctor system. During this speechy I will draw upon some Of my first-hand experiences which I believe are pertinent to the point I am making in the speech. Self—reference by the communicator is under- lined throughout the message. 170 You might ask why a college should have the honor system for examinations. The fundamental purpose for any system of this nature is to attempt tO curb cheating. Academic cheating, in one form or another, has been with us through theages. In the Oriental Library at Princeton University, an interesting arti- fact from ancient China, a cribbing shirt, is dis- played. I guess this only proves that exam pressure was evident even then and that short cuts were employed to meet this pressure. What about the extent Of cheating in college today? Columbia University, in February 1965, released the results Of what has been reported to be the most comprehensive study ever made among college students. Nearly half Of the 5,000 students questioned, in strict confidence, at 99 colleges and universities admitted they had engaged in some form Of cheating since enter— ing college. Nearly fifty percent cheat! We can all agree that is quite a staggering figure. Does this mean that today's college student is excessively im— moral? I was fortunate as a college student to have served on a faculty:student committee which studied cheating at New York State Universities. From what I was able to learn from this valuable personal exper— ience, I do not believe that is a fair accusation. The 171 same Columbia study just cited found that ninety per- cent Of the students, including many who admitted to cheating,said they were Opposed to the practice on moral grounds. This information would seem to cast doubt on the label Of "excessively immoral" being placed on college students. The question then arises, if stu— dents are basically Opposed to cheating, why dO they do it. From my interviews with students for the New York State Universities study on cheating, my impres— sion is the main hope of most students is to be given a fair chance of showing their best. I have come to the conclusion that many students are stronglyyinfluenced byytheir college environment in matters of college honesty. Within the context Of the environment and pur- pose Of a university, let's examine the twO systems aimed at eliminating the incidence Of cheathrgon exam— inations—-the proctor system and the honor system. TO seek, guard, and speak for the truth are considered the duties Of any institution of higher learning. We would hope college seeks, through the exchange be— tween students and faculty, not merely to convey know- ledge and gain skills, but also to develop under- standings, insights,and character. How do the proctor and honor systems fit in with this philosophy Of a university? 172 The theory behind the proctor system for exam— inations is that, with the instructor watching you, the threat of his detecting cheating on your part will dis— courage you from cheating. How does the proctor sys- tem work in practice? Let's lOOk at the examination situation in our attempt to answer this question. Don't h we often think Of an examination as a matching of wits between student and instructor? In effect, through the material on the test, the instructor is Offering a chal— lenge tO the student and the student attempts to meet this challenge. This is a healthy situation. However, now add the proctor to the examination. Hasn't the area of challenge broadened? Since the instructor is present to detect cheating, isn't his physical presence an invitation tO students to attempt to "beat the sys— tem?" Such a system does not help to maintain an en- vironment where a free exchange between student and faculty can truly take place. The theory Of the honor system is based on the premise that tO attack the problem Of cheating one must foster an atmosphere Of learning in which cheating is impractical. In other words, the university should develop a setting in which it is expected that cheat— ing will not take place. The honor system simply means no instructor is obliged tO be present during 173 examinations. I remember the day we had a visitor to the class I was attending who wished to see our library. Our class was taking its final exam, but our instructor said he would go with him. I recall the look of amaze- ment on the visitor's face upon hearing that our in— structor could leave his class during an examination. It can be seen that with the faculty—proctor absent, one of the motives for cheating is removed. There is nO longer the same challenge to try to outsmart the instructor. Some of you may say that some students will still attempt to cheat. However, a key factor is working under the honor system which is not present under the proctor system. Students taking examinations with a proctor depend upon him to deal with cheating. Students may assume the attitude of dissociating them— selves from the problem. Let the instructor worry about it! It is his problem! The honor system places cheat— ing in a different light. Suddenly, the students as a group and as individuals have the responsibility placed ontheir shoulders. The student who may have cheated under the proctor system now has second thoughts. He is more reluctant, whether for personal reasons or a desire not to go against his peers. Havingyknown familiarly a campus with a working honor system, I have been in classrooms where the student merely in— sisted that there be fairness in the examination room whether the professor was present or not. 174 The question inevitably arises: Does the honor system actually curb cheating? It is fine in theory but do students take it seriously. Thinking back on my own ex- periences under the honor system, I am reminded of how the faculty, in the beginning, didn't think the stu— dents would do much better than faculty supervision; Ek but now, after four yearsy the faculty holds the Opinion that, though matters are not perfect, they have improved greatly. There is considerable evidence from students and faculty alike that the honor system is effective. at We can see results at a specific university. A five— year study was conducted prior, during, and after the adoption Of the honor system at Brigham Young University. It was discovered that rates Of cheating in the classes were reduced by nearly two—thirds. While these findings do not necessarily prove a direct causal relationship between the honor system and a reduction of cheating, the liklihood is certainly present. Some of you may question the typicalness Of such a specific instance. Yet, the study of 99 colleges by Columbia University cited earlier, noted that cheating is most evident at schools where the faculty alone tries to control the problem. It occurs far less Often at colleges with an honor system. Thus, it would seem reasonable to con— clude that the honor system is effective in achieving one of its primary purposes—-coping with the problem of cheating. 175 In addition to being a workable solution, the honor system has other distinct advantages. Its great— est benefit is that it is educational as well as reg— ulatory. Shouldn't a university help to create an en- vironment which will be most favorable to developing responsible individuals? What about the proctor system for examinations? In the most favorable setting, it is a kind of supervision. Most of the time it is unfort- unately an inspection system and, at its worst, it is a policing system. Is this the best our colleges and universities can do for us? The development of respons- ibility and character should not stOp at an inspection or police stage. I remember feeling confident afrer taking myrRhOdes Scholar examinations thinking to my— self: The honor system gave me an Opportunity to do my best. We all do our best when we really believe some- body has faith in us. It's the indifference under the proctor system that makes rationalization for cheating so much easier. It should be possible for you, as col- lege students,to assume the responsibility Of maintain— ing fairness in examinations. Now it is important to realize that the honor system is not without criticism. The Air Force Academy incident Of 1965 is pointed to as an example Of the failure of the honor system to curb cheating. But, do you realize that only four percent of all Of the cadets 176 were implicated in the cheating incident. Compare that figure with the nearly fifty percent rate cited