THE EFFECT ON ATTH'UDES AND RETENTION OF MESSAGE CQRDER IN CGNTRQVERSIAL MATERIAL Thais for flu Ilia-gt“ of Ph. D. MiCHiGAN STATE UNWERSZTY Terry A. Weiden 196E "-HJ‘Q 0"!" ' ' ( ___._—————--'...~—-—— - ’ W Wm WM \ \ _— IWT i \ W N m 3 1293 10716 4588 This is to certify that the thesis entitled THE EFFECT ON ATTITUDES AND RETENTION OF MESSAGE ORDER IN CONTROVERSIAL MATERIAL presented by TERRY A. WELDEN has been accepted towards fulfillment of the requirements for Speech jig; degree in \ Major prvofessor Date November 10, 1961 0-169 LIBRARY Michigan State University MSU RETURNING MATERIALS: Place in book drop to remove this checkout from LIBRARIE 4;!5.c...:L_ your record. FINES wil] be charged if book is returned after the date stamped below. ABSTRACT THE EFFECT ON ATTITUDES AND RETENTION OF MESSAGE ORDER IN CONTROVERSIAL MATERIAL by Terry A. Walden This study developed from consistent research findings on the relationship of attitudes and retention in two-sided presentations of controversial material. “The problem under in- . vestigation generated three hypotheses: 1) When an audience is presented with a message containing equal discussion of both sides of an issue, initial exposure to the side compatible with the audience's existing attitude results in a more extreme atti- tude toward the tOpic of the speech than results if there is initial exposure to the incompatible material. Corollary A. Given that the audience has a favorable initiw. attitude toward the message tOpic, initial exposure to the compatible material results in a more favorable attitude toward the tOpic of the speech than results if there is initial exposure to the incom- patible material. Corollary B. Given that the audience has an unfavorable initial attitude toward the message tapic, initial exposure to the compatible material results in a more unfavor- able attitude toward the t0pic of the speech than results if there is initial exposure to the incompatible material; 2) When an audience is presented with a message containing equal Terry A. Walden discussion of both sides of an issue, the material compatible with the audience's existing attitude will be better retained than will the material which is incompatible; 3) When an audience is pre- sented with a message containing equal discussion of both sides of an issue, the difference between the retention of compatible and incompatible material will be greater when the compatible material is presented first than when the incompatible material is presented first. While specific hypotheses were not advanced, questions of speaker credibility and message saliency were considered in the investigation. A pre-test-post-test design utilized an experimental mes- sage to present both pro and con material on the tapic of college fraternities and sororities to an eXperimental population of 225 college students. A corresponding control group consisted of 112 students. The message was presented to one segment of the pepu- lation in a pro-con order and to another segment in a con-pro order. The measure of initial attitude determined three groups of subjects: favorable, neutral, and unfavorable. Thus, it was possible to characterize message order as compatible-incompatible or incompatible-compatible for both the initially favorable and unfavorable groups of subjects. With the measure of post-attitu de, it was possible to determine and compare attitude change scores for subjects receiving different message orders. An index of re— tention within the post-test was used to measure differences in retention between compatible and incompatible material and Terry A. Welden differences in retention attributable to message order. With respect to Hypothesis I, significant differences were obtained fOr the initially favorable group of subjects. Order appeared to make no difference for the initially unfavorable group of subjects. However, this group did not remain unfavorable toward the experimental tapic, and thus did not meet the requirements of the rationale for the predictions made by any of the three hypo- theses. Again, for the initially favorable subjects, significant differences were obtained in the test of Hypothesis II. Signifi- cantly more pro material than con material was retained. This relationship held regardless of the order of presentation. No support was obtained for Hypothesis III. Also, the questions of speaker credibility and message saliency as phrased . in the present study received no clarification. THE EFFECT ON ATTITUDES AND RETENTION OF MESSAGE ORDER IN CONTROVERSIAL MATERIAL By 6:.“ Terry A\.‘\ Walden A THESIS Submitted to Michigan State University in.partial fulfillment of the requirements for the degree of DOCTOR OF PETIDSOPHY Department of Speech 1961 ACKNIMIEII‘IMENTS The author is grateful for this Opportunity to express his appreciation to the members of his guidance committee: Dr. Kenneth G. Hence, Chairman; Dr. David K. Berlo; Dr. Archie O. Heller; Dr. Leo A. Martin; Dr. David C. Ralph; and Dr. Gordon 1.. Thomas. Their help and attention throughout his whole pro- gram of study was basic to the develOpment of his enthusiasm for the challenge they outlined. Both the chairman of the author's committee, Dr. Kenneth G. Hence, and the thesis advisor, Dr. David K. Berlo, deserve and are offered a special thanks. They provided a much needed double dose of patience and confidence. It is the author's hepe that they will continue to so generously fortify him. 7 TABLE OF CONTENTS ACWNIENTS O O O 0 O O O O O O O O 0 IIISTOFTABIESOOOOOOOOOOOOO LIST OF APPENDICES. O O O O O O 0 O O O 0 Chapter I. mowCTION O O O O O O O O O 0 Statement of the Problem Setting in the Literature Justification View of the Dissertation II. HIPOTHESES AND RATIONAIE . . . . III. HPERD’IENTALPROCEDURES . . . . . Subjects Experimental Design Materials . Data Collection IV. RESULTSANDANAIJSIS . . . . . . V. DISCUSSION ........... Initially Unfavorable Subjects Initially Neutral Subjects Initially Favorable Subjects Related Questions Research Implications BIBLIOGRAPHY O O O O O O O 0 O O O O 0 O WICESOOOOOOOOOOOOOO. iii Page iv 21: 30 39 6O 62 Table l. 2. 3. h. 5. 6. 7. 8. 9. 10. 11. LIST OF TABLES A Sumary Count of Subjects According to the Factors of Order, Saliency, Attitude, and MessagePresentation..... oeooeoeeoo Chi-square Analysis of Relationships of Close Ability and Pre-attitude and the Variables of Sex and Fraternity Membership . . . . . . . . . Distribution of Four Variables Across Four Experimental and Four Control Groups . . . . . . . Percein Relevance Differences Among Different Message Orders and Saliency Differing Messages . . Attitude Change Differences between Non-Salient andSalientGroups................ Retention Differences between Non-Salient and SalientGroups.................. Pre to Post-attitude Change Differences between Experimental and Control GNUPS o o e o e e e o o Retention Differences for Both Pro and Con Material between Experimental and Control Groups . Pro to Post-attitude Change Differences between Pro-Con and Con-Pro Ordered Treatment Groups Arranged According to Initial Attitude o o o o o Retention Differences between Pro and Con MessageSegments.......o......... Retention Differences between Pro-Con Ordered Treatments and Con-Pro Ordered Treatments . . . . iv 50 LIST OF APPENDICES Appendix Page Acooooeoooceoocccoooooococo. 62‘ Non-Salient Message: Pro-Con Order Salient Message: Con-Pro Order BO...OOOOOOOOOOOOOOOOOOOOOO. 7h Pre-Test Form: Experimental and Control Groups Post-Test Form: Experimental, Non-Salient, Pro-Con Order Post-Test Form: Control, Salient, Con-Pro Order CHAPTER I DWTRONCTION Aristotle contends in his Rhetoric that a speech consists of stating the case and proving it. while perhaps the |'sorry na- ture of audiences“ will necessitate, he says, the inclusion of an introduction and s conclusion, these are incidental to an under- standing of what constitutes a speech. The speech was for Aristotle only one of the three elments of a speech situation. It contains, in addition, the speaker and his audience. Tradi- tionally, writers on public speaking have worked from these classi- cal divisions. Speakers, though possibly unaware of such formali- sstions, are unlikely to respond in much different terms if (period along these lines. that speakers need is advice on unipulnting the given elements in a speech situation so as to increase the probability that their specific purpose for speaking is achieved. Studies reported from enerience through observation and througl observation with controls attempt to provide needed advice. The variable of 3% within a mango is easily under the control of a speaker before he blivers his mssage ad is easily controlled by the experimenter wanting to test hypotheses related to message effectiveness. Unfortunately, while this fact has spurred Itch quantitative research, potential speakers still lack clear advice as to "effective manipulation." An examination of the implications 1 2 of Aristotle's basic assertions points up the difficulty and lays the foundation for the present investigation. First of all, Aristotle, in $fining a speech in terms of ten distinct elements, has helped the variable of order to be viewed as a message variable. Obviously, any variation of these elements is a vu-iation of the message order. Secondly, he pro- vided labels for other elements in the total speech situation. is they are not inbpenknt elements, it night be useful to inves- tigate a variable like order, not only among coordinate elements, but among subordinate and superordinate elements as well. while a less apparent relationship exists, the little seemingly consis- tent advice available night Justify its examination. The first implication requires considerable explication, however, before the second one my be appropriately detailed. The Vuiable of Order Wrists findthe 'idealtype' to be auseful concept for theory development. The concept is necessary to wcount for the perfectly organised speech. However, the random urangenent of all the elements of a message is quite attainable and would lead to a significant lowering of the commnication of the message. A task of the student of orgmisation is to help determine the effect on ccumnication of various orders of message elments within the range of possible orders. this is no small task. Beginning with the eluents I'sta‘tanent of the case" and "proof,'I it is possible inmo- tively to arrange these elemnts by first stating the selected swim and leading up to the assertion it supports. It is also possible to state the assertion and proceed to support it. It is 3 also quite proper to present evidence leading to an assertion, to state the assertion, and then to continue to support it. To the extent that the evidence selected is believed to be, or demonstrated to be, ranked as to stnngth, it is possible to vary the order of evidence along a strong-weak dimension with all the orders involved according to the number of elements of evidence. Since assertions, in the view of audiences, can serve a dual role and support other assertions or even themselves in some cases, sub- assertions can be ranked as to strength with the same potential of order variation. Even sets of sub-as sertiens with attending evi- bnce enter the realm of quite possible order choices. Speakers are not restricted to the above vm-iations. New messages are constructed to treat of controversial subjects. in issue is recognised, and a side is taken. If, for reasons of objec- tivity, fairness, or strategy, the speaker chooses to include asser- tions and evidence from both sides of the issue, he faces the deci- sion as to where to place the arguments that are incompatible with his views. Should he place tha first and get them out of the m orplncethenlestsoastoarnhismdiencewithhisversionat the outset! Perhaps he should scatter than about so as to weaken their impact? m above orcbr choices do not exhaust the possible order variations, but they point up a cosplexityof choices. Reference to the eluents of evidence as believed or demonstrated to lie along a strong-weak dimension signals the second implication of Aristotle's divisions. Any order change is a simificant order change only to the «tent that it affects a significant difference in h the retention and attitucbs of the mdience. They may retain more or less of the message, and their attitudes toward the speaker or his topic may become more or less favorable. For the present study, st least, such changes would constitute significant changes attribu- table to order differences. If it is thus hypothesized that order be treated as an interaction variable affecting messages and receivers, pm'ticular speech situations can be visualized in which possible sig- nificant differences in retention md attitudes can be inbred accord- ing to the order of audience defined elements. 