at most colleges-—particularly those with no honor system. The honor system is not perfect but, according to this com- parison alone, it is certainly more effective than the proctor system. F Another criticism is that, while the honor system I may work in certain small private colleges, it is not practical when applied to larger state universities. i Yet, do you consider Dartmouth College, the University H Of North Carolina, the University of Texas, or the Un— iversity of California small colleges? Of course not, and these universities have honor systems! Finally, the honor system is accused Of being fine in theory but unworkable in practice. However, I hOpe myrfirst—hand erperiences with both the proctor system and the honor system, to which I referred during this speech, have shown you that the honor system not only can but does work. The honor system is no Overnight paneca to elim- inating cheating but one basic fact must be emphasized. The honor system is effective in curbing cheating be- cause it establishes an atmosphere of trust and confid- ence. The individual responsibility and integrity of the student are given a chance. I don't think I will 177 ever forget the year, when the university I was attend- ing, considered the adoption of the honor system. Al- though somewhat skeptical Of it in the beginning, I felt it should be given a fair trial. Belive me, I an now glad I took that position. Now, I say to you. Give it a fair trial. AdOpt the honor system! IR E. LOW PRESTIGE—REFERENCE MESSAGEl Consider, if you will, the following statement from the College Of William and Mary Quarterly. "While nothing is required of a college student but attention to his studies, nothing is forbidden but dueling which IR might be fatal to his life, and gambling and drunken- ness and tavern haunting, which must be pernicious to his health and morals." This unusual statement was an attempt to formulate a system Of student conduct, a code of honor, at that college in the late 1800's. It is generally held that the first attempt in American colleges to phrase a code Of honor for stu- dents was inaugurated by Thomas Jefferson in the found— ing of the University of Virginia in 1825. Through the years codes of student conduct have taken on many forms and have been applied to colleges Of all classifications. At Barnard, principles of student conduct have per— tained only to quizzes and written reports. At Beloit College in Wisconsin, a code has been set up for the student use of reserve books in the college library. All student activity fell under a system Of honor at South Carolina while, as we have seen, only moral conduct was included at William and Mary. On the other hand, lApproximate message length — eleven minutes and forty-five seconds. 178 179 an honor code applied to all rules made by the stu— dent association at Wilson College in Pennsylvania. These colleges illustrate the unusual and many diverse attempts to set down principles Of student conduct. They are not, however, typical of what we know as the honor system today. For an idea Of the form Of the honor system used most widely, place yourself in the Pi. following situation. This semester you, along with thousands of other college students, face examinations in your various courses of study. If you were to walk into your next exam and find no proctor present, no instructor watching you during the examination to discourage and note any attempts at cheating, you in effect would be under the honor system. The use of unproctored tests is the most prevalent form of the honor system and is the topic which I would like to discuss with you today. A personal friend of my famiry's, Professor William Canning of Stanford University, has taught for twenty years at five different colleges, under both the honor system and faculty:proctor system. During this speech, I will draw gpon some of his first-hand experiences as well as the first—hand experiences of others I have,personally known.l lPrestige—reference by the communicator is under- lined throughout the message. 180 You might ask why a college should have the honor system for examinations. The fundamental purpose for any system of this nature is to attempt to curb cheat- ing. Academic cheating, in one form or another, has been with us through the ages. In the Oriental Library at Princeton University, an interesting artifact from !\ ancient China, a cribbing shirt, is displayed. I guess this only proves that exam pressure was evident even then and that short cuts were employed to meet this pres— 5 “F! 7,; ._. sure. What about the extent Of cheating in college today? Columbia University, in February 1965, released the results of what has been reported to be the most comprehensive study ever made among college students. Nearly half of the 5,000 students questioned, in strict confidence, at 99 colleges and universities admitted they had engaged in some form of cheating since enter- ing college. Nearly fifty percent cheat! We can all agree that is quite a staggering figure. Does this mean that today's college student is excessively im- moral? It has been said that is not a fair accusation. The same Columbia study just cited found that ninety percent of the students, including many who admitted to cheating, said they were Opposed to the practice on moral grounds. This information would seem to cast doubt on the label Of "excessively immoral" being 181 placed on college students. The question then arises, if students are basically Opposed to cheating, why do they do it. The environment at their particular col- 1eges might be a key to understanding this problem. Within the context of the environment and pur- pose Of a university, let's examine the two systems aimed at eliminating the incidence of cheating on exam- inations——the proctor system and the honor system. To seek, guard, and speak for the truth are considered the duties Of any institution of higher learning. We g would hope college seeks,through the exchange between students and faculty, not merely to convey knowledge and gain skills, but also to develOp understandings, insights, and character. How do the proctor and honor systems fit in with this philosophy of a university? The theory behind the proctor system for examin- ations is that, with the instructor watching you, the threat Of his detecting cheating on your part will dis— courage you from cheating. How does the proctor system work in practice? Let's look at the examination situa— tion in our attempt to answer this question. Don't we Often think of an examination as a matching Of wits between student and instructor? In effect, through the material on the test, the instructor is offering a challenge to the student and the student attempts to meet this challenge. This is a healthy situation. 182 However, now add the proctor to the examination. Hasn't the area of challenge broadened? Since the instructor is present to detect cheating, isn't his physical pres— ence an invitation to students to attempt to "beat the system?" Such a system does not help to maintain an environment where a free exchange between student and fix“ faculty can truly take place. The theory Of the honor system islfised on the premise that to attack the problem of cheating one must foster an atmosphere of learning in which cheating is impractical. In other words, the university should develop a setting in which it is expected that cheating will not take place. The honor system simply means no instructor is obliged to be present during examinations. My good friend, Dr. Canning, once spoke to me of the day he had a visitor to class who wished to see our library. Professor Canning was conductingythe final exam of the class but said he wouldrgo with him. He said the visitor showed amazement that he could leave class during an examination. It can be seen that with the faculty— proctor absent, one Of the motives for cheating is removed. There is no longer the same challenge to try to outsmart the instructor. Some of you may say that some students will still attempt to cheat. However, a key factor is working under the honor system which is not present under the proctor system. Students taking 183 examinations with a proctor depend upon him to deal with cheating. Students may assume the attitude Of dissociat- ing themselves from the problem. Let the instructor worry about it! It is his problem! The honor system places cheating in a different light. Suddenly, the