1 speaker wanting "objectively" to present equal discussion of both sides of an issue has as one available choice, the develop- ment of pro and con segments. Much as in a one-man debate, he nay introdlce his tapic and then present the arguments for and then the arguments against the topic. Unlike traditional debate, he may re- verse the order of presentation. If this should significantly alter either or both retention and existing attitucbs, will he in- sure or dew his desired "objectivity"? It is also possible that a speaker in wishing to "appear“ objective, decides to allot the same amount of time to both sides of a controversial issue on which he has definite feelings. Assuming that he still wishes effectively to favor his I'side," he is faced with the decision as to which order to employ. Even more basic is the cpsstion, will it make any difference? in answer to this qzestion seems to necessitate some reference to I the existing attitudes of the audience. The statement of the probleea for the present study encompasses this question. 5 Statement of the Problem Given a two-sided message on a controversial issue 3 given equal treatment of both sides of the message; given that one sick of the message is compatible with existing audience attitudes and that one side is incompatible with existing attitudes; the specific problem to be investigated is, what effect does the reversing of the order of presentation have upon attitudes and retention? This question generates three specific hypotheses which are enplicated in Chapter two. Related qaestions, capable of in- vestigation, can be asked within the context of the present study. Two such questions, while not treated as hypotheses in the stuck, may be attacked through a modification of the experimental design. Chapter three presents the experimental design adapted for the in- vestigation of the effect of message saliency and speaker credi- bility. Here, in order to define meaningfully and to limit the problem, the basic vn‘iables must be presented. Exi__s‘_ti£g_ attitn des m be differentially altered as a function of ordering compatible and Matias material. Similarly, retention may be altered. Attitudes and retention may be altered conjointly or independently. Hessages, identical except for the degree of salie_n_cz for the audi- ence, may differentially alter both attitudes and retention. The credibility assigned by m mdience to an unknown speaker may vary as a function of ordering compatible and incompatible material. Definition If a stated issue is to be regarded as a "controversial issue,‘I the stntuent of that issue alone will evoke an expression 6 of attitude other than neutral by the majority of an indended audience and the expression will not be a unanimous favoring or dis- favoring. For the present stuw, the topic, "Social Fraternities and Sororities at Kansas State University," fulfills the criteria for a controversial issue. "Equal treatment of material" is bfinsd as presenting both pro and con material in approximately the same mmber of words, with the same mmber of major assdrtiens and with similar kinds of support for the major assertions. "Compatible" and "incompatible" material requires extended clarification. its subject of a speech may be viewed by an indivi- dual in a favorable or unfavorable manner. Of course, he could also be neutral toward the subject. If he is a potential member of anaudiencetobe uposedto the speech, he mybe saidtohave an initially favorable, neutral, or unfavorable attitude toward the speech tapic. the speaker who includes in his message material that is favorable and unfavorable to the topic of the message is said to present both the pro and the can of the matter at hand. Given two individuals, one with an initially favorable attitude toward the topic and the other with an initially unfavorable attitude, the pre- sentation of the queech in a pro-con order is considered ordered in a compatible-incomadble manner for the I'pro'l individual and ordered in an inmmpatible-cmupatible manner for the second or "con" individial. Limitations ' ~ The following conditions are «eased necessary to construct an adequate experimental design. In order to isolate the effects of order, no information is 7 contained in the introduction of the experimental message to suggest that a discussion of both sides of the subject will follow. Initially neutral subjects are eliminated from the major analyses although incidental analysis is made of their data. In order to minimize and control the factor of credibility, a speaker is used who is unknown and unidentified to the experimental audieme. The minimizing seems necessary to prevent a highly credi- ble speaker from retaining "enough" credibility while presenting in- compatible material to cloud existing order differences. V Settig in the Literature There is a general framework within which order differences in controversial material can be viewed. The effect of order differ- ences on both attitudes and retention can be examined in light of the hypothesis that for controversial material, attitudes very direct- 1y with retention. The present stuw developed from the literature in support of this hypothesis. Frame of Reference in early stuw by Watson and Hartmannl tended to confirm the hypothesis. Their interest in the question grew out of the work by Sherifz on the auto-kinetic effect. Sherif noted the influence of Bartlett on his thinking. Bartlett3 reports on his own lengthy 1V. S. Watson and O. Hartmann, "The Rigidity of a Basic Atti- tudinal Frans," Journal of Abnormal and Social ijm (1939) s 311F635 e 2M. Sherif, The Psychology of Social Norms, (New York: Harper. 1936)- 3F. c. Bartlett, Social Elan, (C - 8 experience and observation of peeple ' s ability to recall past events. His early theoretic concern with the roles of attitude md learning is cut-ed up in the statement, . . . when a subject is being asked to remember, very often the first thing that emrges is something of the nature of attitude. The recall is then a construction, lads largely on the basis of this attitude, and its genes-cg effect is that of a justification of the atti- tude. Later thinking on the "frame of reference" that an indivimal em- ploys when viewing his envircnaent grows out of Bartlett's early theorizing. watson and Hartmann assert, "The Gestalt concept of a nontal structure or pattern that acts as a reference field or frame in conditioning local behavior has been placed upon a new superinental foundation since the work of Sherif in his stucb' of the auto-kinetic movement.'5 According to Sherif, frames of re- ference, once established in the individual, " . . . enter as im- portant factors to determine or modify his reactions to situations that he will face 1ater."6 Watson and Hartmann developed an experimental study based on the assumption that, "The chief concrete manifestation of a frmwork is a guiding pattern of belief and action.” They pre- pared suitable controversial written material for both theistic and atheistic s‘lmcbnts. "For the group under investigation . . . 1933., 20?. 5Watson and Hart-sun, M" 315 . ssherif, 22.933.” 