students as a group and as individuals have the respons— ibility placed on their shoulders. The student who may have cheated under the proctor system now has second thoughts. He is more reluctant, whether for personal reasons or a desire not to go against his peers. Grad- ually, it becomes a personal Obligation on the part Of each student to see that fairness is maintained. The question inevitably arises: Does the honor system actually curb cheating? It is fine in theory but do students take it seriously. It is possible the faculty at first may doubt whether the students can do much better than themsleves but after awhile they will likely say that, though there are problems, the situation has improved. There is considerable evidence from students and faculty alike that the honor system is effective. We can see results at a specific univer- sity. A five-year study was conducted prior, during, and after the adoption Of the honor system at Brigham Young University. It was discovered that rates of cheating in the classes were reduced by nearly two-thirds. While these findings do not necessarily prove a direct 184 causal relationship between the honor system and a reduction Of cheating, the liklihood is certainly pres- ent. Some of you may question the typicalness Of such a specific instance. Yet, the study Of 99 colleges by Columbia University cited earlier, noted that cheating is most evident at schools where the faculty alone tries to control the problem. It occurs far less Often at colleges with an honor system. Thus, it would seem reasonable to conclude that the honor system is effective in achieving one of its primary purposes——coping with the problem Of cheating. In addition to being a workable solution, the honor system has other distinct advantages. Its great- est benefit is that it is educational as well as reg- ulatory. Shouldn't a university help to create an environment which will be most favorable to develOping responsible individuals? What about the proctor sys- tem for examinations? In the most favorable setting, it is a kind of supervision. Most of the time it is unfortunately an inspection system and, at its worst, it is a policing system. Is this the best our col- leges and universities can do for us? The development of responsibility and character should not stop at an inspection or police stage. A fraternity brother of mine, feeling confident after taking his Rhodes Scholar examinations, told me: "The honor system gave me an 185 opportunity to do my best. We all do our best when we really believe somebody has faith in us. It's the in— difference under the proctor system that makes rationalization for cheating so much easier." It should be possible for you, as college students, to assume the respons- ibility Of maintaining fairness in examinations. Now it is important to realize that the honor system is not without criticism. The Air Force Academy incident of 1965 is pointed to as an example of the failure Of the honor system to curb cheating. But, do you realize that only four percent of all of the cadets were implicated in the cheating incident. Com- pare that figure with the nearly fifty percent rate cited at most colleges——particularly those with no honor system. The honor system is not perfect but, according to this comparison alone, it is certainly more effective than the proctor system. Another criticism is that, while the honor sys- temrmgzwork in certain small private colleges, it is not practical when applied to larger state universities. Yet, do you consider Dartmouth College, the University Of North Carolina, the University of Texas, or the University of California small colleges? Of course not, and these universities have honor systems! Finally, the honor system is accused of being fine in theory but unworkable in practice. It can be seen, however, that the honor system not only can but does work. 186 The honor system is no overnight paneca to elim- inating cheating but one basic fact must be emphasized. The honor system is effective in curbing cheating because it establishes an atmosphere of trust and confidence. The individual responsibility and integrity of the stu— dent are given a chance. Even those skeptical at first Of the honor system have said it should be given a fair trial. Now, I say to you. Give it a fair trial. Adopt the honor system! F. MODERATE PRESTIGE—REFERENCE MESSAGEl Consider, if you will, the following statement from the College of William and Mary Quarterly. "While nothing is required of a college student but attention to his studies, nothing is forbidden but dueling which E might be fatal to his life, and gambling and drunken- ness and tavern haunting, which must be pernicious to his health and morals." This unusual statement was an E attempt to formulate a system of student conduct, a code Of honor, at that college in the late 1800's. It is generally held that the first attempt in American colleges to phrase a code Of honor for stu— dents was inaugurated by Thomas Jefferson in the found- ing of the University Of Virginia in 1825. Through the years codes of student conduct have taken on many forms and have been applied to colleges of all classifications. At Barnard, principles Of student conduct have per— tained only to quizzes and written reports. At Beloit College in Wisconsin, a code has been set up for the student use of reserve books in the college library. All student activity fell under a system of honor at South Carolina while, as we have seen, only moral con- duct was included at William and Mary. On the other lApproximate message length - twelve minutes and five seconds. 187 188 hand, an honor code applied to all rules made by the student association at Wilson College in Pennsylvania. These colleges illustrate the unusual and many diverse attempts to set down principles Of student conduct. They are not, however, typical of what we know as the honor system today. For an idea of the form of the honor system used most widely, place yourself in the following situation. This semester you, along with thousands of other college students, face examinations in your various courses of study. If you were to walk into your next exam and find no proctor present, no instructor watch- ing you during the examination to discourage and note any attempts at cheating, you in effect would be under the honor system. The use of unproctored tests is the most prevalent form Of the honor system and is the tOpic which I would like to discuss with you today. A per- sonal friend of my family's, Professor William Canning of Stanford University, has taught for twenty years at five different colleges, under both the honor sys- tem and faculty-proctor system. During this speech, I will draw upon some of his first-hand experiences as well as the first—hand experiences of others I have per- sonallyknown.l l Prestige—reference by the communicator is under— lined throughout the message. 189 You might ask why a college should have the honor system for examinations. The fundamental purpose for any system of this nature is to attempt to curb cheat- ing. Academic cheating, in one form or another, has been with us through the ages. In the Oriental Library at Princeton University, an interesting artifact from ancient China, a cribbing shirt, is displayed. I guess this only proves that exam presure was evident even then and that short cuts were employed to meet this pressure. What about the extent of cheating in college to— day? Columbia University, in February 1965, released the results Of what has been reported to be the most comprehensive study ever made among college students. Nearly half Of the 5,000 students questioned, in strict confidence, at 99 colleges and universities admitted they had engaged in some form Of cheating since enter- ing college. Nearly fifty percent cheat! We can all agree that is quite a staggering figure. Does this mean that today's college student is excessively immoral? My major advisor in collegeJ Dr. Graham Harrington, served on a faculty—student committee which studied cheating at New York State universities. I know him very well and respect his opinion. In a conference, he once said to me that based on his experience that is not a fair accusation. The same Columbia study just cited found that ninety percent Of the students, includ- ing many who admitted to cheating, said they were Opposed 190 to the practice on moral grounds. This information would seem to cast