1:3. 7Watson and Hartmann, M... 311;. 9 noterial which supported the subject's attitudinal frame was re- tained better than material which Opposed it."8 ldwu'ds9 tested, in a different manner, essentially the same hypothesis as that of Watson and Hartmann. " . . . specifi- cally, the hypothesis is that experiences which are in harmom with an existing frame of reference will tend to be learned and remnbered better than those which are in conflict with it."10 Be prepared a ten minute speech on the "New Deal" containing an equal number of pro and con arguments. His experimental subjects consisted of college students whose initial nttitudes varied from! favorable to unfavorable. It is interesting to note that the speech resulted in little or no change in the subjects' attitude toward the New Deal. The favorable group is still. favorable, the neutral poup is still neutral, andiihe opposed group is still opposed to the New Deal. Edwards' findings were compatible with the earlier studies. "The very same material which is learned by one group is not learned by another, while the material which is not learned by the first group is learned by the second." 1‘2 Levine and nanny” were interested in developing learning 8Ibid., 319. 9L. 1.. Edwards, "Political Frames of Reference as a Factor Influencing Recognition," Journal of Abnormal and Social Psycholog, MI (19111): Bit-5°- 1°_I_'_b_i_._d., 35. 311.1214" 38-39. ”aid” M. 13.1. n. Levine and c. nuphy, "The learning and Forgetting of Controversial Material," Journal of Abnonal and Social Psycholog, mun (191:3). 507-517. 10 (and forgetting) curves for controversial material. Working with individual subjects, they presented a ppo and can set of written messages dealing with the Soviet Union several times at each of four weekly sessions. The subjects consisted of both pro and anti- communist students (determined by a measure of their social phil- osOphy) . The Isvine-lhu-phy theoretic question reads, l'Is material congruent with our social attitudes assimilated in such fashion as to yield a different shape of curve from that which appears when the material conflicts with one's hieszun‘ It was determined that the subjects learned the conflicting material at a significantly lower rate than they did the compatible material. In 1953, Gilkinson, Paulson, and Sikkim:2L5 essentially replicated the study by Edwards” in three separate emeriments. In all three instances they supported his findings. In one of the three instances they attempted to control for the possibility that more comatible material is retained by an individual because he exposes hinself and is exposed more frequently to material which conforms to his bias than to material which conflicts with it. Here, they found no difference between two experimental treatments, but both treatments were consistent with Edwards' findings. There is consistent evidence, then for the statement that for controversial material, attitudes vary directly with retention. Since the present study is concerned with order differences within ”his. 508. 15Howard Gilkinscn, S. F. Falcon, and D. E. Sikkink, "Con- troversial Statements in Connected Discourse, Their Form and Politics ,' §_peech Harlem, 11 (November, 1953), 253-260. 2L6Eaterds, Op. cit., 3h-50. ll. this general frnnework, it is relevant to look at some of the literature dealing with attitudes and/or retention as a function of order. Order Differences Two kinds of order differences mark the litera'mre . Primacy-recency studies are directed to the question of whether attitudes and retention differ significantly between material pre- sented early and late in a mesage. Climax-anti-clinax studies are directed to early or late placement in a message of rhetori- cal'ly strong and weak assertions md attending supporting material. Here, again, both attitudes mid retention are indexed as a func- tion of order. For clarity, a further breakdown of the prinecy-recency studies is useful. Studies concerned with attitude changes uploy mats on both sides of an issue. Studies concerned with re- tention employ a set of statements with reverse clusters of early and late statements. In 1925, Land” postulated a "law of primacyn out of a study of two-sided written messages. He indexed the attitudes of his subjects prior to the presentation of the pro and con argu- ments. He noted that attitudes shifted in the direction advocated first and did not shift back completely after the other side was . presented. This primacy effect was noted whether the pro or con arguments were presented first, although levels of significance were not computed. 17F. H. Lund, ”The Psychology of Belief: IV. The Law of Priucy in Persuasion, " Journal of Abnormal and Social Psycholog, n (1925), 183-191. 12 Cromwell18 obtained significant and contradictory findings using oral messages on both sides of an issue. I'When strong affir- native and negative arguments on the same proposition are presented» in recorded speeches rated as equally effective the speech pre- sented in the second position has the greater influence on the attitude of the listeners."19 This recency effect was noted whether the pro or con arguments were presented first. Contradictory findings are - also reported on the variable of retention as a function of order. Jersild2° supports the Lund findings. He prepared 70 distinct statements of fact as a bio- graphical account of a person with 10 different arrangements of the 70 statements. For all ten arrangements he compared the reten- tion of the first three statements presented with the last three statements presented. He concluded, "Primacy stands out as far superior to recency as an aid to the recall of meaningful material . . ."21 In essentially a replication of the Jersild study, mirensberger r22 found a favoring of recency for retention. He pre- pared a list of 52 narrative statements about the economic, poli- tical, geographical, and religious background of a people from a 18Jamey Cromwell, "Relative Effect on Audience ittitude of the First Versus the Second irgume ntative Speech of a Series," Mob Homage , XVII (June, 1950), 105-122. J9Ihid., 121. 20A. T. Jersild, "Primacy, Recency, Frequency and Vivid- ness," Journal of Experimental Wong, III (1929) , 58-70. 21min, 69. ”hey Ehrensberger, uh Experimental Study of the Relative Effectiveness of Certain Forms of Emphasis in Public Speaking, Speech New III (191:5). 9h-lll. 13 country unfamiliar to his experimental subjects. He selected five statements appearing within the first ten statements to index primacy and five statements within the last ten statements to index recency. According to his results, "Recency stands out as defi- nitely superior to primacy as an aid to recall."23 me enerimental findings with regard to a climm: versus an anti-climax order are equally contradictory. Here, however, there are also differing conceptualizations of climmc-anti-climax order. Sponberg defined climax order as " . . . that arrangement of materials in which the argument receiving the least space- emphasis is placed first, and the argument receiving the most space-emphasis is placed last."2h The definition of an anti- clinax order is, of course, the reverse of the above. He pre- pared a recorded speech alloting 8 minutes and 20 seconds to the large argument and 3 minutes to the small argument. For both attitude change and retention, he found the anti-climmc order more effective. In essentially two replications of the Sponberg stuck, however, Gilkinson, Paulson, and Sikkim:25 detected no significant differences between the two orders. Cromwell, in another section of the study noted earlier, bfined climax and anti-cling order in terms of a weak argumenta- tive speech and a strong argumentative speech on the same side of 23Ibid., 111. zhflarold Sponberg, "The Relative Effectiveness of Climn and Anti-Climax Order in an Argumentative Speech," Speech New XIII (191:6). 37. 25Howard Gilkinson, S. F. Paulson, and D. E. Sikldnk, "Effects of Order and Authority in an Argumentative Speech," garterly Journal of Speech, 11. (April, 1951;), 183-192. It a proposition. Here, the weak speech followed by the strong speech would be a climn order sequence. His criterion of strength was based on judgments of effectiveness w one group of his eXperimen- tal subjects. Cromwell was concerned only with order effects on listener attitudes. His findings support a climax order for effec- tiveness. He reports, "When oral arguments are presented on the same side of the proposition in recorded speeches rated as "possess- iug unequal strength, the greater cmnulative effect on the atti- tude of the listeners occurs when the weaker speech is presented first and the strong speech is presented second."26 Culley and Berle” defined climax-anti-climax order in terms of assertion strength within a single message. They pre- pared a number of assertive statements in support of a proposi- tion; selected 10 assertions on the basis of sorting by teachers of public address; and chose five of these on the basis of rank- ordering by 50 subjects chosen from the same papulation as the experimental subjects. They prepared three recorded messages with the assertions and attending evidence placed in a l-2-3-h-5 order, S-h-a-z-i order, and a 5.3.1.24; order ("1" defined as strongest)‘, may found no significant differences for either a climax, pyra- midal, or anti-climax order on the variable of attitude change or retention. No clear evidence erxists for the effect of order differences on attitude and retention when they are viewed either in terms of “Citation, op. cit., 121. 27Halbert E. Culley and David K. Berle, "Effect of Inter- cellnlar and Intracellular Speech Structure on Attitude Change and Learning,a Speech 110mm, XXIII (August, 1956), 288-297. 15 primacy-recency or climax-anti-clinax structure . Other background literature for the present study exists, and a more complete pic- ture of the emergence of the present study is obtained by examining research dealing with the presentation of one or both sides of a controversial issue. Related Studies A study on the effects on attitude change of a one sicbd or two sided presentation of a controversial issue is reported by Hovland, Lumsdaine, and Sheffield.28 They were interested in differential effects due to initial attitudes favoring or Opposing the position advocated in the experimental message. Stated in terms of questions investigated, there were two: Do epposed lis- teners mentally construct arguments to refute one-sided speakers? Does stating arguments on the |'other sitb' weaken the effective-1 ness of the side advocated? In constructing their experimental message they assumed that arguments on the nother side” should be presented early in the speech, and attempts to refute those argu- ments made late in the speech. he stumr was conducted during the Second World War, and United States Am information policy required the omission from the opposed arguments certain material that proved, through later analysis, to be relatively salient for those subjects initially Opposed to the position advocated. Three conclusions drawn from the study bear on the present investigation. 2-8 0 I0 HOVMd, A. ‘0 mm, md Fe Do Sheffield, PEMB on Mass Cominication Studies in Social Psycholo in or ar , , atom ton University Press, , 16 l. Presenting the arguments on both sides of an issue was found to be more effective than giving only the arguments supporting the point being made, in the case of individuals who were initifl omsed to the point of view being pres e . 2. For men who were already convinced of the point of view presented, however, the inclusion of arguments on both sides was less effective, for the group as a whole, than presenting only the arguments favor- ing the general position being advocated. 3. An important incidental finding was that the absence of one relevant argument against the stand taken by the programs was more noticeable in the presentation using arguments on both sides than in presentations in which only one side was discussed. Furthermore, advantage of the program giving both sicbs among men initially opposed was less for those who regarded the omitted argument as an important one.” l-"aulson,30 in a study of one and two-sicbd speeches, also constructed an experimental message ordered with a presentation of opposing arguments followed by favorable arguments strongly sup- ported. The one-sided speech simply omitted the opposing arguments. This order parallels the procedure of the Hovland, Inmsdaine, and Sheffield etuov. Paulson detected no significant difference be- tween the two speeches in producing a shift of opinion toward the side advocated. For women subjects, retention scores did not differ significantly. However, for the men in the audience, the two-sicbd speech produced significantly higher retention; and the greatest difference in retention between the two speeches was noted for those male subjects initially opposed to the position advocated. 29J:Ioid., 223-225. 30Stanley F. Paailson, «me Effects of the Prestige of the Speaker and Acknowledgnent of Opposing Arguments on Audience Reten- tion and Shift of Opinion,“ Speech Maxim XII (November, 19514), 267-271. . 