doubt on the label Of "excessively immoral" being placed on college students. The question then arises, if students are basically Opposed to cheating, why do they do it. The environment at their particular colleges might be a key to understanding this problem. :1 Within the context of the environment and purpose I Of a university, let's examine the two systems aimed at eliminating the incidence of cheating on examinations—- I the proctor system and the honor system. To seek, guard, L4 and speak for the truth are considered the duties Of any institution of higher learning. We would hope college seeks, through the exchange between students and faculty, not merely to convey knowledge and gain skills, but also to develop understandings, insights, and character. How do the proctor and honor systems fit in with this phil— osophy Of a university? The theory behind the proctor system for examin- ations is that, with the instructor watching you, the threat Of his detecting cheating on your part will dis— courage you from cheating. How does the proctor system work in practice? Let's look at the examination situa— tion in our attempt to answer this question. Don't we often think of an examination as a matching Of wits between student and instructor? In effect, through the material on the test, the instructor is offering a 191 challenge to the student and the student attempts to meet this challenge. This is a healthy situation. However, now add the proctor to the examination. Hasn't the area of challenge broadened? Since the instructor is present to detect cheating, isn't his physical presence an invitation to students to attempt to "beat the sys— T! tem?" Such a system does not help to maintain an en- vironment where a free exchange between student and fac— ulty can truly take place. 1mm: The theory Of the honor system is based on the premise that to attack the problem of cheating one must foster an atmosphere Of learning in which cheating is impractical. In other words, the university should develop a setting in which it is expected that cheating will not take place. The honor system simply means no instructor is obliged to be present during examinations. My good friend, Dr. Canning, once spoke to me of the day he had a visitor to class who wished to see our library. Professor Canning was conducting_the final exam of the class but said he would go with him. He said the visitor showed amazement that he could leave class during an examination. It can be seen that with the faculty—proctor absent, one of the motives for cheating is removed. There is no longer the same chal- lenge to try to outsmart the instructor. Some of you Inay say that some students will still attempt to cheat. 192 However, a key factor is working under the honor system which is not present under the proctor system. Students taking examinations with a proctor depend upon him to deal with cheating. Students may assume the attitude of dissociating themselves from the problem. Let the instructor worry about it! It is his problem! The honor system places cheating in a different light. Sud— denly, the students as a group and as individuals have the responsibility placed on their shoulders. The stu- dent who may have cheated under the proctor system now has second thoughts. He is more reluctant, whether for personal reasons or a desire not to go against his peers. Gradually, it becomes a personal Obligation on the part of each student to see that fairness is maintained. The question inevitably arises: Does the honor system actually curb cheating? It is fine in theory but do students take it seriously. A former professor of mine, Dean Carl H. Fine, who is now at the Univer— sity of Michigan, told me that the faculty, in the be— ginning, didn't think the students would do much better than faculty supervisioni,but, after four years, the faculty holds the gpinion thatJ though matters are not perfect,ythey have improved greatly. There is consider- able evidence from students and faculty alike that the honor system is effective. We can see results at a specific university. A five-year study was conducted prior,chring, and after the adoptioncf the honor system 193 at Brigham Young University. It was discovered that rates Of cheating in the classes were reduced by nearly two-thirds. While these findings do not necessarily prove a direct causal relationship between the honor system and a reduction Of cheating, the liklihood is certainly present. Some of you may question the typical— FL ness Of such a specific instance. Yet, the study of 99 colleges by Columbia University cited earlier, noted that cheating is most evident at schools where the _[ faculty alone tries to control the problem. It occurs 1 far less Often at colleges with an honor system. Thus, it would seem reasonable to conclude that the honor sys- tem is effective in achieving one of its primary pur- poses-—c0ping with the problem Of cheating. In addition to being a workable solution, the honor system has other distinct advantages. Its greatest bene— fit is that it is educational as well as regulatory. Shouldn't a university help to create an environment which will be most favorable to developing responsible individuals? What about the proctor system for exam— inations? In the most favorable setting, it is a kind of supervision. Most Of the time it is unfortunately an inspection system and,at its worst, it is a policing system. Is this the best our colleges and universities can do for us? The develOpment Of responsibility and 194 character should not stOp at an inspection or police stage. A fraternity brother Of mine, feelipg confident after taking his Rhodes Scholar examinations, told me: "The honor system gave me an opportunity to do my best. We all do our best when we really believe somebody has faith in us. It's the indifference under the proctor system that makes rationalization for cheating so much a! eaSier." It should be possible for you, as college students, to assume the responsibility Of maintaining fairness in examinations. : Now it is important to realize that the honor system is not without criticism. The Air Force Acad- emy incident Of 1965 is pointed to as an example of the failure of the honor system to curb cheating. But, do you realize that only four percent Of all of the cadets were implicated in the cheating incident. Compare that figure with the nearly fifty percent rate cited at most colleges--particularly those with nO no honor system. The honor system is not perfect but, according to this comparison alone, it is certainly more effective than the proctor system. Another criticism is that, while the honor system may work in certain small private colleges, it is not practical when applied to larger state universities. Yet, do you consider Dartmouth College, the University of North Carolina, the University of Texas, or the 195 University Of California small colleges? Of course not, and these universities have honor systems! Finally, the honor system is accused of being fine in theory but unworkable in practice. However, I hope the first—hand experiences with the proctor system and honor system related during this speech by individ— uals I haveypersonally known, have shown that the honor system not only can but does work. The honor system is no overnight paneca to elim- inating cheating but one basic fact must be emphasized. ! The honor system is effective in curbing cheating be- I cause it establishes an atmosphere of trust and con- fidence. The individual responsibility and integrity Of the student are given a chance. Even those skept— ical at first of the honor system have said it should be given a fair trial. Now, I say to you. Give it a fair trial. Adopt the honor system! G. HIGH PRESTIGE-REFERENCE MESSAGEl Consider, if you will, the following statement from theCollege Of William and Mary Quarterly. "While nothing is required Of a college student but attention to his studies, nothing is forbidden but dueling which might be fatal tO his life, and gambling and drunken— ness and tavern haunting, which must be pernicious to his health and morals." This unusual statement was an attempt to formulate a system of student conduct, a code Of honor, at that college in the late 1800's. It is generally held that the first attempt in American colleges to phrase a code of honor for students was inaugurated by Thomas Jefferson in the founding of the University of Virginia in 1825. Through the years codes of student conduct have taken on many forms and have been applied to colleges Of all classifications. At Barnard, principles of student conduct have per- tained only to quizzes and written reports. At Beloit College in Wisconsin, a code has been set up for the student use of reserve books in the college library. All student activity fell under a system Of honor at South Carolina while, as we have seen, only moral con— duct was included at William and Mary. On the other lApproximate message length - twelve minutes and fifty seconds. 