17 One result of a study by Ihistlethwaite, Kemenetsky, and Schmidt31 was the test of the assunxption that material which re- futes opposing arguments should be placed late in a speech contain- ing opposing arguments. They attacked the problem through a con- corn with the question, “Is a speech more effective when the speaker avoids explicit refutation of the arguments of apposed members of the audience, or is it better to present direct refu- tation of opposed arguments?"32 They selected assertions which refuted arguments known to be relatively salient among members of the audience, and for the one-sided speech left unmentioned the arguments to be refuted. I'In the two-sided presentation each argu- ment to be refuted was explicitly formulated and briefly elaborated before presenting the relevant refutative material."33 In a second experiment, also reported, they were concerned with the effects of presenting explicit refutation early or late in the speech. lime, their both-sided versus one-sided speech with the placement of explicit refutation early or late in the speech paral- lels the studies by Hovland, Lumsdaine, and Sheffield and Paulson. They concluded from their stuchr, "Although the speeches as a whole were effective in modifying Opinions, there was no evidence that explicit and implicit refutations, or one-sided and two-sided presentations, were differentially effective in changing opinions ”Donald 1.. Thistlethwaite, J. Kemenetsky, and H. Schmidt, "Factors Influencing. Attitude Change Through Refutative Comica- tinns," gmech Monograms, XXIII (March, 1956), lh-25. 321mm, 11;. 331b1d., 17. 18 of Opposed subjects."3h The authors did recognize that the short (five and one half minutes) time span between earlyr and late place- ment of explicit refutation could account for the lack of signifi- cant differences. Conflicting evidence, then, is available for - the thesis that two-sided presentations with late support for the side advocated is more effective in changing attitudes and in- creasing retention. It is from the foregoing literature that the present study has emerged. Consistent findings as to the effect of an indivi- dual's frame of reference on his attitudes toward and retention of controversial material seem to Justify anmrexamination of the variables of attitude and retention as a function of order, given, in general, that they vary directly with one another. The litera- ture reveals unsuccessful efforts to index consistently the effect of order on attitudes or retention. This problem confronts the early reports of primacy-recency differences and later reports of climax-anti-climax differences. Other studies related to climax- anti-climax order, because of a concern with one or two-sided pre- sentations, fail consistently to index order differences. It seems 'that continued efforts to index order differences will prove fruit- less unless a new approach is taken. Justification [orthwhileness of the Present Shggy ‘ J The purpose of commnication is to change, ovartly or co- vertly, the behavior of the receiver. In am given speech situation 315932., 2h—25. 19 many complex variables affect the speaker's successful accomplish- ment of his specific purpose. Some of these variables are poten- tially under the control of the speaker. A legitimate end of re- search is the explication of how such variables do Operate. Ellie paucity of consistent findings reported in the literature reveals through their number a continuing interest, but also a lack of success, in explication. Partial Justification for the present study rests in the additional effort to index the variable of order. The utility for potential speakers remains obvious if significant differences in attitudes and retention attributable to order exist. Further jus- tification seems to demand that a new approach be taken. The pre- sent study is an attempt to investigate the variable of order within a different theoretic context. If the logic of this approach is Justified, then the study seems justified. . -' f be validity of the chosen theoretic approach emerges from the inquest of a "frame of reference" on attitudes and retention. For individuals with a marked attitude toward an issue, exposure to a two-sided message on that issue results in at least two measure- able phenomena. Attitudes toward the issue remain much as before exposure. Retention of the content Of the compatible side is con- sistently higher than of the incompatible side. Order differences, if they exist, do not alter the relationship between attitucb and retention on this gross level. it a finer level of analysis, how- ever, it is obvious that attitudes can vary within a favorable range, for instance, and retention Of compatible material can vary and still remain higher than retention of incompatible material. 20 Here, order differences might be noticeable. For example, favor- able attitudes toward an issue as measured after aposure to a two-sided message on that issue might be significantly different for two comparable audiences who heard the message in a different order. Similarly, retention of compatible material might be sig- nificantly different. These possibilities are consistent with the 'frarne of reference" studies and still justify an attenmt to in- dex the effect on attitudes and retention of order differences in controversial material. Distinctiveness of the Present Stacy A brief analysis of the existing literature reveals the potential value of examining order from a different point of view. Beginning with the consistent findings of the "frame of reference" studies, it is possible to hypothesize at least partial resolution of the inconsistent findings on order. The studies on "frame of reference" were not interested in the order of presentation as an experimental variable. Watson and Hartmahn35 used ten sets of paired pro and con arguments. Edwards36 did not vary the order of presentation. Levine and Murphy used pro and con arguments of unequal difficulty and stated, " . . . results on the two paragraphs cannot be meaningfully compared."37 while Gilkinson, Paulson, and Sikkink38 presented their experimental 35Watson and Hartmann, Op. cit., Blur-335. 363m31‘d8, EEO Cite’ 314-50. 37Levine and mm, op. cit., 509. 38dilhinson, Paulson, and Sikkink, "Effects of Order and Authority 0 e a," CEO Cite, 183.1920 21 message in various orders (mixed, pro-con, con-pro), they did not obtain an initial measure of attitude, and thus no apprOpriate breakdown of their data is possible. The studies by Lund and Cromwell of primacy-recency effects on attitudes were interested in the variable of order, but did not consider pro and con material as compatible or incompatible for the audience. It is possible that such an analysis would reveal a con- sistency between the findings of hand39 and Cromwell.ho Both in- vestigators considered the reversal of pro and con arguments as a message control factor rather than an audience experimental factor. The analysis of the latter view of order is, as mentioned earlier, a major justification for the present stuch'. One experimental treatment of a study of attitudes by Knowerm' is relevant to the Lund and Cromwell findings and to the view of order in the present stucw. Unfortunately, as Knower states, no conclusions can be drawn from his report. However, as one variation in a continuing stucw of attitudes, Knower did sub- divide his experimental subjects according to initial pro or con attitude toward a preposition aid tested for attitude changes attributable to pro-con and con-pro orders. He did not test for differences in retention; and his data on attitudes, according to him, are only "suggestive." He interpreted the method of treatment as primacy-recency. However, it can be interpreted as a precedent 39Lund, loc. cit. hocronnell, op. cit., 105-322 . “arsenals H. Knower, ”Experimental Studies of Changes in Attitude: II. A Study of the Effect of Printed Argument on Changes in Attitude," Journal of Abnormal and Social 13mm m (1936), 522-532. 22 fer a study'of‘the effect of varying compatible and incompatible material on attitudes and retention. The inconsistent findings of Jersildh2 and.Ehrensbergerh3 also seem to Justify a different approach to the effect of order differences on.retention. Neither study dealt with controversial material. It is possible that fer the Jersild.study, interest 'waned.as 7O biographical statements were read, and.for the iEhrensberger study, interest picked up as more information on an unknown country was read, thus accounting for the difference in primacy and.recency. Such an interpretation lends support to a viewing of'order as a.message-receiver interaction variable. The present study, while related.to the studies on climaxpanti-climax and one or two-sided.presentations, is Obviously different in approach. It is nmdnmained.that the preceding analysis justifies the present effort to index some effects of order differences on atti- tudes and.retention. View of the Dissertation The statement of the problem'undsr study, its placement in context'within.the literature, and a justification.for its investi- gation serve as a foundation for this dissertation. The ensuing chapters detail the rationale, methodology, and.outcome of the study. ‘Within.Chapter two is fOund.a.fonmal statement of three lti’Jersild, g. cit., 58-70. l‘3‘Ehrensberger, gp. cit., 9h—lll. 23 hypotheses. The rationale fer each hypothesis includes an analysis of’relevant literature and reasoning based.on that analysis. In Chapter three the methodology and.the experimental design.utilized are explicated. The experimental messages, the construction of attitude and.retention forms, and the process of data.collection are reported in detail. Chapters feur and five, respectively, treat of the results of the statistical analysis of the data and.the COD! clusions warranted.by this analysis. Included.in the appendices accompanying the report are manuscripts of the experimental.mes- sages and facsimiles of the attitude and.retention schedules. It is heped.that the design and.rationale of the present study are sufficient to provide some clarification of the effect of certain order differences within a.message on the attitudes of the receiver toward.the subject of the message and on the retention.by the receiver of the content of the message. CHAPTER II HIPOTHESES AND RATIONAIE The purpose of this chapter is to explicate the three hypo- theses of this study and the particular rationale for each. The hypotheses predict specific order effects under differing conditions. he rationale for each hypothesis is derived from the literature, a pilot study, and relationships among impotheses. mmeses I. when an audience is presented with a message containing equal discussion of both sides of an issue, initial exposure to the side compatible with the audience's existing attitude results in a more extreme attitude toward the topic of the speech than results if there is initial exposure to the incompatible material. Corollary A. Given that the audience has a favorable initial attitude toward the message tapic, initial exposure to the compatible material results in a more favorable attitude toward the tepic of the speech than results if there is initial exposure to the incompatible material. Corollary B. Given that the audience has an unfavorable initial attitude toward the message topic, initial eaposure to the cmpstible material results in a more unfavorable attitude toward thetopicefthe speech thanresultsifthereisinitialezposureto the incompatible material. 21; 25 II. when an audience is presented with a message containing equal discussion of both sides of an isms, the material compatible with the audience's existing attitude will be better retained than will the material which is incompatible. III. When an audience is presented with a message containing equal discussion of both sides of an issue, the difference between the retention of compatible and incompatible material will be greater when the compatible material is presented first than when the incompatible material is presented first. Rationale Hypothesis I accounts for certain attitude changes for individuals initially favorable or unfavorable to the topic of a controversial issue. It does not treat of initially neutral atti- tudes for individuals. For two individuals, favorably disposed to the topic, both of whom are exposed to a two-sided speech on the tapic, the first in a pro-con order and the second in a con- pro order, the hypothesis prédicts that the attitude of the first individual will tend to become more favorable while the attitude of the second individual will tend to become less favorable. For two individuals unfavorably disposed to the tepic, both of whom are exposed to a two-sided speech on the topic, the first in a con- pro order and the second in a pro-con order, the Impothesis pre- dicts that the attitude of the first individual will tend to become less favorable while the attitude of the second individual will tend to become more favorable. Thus, the pro material is compatible for the initially favorable individual; and the con material is com- patible for the initially unfavorable individual. For both indi- vidu als, the compatible-incompatible order results in a tendency 26 toward a strengthening of the initial attitude. Several factors seem to provide a rationale for this hypothesis. "Objectivity" might