196 197 hand, an honor code applied to all rules made by the student association at Wilson College in Pennsylvania. These colleges illustrate the unusual and many diverse attempts to set down principles of student conduct. They are not, however, typical of what we know as the honor system today. For an idea Of the form of the honor system used most widely, place yourself in the lie following situation. This semester you, along with thousands of other college students, face examinations in your various courses of study. If you were to walk into your next exam and find no proctor present, no instructor watch- ing you during the examination to discourage and note any attempts at cheating, you in effect would be under the honor system. The use of unproctored tests is the most prevalent form of the honor system and is the topic which I would like to discuss with you today. A personal friend of my family's, Professor William Canning_of Stanford University, has taught for twenty years at five different colleges, under both the honor system and faculty-proctor system. During this speech, I will draw upon some of his first-hand experiences as well as the first—hand experiences of others I have personally known. l Prestige-reference by the communicator is under- lined throughout the message. 198 You might ask why a college should have the honor system for examinations. The fundamental pur— pose for any system of this nature is to attempt to curb cheating. Academic cheating, in one form or another, has been with us through the ages. In the Oriental Library at Princeton University, an interest- ing artifact from ancient China, a cribbing shirt, is displayed. I guess this only proves that exam pressure was evident even then and that short cuts were employed to meet this pressure. What about the extent of cheating in college today? Columbia University, in February 1965, released the results Of what has been reported to be the most comprehensive study ever made among college students. Nearly half Of the 5,000 students questioned, in strict confidence, at 99 colleges and universities admitted they had engaged in some form Of cheating since enter— ing college. Nearly fifty percent cheat! We can all agree that is quite a staggering figure. Does this mean that today's college student is excessively im— moral? My major advisor in college, Dr. Graham Harrington, served on a faculty—student committee which studied cheating at New York State univerSLties. I know him very well and respect his Opinion. In a conference, he once said to me that based on his experience that is not a fair accusation. The same Columbia study just cited 199 found that ninety percent of the students, including many who admitted to cheating, said they were Opposed to the practice on moral grounds. This information would seem to cast doubt on the label of "excessively immoral" being placed on college students. The question then arises, if students are basically Opposed to cheat- ing, why do they do it. From his interviews with stu— dents for the New York State universities study_on cheating, my advisor told me his impressflN1 was that the main hope of most students was to be given a fair chance of showing their best,, He came to the conclusion that many students are strongly influenced by their college environment in matters of college honesry. Within the context of the environment and pur— pose Of a university, let's examine the two systems aimed at eliminating the incidence Of cheating on ex— aminations-—the proctor system and the honor system. TO seek, guard, and speak for the truth are considered the duties of any institution of higher learning. We would hope college seeks, through the exchange between students and faculty, not merely to convey knowledge and gain skills, but also to develOp understandings, insights, and character. How do the proctor and honor systems fit in with this philosophy of a university? 200 The theory behind the proctor system for exam- inations is that, with the instructor watching you, the threat Of his detecting cheating on your part will dis— courage you from cheating. How does the proctor system work in practice? Let's look at the examination sit— uation in our attempt to answer this question. Don't h! we often think of an examination as a matching Of wits between student and instructor? In effect, through the material on the test, the instructor is Offering a challenge to the student and the student attempts to 1! ‘1' meet this challenge. This is a healthy situation. How- ever, now add the proctor to the examination. Hasn't the area of challenge broadened? Since the instructor is present to detect cheating, isn't his physical pres— ence an invitation to students to attempt to "beat the system?" Such a system does not help to maintain an environment where a free exchange between student and faculty can truly take place. The theory of the honor system is based on the premise that to attack the problem Of cheating one must foster an atmosphere of learning in which cheating is impractical. In other words, the university should develop a setting in which it is expected that cheating will not take place. The honor system simply means no instructor is obliged to be present during examinations. jMy good friend, Dr. Canning, once spoke to me of the dey 201 he had a visitor to class who wished to see our library. Professor Canning was conducting the final exam Of the class but said he would go with him. He said the visitor showed amazement that he could leave class during an examination. It can be seen that with the faculty- proctor absent, one of the motives for cheating is removed. There is no longer the same challenge to try to outsmart the instructor. Some of you may say that some students will still attempt to cheat. However, a key factor is working under the honor system which is not present under the proctor system. Students taking examinations with a proctor depend upon him to deal with cheating. Students may assume the attitude of dissociating themselves from the problem. Let the instruc- tor worry about it! It is his problem! The honor sys— tem places cheating in a different light. Suddenly, the students as a group and as individuals have the responsibility placed on their shoulders. The student who may have cheated under the proctor system now has second thoughts. He is more reluctant, whether for personal reasons or a desire not to go against his peers. Having known familiarly a campus with a working honor system, Professor Canning, one evening while visiting with my family, pointed out to me that he had been in classrooms where the students merely insisted that there be fairness in the examination room whether the_professor was present or not. 202 The question inevitably arises: Does the honor system actually curb cheating? It is fine in theory but do students take it seriously. A former,professor Of mine, Dean Carl H. Fine, who is now at the Univer- sity Of Michigan, told me that the faculty, in the be- ginning, didn't think the students would do much better than faculty supervision, but, after four years, the [L faculty holds the Opinion that, though matters are not perfect, they have improved greatly. There is consider— able evidence from students and faculty alike that the honor system is effective. We can see results at a specific university. A five—year study was conducted prior, during, and after the adOption Of the honor sys- tem at Brigham Young University. It was discovered that rates of cheating in the classes were reduced by nearly two—thirds. While these findings do not neces— sarily prove a direct causal relationship between the honor system and a reduction of cheating, the liklihood is certainly present. Some Of you may question the typ- icalness of such a specific instance. Yet, the study of 99 colleges by Columbia University cited earlier, noted that cheating is most evident at schools where the faculty alone tries to control the problem. It occurs far less Often at colleges with an honor system. Thus, it would seem reasonable to conclude that the honor system is effective in achieving one Of its pri- mary purposes--COping with the problem of cheating. 203 In addition to being a workable solution, the honor system has other distinct advantages. Its great- est benefit is that it is educational as well as regulatory. Shouldn't a university help to create an environment which will be most favorable to develOping responsible individuals? What about the proctor system for examinations? In the most favorable setting, it is a kind of supervision. Most of the time it is un- fortunately an inspection system and, at its worst, it is a policing system. Is this the best our colleges and universities can dofbr us? The development of responsibility and character should not stop at an inspection or police stage. A fraternity brother Of mine, feelipg confident after taking his Rhodes Scholar examinations, told me: "The honor system gave me an Opportunity to do my best. We all do our best when we really believe somebody has faith in us. It's the indifference under the proctor system that makes ration- alization for cheating so much easier." It should be possible for you, as college students, to assume the responsibility of maintaining fairness in examinations. Now it is important to realize that the honor system is not without criticism. The Air Force Academy incident of 1965 is pointed to as an example of the failure of the honor system to curb cheating. But, do you realize that only four percent of all of the cadets ‘IFK- . - 204 were implicated in the cheating incident. Compare that figure with the nearly fifty percent rate cited at most colleges-—particularly those with no honor system. The honor system is not perfect but, according to this comparison alone, it is certainly more effective than the proctor system. Another criticism is that, while the honor system may work in certain small private colleges, it is not practical when applied to larger state universities. Yet, do you consider Dartmouth College, the University of North Carolina, the University Of Texas, or the Un- iversity of California small colleges? Of course not, and these universities have honor systems! Finally, the honor system is accused Of being fine in theory but unworkable in practice. However, I hope the first-hand experiences with the,proctor sys- tem and honor system related during this speech by in— dividuals I have personally known, have shown that the honor system not only can but does work. The honor system is no overnight paneca to elim— inating cheating but one basic fact must be emphasized. The honor system is effective in curbing cheating be- cause it establishes an atmosphere Of trust and con— fidence. The individual responsibility and integrity of the student are given a chance. Recently, I saw a 205 close college friend of mine who is a Merit Scholar at the University Of Wisconsin. He said the University is considering the adoption of the honor system. Although somewhat skeptical of it, he strongly believes it should be given a fair trial. Now, I say to you. Give it a fair trial. AdOpt the honor system! l, APPENDIX III Table 23.-—Data from three dimensions Of perceived source credibility and attitude change for subjects who heard the No—Reference Message 1 Attitude- Competence Trustworthiness Dynamism Difference Score Score Score Score 10 11 8 10 13 12 9 — 3 ll 9 10 -23 10 13 10 -19 13 13 8 -23 6 10 9 -27 13 17 18 — 5 13 17 5 0 10 9 12 - l 16 10 9 — 1 15 10 11 0 11 12 13 - 9 9 11 12 3 8 7 9 - 6 10 9 6 - l 6 ll 11 - l 11 10 7 -12 9 14 ll — 9 15 20 16 - 8 ll 11 14 -13 9 8 9 —10 10 13 11 — 8 8 8 10 — 1 12 11 7 10 13 15 11 6 Total272 291 256 -151 Mean 10.88 11.64 10.24 - 6.04 Vari— 6.78 9.66 8.69 94.12 ance 1 This treatment served as an index of nO-reference in the series Of messages employing an increasing fre- quency Of self-reference. 2This score was the difference between the pre- test and post-test scores for each subject. A minus sign indicated attitude change in the direction advocated in the message. The highest score prefaced with a minus sign indicated the greatest degree Of attitude change. 207 208 Table 24.-—Data from three dimensions of perceived source credibility and attitude change for subjects who heard the Low Self-Reference Message Attitude— Competence Trustworthiness Dynamism Difference Score Score Score Score 10 8 10 ~17 11 11 8 -26 7 l2 9 2 9 13 7 -15 7 5 7 -20 10 11 8 - 8 11 8 8 1 11 ll 10 4 ll 9 14 0 12 ll 8 6 5 9 ll -13 9 8 ll — 2 l2 12 19 - 5 8 12 10 - 4 11 7 10 -24 11 19 10 1 11 9 7 1 10 10 13 13 11 15 9 4 15 10 4 -34 9 9 ll 0 9 11 10 4 9 17 ll 3 5 8 9 0 10 9 12 ~15 Total244 264 246 -l44 Mean 9.76 10.56 9.84 - 5.76 Vari- 4.86 9.51 8.14 136.02 ance 209 Table 25.--Data from three dimensions Of perceived source credibility and attitude change for subjects who heard the Moderate Self—Reference Message Attitude— Competence Trustworthiness Dynamism Difference Score Score Score Score 7 7 8 -29 6 6 9 ~13 _ 6 4 7 -12 Ex 6 7 12 -14 - 9 l2 9 — 6 7 8 6 - 4 5 4 8 —l4 7 11 9 —13 9 9 8 -15 9 9 7 5 . 4 7 14 -14 1 8 9 8 -32 8 7 10 -16 5 8 12 -23 6 7 8 - 3 8 11 10 - 6 10 14 18 -26 7 8 7 - 4 7 11 10 -14 7 7 8 — 7 6 8 l7 4 l4 8 ll — 6 5 6 5 - 8 6 7 7 —15 8 8 11 -18 Totall80 203 239 -303 Mean 7.20 8.12 9.56 — 12.12 Vari- 4.17 5.36 9.92 83.19 ance 210 Table 26.——Data from three dimensions Of perceived source credibility and attitude change for subjects who heard the High Self—Reference Message Attitude- Competence Trustworthiness Dynamism Difference Score Score Score Score 9 -17 10 2 11 — 8 9 — 6 17 —18 7 —l7 6 O 6 - 8 9 —12 9 - 6 6 9 10 10 13 9 11 8 f—J — 8 - 9 —16 -12 - 3 -13 -16 3 - 4 —24 — 4 —21 —16 -22 -16 H HF4 00m\Ieoomxombqr—Jomowmxnobmcoxoxoexooo H H (DOWCDrDWCDmmOWOKDOW\IGDQHmmKOfiCDrmeD KOUOOWmflU'IKD Totall83 176 226 —271 Mean 7.32 7.04 9.04 - 10.84 Vari- 5.23 3.96 7.04 56.06 ance 211 Table 27.—-Attitude change data for subjects who heard no message Attitude- Difference Score I I ;_n Ni—‘QHUTOHQUWUWCDQWOHWHWQUWUWOWCDH I HF4 1 —13 Total 9 Mean .36 Variance 47.82 1This treatment served as the control group for four self—reference experimental groups. “a... A. ' ‘ 212 Table 28.-~Data from three dimensions Of perceived source credibility and attitude change for subjects who heard the NO-Reference Message 1 Attitude- Competence Trustworthiness Dynamism Difference Score Score Score Score l7 14 19 6 12 14 8 - 7 7 12 10 -10 15 7 9 4 13 14 7 —10 13 13 10 - 8 6 5 6 - 8 10 ll 11 —10 15 8 8 5 ll 13 10 -11 12 14 10 -23 8 19 9 10 8 10 12 - 5 6 10 6 0 15 16 16 - 2 8 17 17 -12 15 13 16 17 6 6 7 -21 17 15 17 - 5 8 8 8 - 1 5 4 7 -l8 5 7 8 -10 9 l7 4 —15 7 9 5 - 7 6 5 7 - 6 Total 254 281 247 -147 Mean 10.16 11.24 9.88 - 5.88 Vari- 15.56 17.77 16.78 87.61 ance 1 This treatment served as an index of nO-refer- ence in the series of messages employing an increasing frequency Of prestige—reference. 213 Table 29.~~Data from three dimensions Of perceived source credibility and attitude change for subjects who heard the Low Prestige—Reference Message Attitude- Competence Trustworthiness Dynamism Difference Score Score Score Score 9 7 9 - 5 6 6 8 ~17 10 7 12 10 12 8 8 ~14 7 6 13 ~ 6 8 16 17 5 7 10 11 ~17 7 12 14 ~17 9 8 15 ~ 3 13 14 ll - 1 6 9 14 ~10 6 5 15 ~11 7 8 9 ~13 6 9 12 4 13 12 16 ~ 2 14 10 10 ~ 6 10 7 12 2 11 17 9 -.7 10 13 8 2 8 12 8 — 9 9 13 17 ~10 12 15 12 ~ 5 9 7 12 ~ 9 l3 9 15 ~11 9 6 9 - 2 Total 231 248 296 ~152 Mean 9.24 9.92 11.84 ~ 6.08 Vari— 6.27 11.83 8.64 52.08 ance "I" ‘. ‘1 214 Table 30.~~Data from three dimensions Of perceived source credibility and attitude change for subjects who heard the Moderate Prestige-Reference Message Attitude- v~ Competence Trustworthiness Dynamism Difference Score Score Score Score 12 12 11 ~ 3 8 9 l4 — 6 9 15 10 ~ 1 11 ll 12 ~10 11 8 8 ~10 11 11 13 ~23 9 7 8 ~15 9 6 16 ~ 1 12 ll 13 ~ 4 6 7 8 ~ 8 10 5 7 6 8 11 8 _7 12 10 9 ~14 8 12 16 ~19 8 6 10 - 5 12 12 19 ~ 1 8 11 9 ~12 4 7 7 ~25 12 13 12 2 9 16 13 ~ 5 8 8 12 ~10 10 12 9 2 10 9 10 14 9 6 9 - l 6 6 8 ~11 Total 232 241 271 ~153 Mean 9.28 9.64 10.84 ~ 6.12 Vari— 4.46 8.91 9.72 84.69 ance 215 Table 31.-~Data from three dimensions of perceived source credibility and attitude change for subjects who heard the High Prestige-Reference Message Attitude- Competence Trustworthiness Dynamism Difference Score Score Score Score 10 7 9 ~12 10 6 7 ~ 6 9 6 10 - 8 6 7 17 --4 10 5 12 ~11 12 6 12 ~10 10 5 6 ~18 ll 5 10 ~ 4 4 10 6 2 6 9 10 ~13 9 6 9 ~ 3 l3 8 8 ~ 3 10 5 8 ~10 8 6 18 ~17 6 7 12 ~14 4 5 8 ~20 5 6 6 ~ 6 6 8 7 ~25 9 9 10 ~20 11 12 10 — 2 12 13 11 ~ 6 ll 9 10 ~10 11 10 13 ~ 8 8 8 21 - 7 15 16 6 ~ 2 Total 227 193 256 ~237 Mean 9.08 7.72 10.24 - 9.48 Vari- 8.66 7.13 14.61 44.51 ance 216 Table 32.-~Attitude change data for subjects who heard no message1 Attitude— Difference Score \IWNNl-‘F-‘OUTU'I I l—‘l—J OO ‘23:.- l—‘ waqt—mehwwwboo Total Mean 1.20 Variance 26.92 (.0 O 1This treatment served as the control group for the four prestige-reference experimental groups. 217 Table 33.-~Resu1ts Of factor analysis for each of the original forty-four attitude statements Standard Deviation Factor Loading Mean Attitude Statement 10. 11. 12. The honor system teaches the lowest form Of competition ~ every man for himself. The honor system represents the best long—term solution to cheating in college. .38 .67 The honor system is a failure..28 The faculty-proctor system works satisfactorily in pre— venting cheating. The honor system gives stu~ dents practice in self- imposed discipline. The honor system should be rejected. The honor system is an out— ward expression Of faculty trust in students. The honor system fails to get at the cause Of the problem Of cheating. Under the honor system, trust by faculty encourages trust- worthiness in students. The honor system helps the dishonest student and pen.— alizes the honest student. The honor system should be adopted. The honor system creates a better public image of the university. .61 .76 .36 .57 .64 .64 .67 .58 -014 2.06 2.55 218 TABLE 33~Continued Attitude Statement Factor Loading Mean Standard Deviation 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. There are better solu~ tions to the problem of cheating than the honor system. .01 The honor system helps to establish an atmosphere of learning in which cheating is impractical. .55 Despite some misgivings about it, I feel the honor system should be given a fair trial. .72 Cheating is encouraged un- der the honor system. .48 Elimination of the faculty— proctor system would intro- duce more serious problems. .01 The honor system encourages greater confidence and co— operation between students and faculty. .46 The general lack Of morality among college students pre- vents the use Of the honor system. .30 The desire of students to get through college by any means makes the honor sys— tem unworkable. .63 The principle of the honor system for curbing cheat- ing is essentially sound. .74 The honor system prepares the student to be person- ally accountable for his own actions. ~.03 219 TABLE 33~~Continued Factor Attitude Statement Loading Mean Standard Deviation 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. The proctor system, in ef— fect, dares the student to cheat. .48 NO honor system can raise the level of moral conduct of students. .58 With the faculty—proctor constantly present, stu- dents feel they are being expected to cheat. .15 The honor system, as a sol- ution to cheating, is sup— erior to all others. .78 The honor system is more an idealistic dream than a workable solution. .53 The honor system makes a significant contribution to the college education Of a student. .68 It is extremely unlikely that the honor system can solve the problem Of cheating. .61 The honor system is essent~ ial for a strong university educational program. .75 The honor system eliminates the battle of wits between the faculty and the student. .62 The honor system does reduce the amount Of cheating. .52 The honor system is another unwarranted attempt to teach morality in the university. .65 3.42 7» Iv.- . "=' J' 220 TABLE 33~~Continued Attitude Statement FaCtSr Mean Stagdafd Loading DeViation 34. The honor system is the greatest influence on a college campus. .43 2.76 1.03 35. The honor system encour~ ages a resentment Of super- vision. .76 2.86 1.11 36. The honor system works bet— ter than the faculty—proctor system. .63 3.24 0.97 37. The honor system encourages in students a sense of in- dividual responsibility. .58 2.55 0.88 38. The honor system intensifies the problem Of cheating. .32 2.29 0.74 39. The proctor system fails to deal with the factors which encourage cheating. .66 2.74 0.90 40. The honor system is inef- fective at large state un- iversities. .56 3.01 0.81 41. Competition for grades in college requires the use of the proctor system. .71 3.00 1.02 42. The honor system is based on unsound educational prin- ciples. .64 2.27 0.75 43. The honor system is fine in theory but unworkable in practice. .78 2.74 1.02 44. The honor system does more harm than good to the univer- sity community. .71 2.34 0.93 BIBL IOGRAP HY 1M)" - * BIBLIOGRAPHY A. BOOKS Abernathy, Elton. The Advocate ~ A Manual Of Persuasion. New York: David McKay Company, Inc., 1964. Aristotle. The Rhetoric of Aristotle. Translated by Lane Cooper. New York: Appleton—Century—Crofts, Inc., 1932. Brembeck, Winston L. and Howell, William S. Persuasion ~ A Means of Social Control. New York: Prentice- Hall, Inc., 1952. Cicero, Marcus Tullius. De Oratore. Translated by E. W. Sutton and Horace Rackham. Cambridge: Harvard University Press, 1948, II. Goode, William J. and Hatt, Paul K. Methods in Social Research. New York: McGraw-Hill Book Company, Inc., 1952. Heider, Fritz. The Psychology of Interpersonal Relations. New York: John Wiley & Sons, Inc., 1958. Hovland, Carl I., Janis, Irving L., and Kelley, Harold H. Communication and Persuasion. New Haven: Yale University Press, 1953. Hovland, Carl I., Lumsdaine, Arthur A., and Sheffield, Fred D. Experiments on Mass Communication. Vol. III Of Studies in Social Psychology in World War II. Princeton: Princeton University Press, 1949. Lindquist, E. F. Design and Analysis Of Erperiments in Psychology and Education. Boston: Houghton Mifflin Company, 1953. Oliver, Robert T. The Psychologyoffi Persuasive Speech. 2nd edition revised. New York: David McKay Company, Inc., 1957. Osgood, Charles E., Suci, George J., and Tannenbaum, Percy H. The Measurement of Meaning. Urbana: University of Illinois Press, 1957. 222 .‘u . r 223 Quintilian, Marcus Fabius. The Institutio Oratoria Of Quintilian. Translated by H. E. Butler. Cam— bridge: Harvard University Press,reprinted 1963, I, Book I. Selltiz, Claire, et a1. Research Methods in Social Relations. 2nd edition reVISed. New York: Henry Holt and Company, Inc., 1960. State Universityrf New York at Buffalo Bulletin - Univer- sity_College, 1965/1966. Thonssen, Lester and Baird, A. Craig. Speech Criticism. New York: The Ronald Press Company, 1948. B. ARTICLES AND PERIODICALS Andersen, Kenneth and Clevenger, Jr., Theodore. "A Summary of Experimental Research in Ethos," Speech Monographs, XXX (June, 1963), 59—78. Arnett, Claude E., Davidson, Helen H., and Lewis, Hallet N. "Prestige as a Factor in Attitude Changes," Sociology and Social Research, XVI (September-October, 1931), 49-55. Bernberg, Raymond E. "Prestige Suggestion in Art as Communication," Journal Of Social Psychology, XXXVIII (August, 1953), 23-30. Berlo, David and Gulley, Halbert. "Some Determinants Of the Effect of Oral Communication in Producing Attitude Change and Learning," Speech Monographs, XXIV (March, 1957), 10~20. Bettinghaus, Jr., Erwin. "The Operation Of Congruity in the Oral Communication Situation," Speech Monographs, XXVIII (August, 1961), 131—42. Brehm, Jack W. and Lipsher, David, "Communicator~ Communicatee Discrepancy and Perceived Communica- tor Trustworthiness," Journal of Personality, XXVII (September, 1959), 352-61. Burdick, Harry and Burnes, Alan. "A Test Of 'Strain Toward Symmetry' Theories," Journal of Abnormal and Social Psychology, LVII (November, 1958), 367-70. 224 Cole, David. "'Rational Argument' and 'Prestige-Sug- gestion' as Factors Influencing Judgment," Socio— metry, XVII (November, 1954), 350-54. Ewing, Thomas N. ”A Study Of Certain Factors Involved in Changes Of Opinion," Journal of Social Psychology, XVI (August, 1942), 63—88. Gilkinson, Howard, Paulson, Stanley, and Sikkink, Donald. "Effects of Order and Authority in an Argumentative Speech," Quarterly Journal of Speech, XL (April, IR, 1954), 183—92. Goyer, Robert S. "Cognitive Dissonance and Communication Theory," Central States Speech Journal, XV (May, 1964), 90-95. Heider, Fritz. "Attitudes and Cognitive Organization," 1 Journal of Peychology, XXI (January, 1946), 107—12. ‘ Hewgill, Murray A. and Miller, Gerald R. "Source Cred- ibility and Response to Fear—Arousing Communications," Speech Monographs, XXXII (June, 1965),95—101. Hovland, Carl I. and Mandell, Walter. "An Experimental Comparison Of Conclusion Drawing by the Commun- icator and the Audience," Journal of Abnormal and Social Psychology, XLVII (July, 19527, 581—88. Hovland, Carl I. and Weiss, Walter. "The Influence Of Source Credibility on Communication Effectiveness," Public Opinion Quarterly, XV (Winter, 1951), 635—50. Kelman, Herbert C. and Hovland, Carl I. "'Reinstate— ment' Of the Communicator in Delayed Measurement of Opinion Change," Journal of Abnormal and Social Psychology, XLVIII (July, 1953), 327-35. Kerrick, Jean. "The Effect of Relevant and Non—Relevant Sources on Attitude Change," Journal Of Social Psychology, XLVII (February, 1958), 15-20. Kulp II, Daniel H. "Prestige; as Measured by Single— Experience Changes and Their Permanency," Journal of EduCational Research, XXVII (May, 1934), 663—72. Lewis, Helen B. "Studies in the Principles of Judg— ments and Attitudes, IV: The Operatian Of"Prestige Suggestion'," Journal of Social Psychology, XIV (August, 1941), 229-56. 225 Lorge, Irving with Curtiss, Carl C. "Prestige, Sug- gestion and Attitudes," Journal of Social Psychol- ogy, VII (November, 1936), 386-402. Moos, Malcom and Koslin, Bertram. "Prestige Suggestion and Political Leadership," Public Opinion Quarterly, XVI (Spring, 1952), 77—93. Osgood, Charles E. and Tannenbaum, Percy H. "The Prin- ciple Of Congruity in the Prediction of Attitude Change," Psychological Review, LXII (January, 1955), 42"550 Powell, Frederic A. "Open and Closed Mindedness and the Ability to Differentiate Source and Message," Journal of Abnormal and Social Psychology, LXV (July, 1962), 61-64. Saadi, Mitchell and Farnsworth, Paul. "The Degrees Of Acceptance of Dogmatic Statements and Preferences for Their Supposed Makers," Journal Of Abnormal and Social Psychology, XXIX (July-September, 1934), 143-150. Sherif, Muzafer. "An Experimental Study Of Stereotypes," Journal Of Abnormal and Social Psychology, XXIX (January-March, 1935), 371-75. Sikkink, Donald. "An Experimental Study of the Effects on the Listener of Anti-Climax Order and Authority in an Argumentative Speech," Southern Speech Journal, XXII (Winter, 1956), 73-78. Thompson, Wayne. "A Study Of the Attitude Of College Students Toward Thomas E. Dewey Before and After Hearing Him Speak," Speech Monogrephs,XVI (August, 1949), 125-34. Wang, Charles K. A. "Suggested Criteria for Writing Attitude Statements," Journal of Social Psychology, III (February, 1932), 367-73. C. UNPUBLISHED MATERIAL Andersen, Kenneth. "An Experimental Study Of the Inter- action Of Artistic and Non-Artistic Ethos in Persuasion." Unpublished Ph.D. dissertation, University of Wisconsin, 1961. 226 Berlo, David K. and Lemert, James B. "An Empirical Test Of a General Construct of Credibility." Unpublished paper presented at the Speech As- sociation of America Convention, New York, December 29, 1961. Bryson, Kenneth D. "An Experimental Study Of the Ef- fectiveness of the 'Denotative' Speech in Per~ suasion." Unpublished Ph.D. dissertation, North- western University, 1952. Haiman, Franklyn. "An Experimental Study of the Ef— , fects of Ethos in Public Speaking." Unpublished _ Ph.D. dissertation, Northwestern University, 1948. Kersten, Barbara. "An Experimental Study to Determine the Effect of a Speech Of Introduction Upon the Persuasive Speech That Followed." Unpublished 1 thesis, South Dakota State University, 1958. F Ludlum, Thomas. "A Study Of Techniques for Influenc— ing the Credibility Of a Communicator." Unpub~ lished Ph.D. dissertation, Ohio State University, 1956. Paulson, Stanley. "Experimental Study Of Spoken Com- munication: The Effects of Prestige Of the Speaker and Acknowledgement of Opposing Arguments on Audience Retention and Shift Of Opinion." Unpub— lished Ph.D. dissertation, University of MinnesOta, 1952. Pross, Edward. "A Critical Analysis Of Certain Aspects Of Ethical Proof." Unpublished Ph.D. disserta- tion, State University of Iowa, 1942. Sattler, William. "Conceptions Of Ethos in Rhetoric." Unpublished Ph.D. dissertation, Northwestern University, 1941. Strother, Edward. "An Experimental Study Of Ethos as Related to the Introduction in the Persuasive Speaking Situation." Unpublished Ph.D. disserta— tion, Northwestern University, 1951. Tannenbaum, Percy H. "Attitude Toward Source and Con— cept as Factors in Attitude Change Through Com~ munication." Unpublished Ph.D. dissertation, Un~ iversity of Illinois, 1953. 227 D. REFERENCES PERTAINING TO HONOR SYSTEM Baldwin, Bird T., et al. The Present Status of the Honor System in Colleges and Universities. Washington: U.S. Government Printing Office, 1915, Bulletin NO. 8. ' Bride, Jr., Joseph. "A Smudge on the Cadet Code," Life, LVIII, No. 5 (February 5, 1965), 82-83. Canning, Ray. "Does an Honor System Reduce Classroom Cheating? An Experimental Answer," Journal of Ex— perimental Education, XXIV, NO. 4 (June, 1956), 291—96. Cole, William Graham. "Cheating Your Way Through Col- lege," Nation, 190, No. 20 (May 14, 1960), 416-18. Colwell, W. A. "The Honor System," Nation, XCVIII, No. 2553 (June 4, 1914), 663-64. Downing, Francis. "The Tragedy at West Point," Com— monweal, LIV, NO. 20 (August 24, 1951), 476-77. Eddy, Jr., Edward D. "The College Influence on Student Character. Washington: American Council on Educa- tion, 1959. Edwards, R. H., Artman, J. M., and Fisher, Galen M. Undergraduates. Garden City, New York: Doubleday, Doran and Company, Inc., 1928. "For the Exam Hurdle: A Cribbing Shirt From Ancient China," Scholastic, LXXVI (February 3, 1960), 17. Gavit, John P. College. New York: Harcourt, Brace and Company, 1925. Geiger, Joseph Roy. "The Educational Value of the Honor System," School and Society, XXI, NO. 540 (May 2, 1925), 516-22. A ‘ . "The Honor System in Colleges," International Journal Of Ethics, XXXII, NO. 4 (July, 1922), 398—409. Glicksberg, Charles I. "Student Ethics and the Honor System," School and Society, LXXXV, No. 212 (May 25, 1957), 181—83. 228 Harrington, E. R. "Observations on Cheating," The Clear- ing House, XXXV, NO. 6 (February, 1961), 354-56. Hawes, James A. Twenty Years Among the Twentinear-Olds. New York: E. P. Dutton and Company, Inc., 1929. "How Would You Stop Cheating on Exams? Forum Discussion," Scholastic, LXIII (January 13, 1954), 7-9. Jacob, Philip E. Changing Values in College. New York: Harper and Brothers Publishers, 1957. Joynes, E. S. "The 'Honor System' in College," Nation, LXXXIII, NO. 2152 (September 27, 1906), 259. Kelly, F. J- The American Art College. New York: The MacMillan Company, 1925. Lyman, R. L. "The Problem Of Student Honor in Colleges and Universities," School Review, XXXV, NO. 4 (April, 1927), 253-71. Mathews, C. 0. Bibliography on the Honor System and Academic Honesty in American Schools and Colleges. Washington: U.S. Government Printing Office, 1931, Pamphlet No. 16. Meinecke, Charlotte Drummond. "Student Government and the Honor Code," Junior College Journal, XXIII, NO. 8 (April, 1953), 426-34. Miner, J. B. "The Control of College Cheating," School and Society, XXXII, No. 815 (August 9, 1930), 199—204. New York Times, February 4, 1965, p. 19. New York Times, February 6, 1965, p. 24. New York Times, March 21, 1965, p. 8E. "On Your Honor? Pro and Con Discussion: Should All High Schools AdOpt the Honor System?," Scholastic, LXXI (January 17, 1958), 6-7. Ordeman, D. T. "Unfinished Editorial," College and University, XXVII, No. 1 (October, 1951), 111—13. 229 Russell, May. "Students Achieve High Moral Values Through College Honor Systems," Junior College Journal, XXXII, NO. 1 (September, 196177 37-39. Schmidt, George P. The Liberal Arts College. New Brunswick, New Jersey: Rutgers University Press, 1957. Smith, C. A. "Honor in Student Life in Colleges and Universities," Educational Review, XXX (November, 1905), 384-95. "A Startling Survey on College Cribbing," Life, LVIII, NO. 5 (February 5, 1965), 84. Stern, Bernard H. "What Should be Done About Cheating in College?," Educational Forum, XXVII (November, 1962), 79-83. Stevens, David H. "Honor Without System," School and Society, XIV, NO. 365 (December 24, 1921), 613-16. Stevens, W. H. "The Honor System," Nation, XCVIII, NO. 2556 (June 25, 1914), 754. Trabue, Ann. "Classroom Cheating ~ An Isolated Phenom~ enon?," Educational Record, XLIII, No. 4 (October, 1962), 309—16. 3.3 .F _ .