function differentially within an indi- vidual. Given that a listener does not know before the conclusion of obs side of a speech that the other side is to follow, his "objectivity" might be influenced by the nature of the first side presented. The listener, on exposing himself to material incompa- tible with his existing attitudes, might be less likely wholly to accept subsequent compatible material. He might be more "objec» tive" having just attended to the "other side" of the issue. After initial exposure to compatible material, "objectivity" might fail him as he is exposed to incompatible material. The immediately preceding exposure to compatible material makes the "other side" vividly incompatible. This reasoning supports the prediction formulated in Hypothesis I. The stucbr by Hovland, Inmsdaine, and Sheffieldl seems to supplement the line of reasoning ennan "objectivity effect." meir study concerned the early placement md‘ late refutation of material that is in conflict with the position advocated. For receivers initially favorable, masdnmm effect occurs when there is no mention of incompatible material. For receivers initially apposed, early though brief mention of compatible material might indicate tru st- worthiness, whereas with further elaboration the arguments migxt "take hold" and thus reinforce attitudes sufficiently to preclude an objective attending to subsequent incompatible material. Their study laminae, Lumsdaine, and Sheffield, op. cit., ash-307. 27 was not concerned with the differential effects of compatible ma- terial or with equal treatment of both sides of an issue, but it indirectly lends some support to the statement of Hypothesis I. A pilot studyr conducted for the present investigation pro- vides further support for Hypothesis I. Two groups of subjects were presented a speech containing equal discussion of compatible and incompatible material. For the group of 21 hearing the speech in a compatible-incompatible order, 15 shifted to a more favorable attitude position or remained as favorable as before the speech to the tepic discussed, while 6 shifted to a less favorable attitude position. For the group of 29 hearing the speech in an incompatible- compatible order, 20 shifted to a less favorable attitude position, while 9 remained as favorable as before the speech or shifted to a more favorable attitude position. The above line of reasoning and evidence seems to provide an adequate rationale for the prediction of Hypothesis I. Hypothesis II predicts that listeners with an existing atti- tude, either favorable or unfavorable toward the subject of a message containing equal discussion of both sides of an issue, will tend to maintain that attitude and will tend to retain better the compa- tible segment of the message. This is a refornmlation of the gen- eral hypothesis that for controversial material, attitudes vary directly with retention. is was noted in the Edwards' study, two- sided presentations of controversial material do not seem to change attitudes . The favorable listeners remain favorable, and the un- favorable listeners remain unfavorable. The general hypothesis has 28 been confirmed by Watson and Hartmann,2 Edwards,3 Levine and Murphy," and Gilkinson, Paulson, and Sikkink.S In eight experi- mental treatments in the pilot study this effect was noted, al- though levels of significance were not computed. The consistency of findings on the general hypothesis might argue that w further testing should be incidental to other con- earns. However, Hypothesis II is considered a major hypothesis of the present study. It provides a logical relationship between the statement of Hypotheses I and III. If, for controversial material, attitudes vary with retention on a gross level, then shifting of attitudes attributable to order as predicted in Hypothesis I might be paralleled by a change in the retention level of compatible material as predicted in Hypothesis III. The above evidence and reasoning seem to provide an adequate rationale for the prediction of Hypothesis II and for its inclusion as a major hypothesis of the study. Hypothesis III predicts that the difference between the re- tention by a listener of compatible and incompatible material will be greater when the compatible material is presented first in a message containing equal discussion of both sides of an issue. The rationale for this hypothesis is based on reasoning from Hypotheses I and II. 2Watson and Hwhnann, Op. cit., 31’4-335. BEMdB, OE. Cite, 3’4-500 1"Levine and Murphy, Op. cit., 507-517. 5(lilkinson, Paulson, and Sikkink, "Controversial Statements 0 o 0" CE. Cite, 253.260. 29 According to Hypothesis II, attitudes (b not shift from favorable to unfavorable when both sides of an issue are presented because attitudes vary directly with retention, which is always greater for compatible material regardless of order. Hypothesis I predicts significant shifting of attitudes within a given range of attitudes as a function of order. Retention measures may not be sensitive enough to detect significant differences attributable to order. Iet, since both compatible and incompatible material are, to some extent, retained, the advantage gained through order- ing should rest in increasing the differences between the reten- tion of compatible and incompatible material. This is the basis for the statement of Purpothesis III. ' In sumnary, three hypotheses have been presented with an attending rationale for each. In the following chapter, the experi- mental procedures employed to test these hypotheses are detailed. CHAPTER III EXPERDIENTAL PROCEDJ RES The purpose of this chapter is to detail the experimental design and the procedures followed for data collection. The study was carried out during May, 1961 at Kansas State University, Manhattan, Kansas. Subjects All the experimental and control subjects were students enrolled in Oral Connmmication I, an all-University requirement usually taken during the freshman or sOphomore year. Almost all of the 337 subjects were either Freshmen or Sephomores and 17-20 years of age. Tuenty-four sections, consisting of approximately 20 students each, were randomly assigned to one of four superinten- _ tel or four control groups. The subjects were tested within regular class hours. Experimental. Desifl The following design was considered adequate to test. the three experimental hypotheses and to treat of the questions on the effects of message saliency and speaker credibility. The question of the effect of message saliency seemed to justify the construction of two experimental messages which differed only in the number of audience reference terms (see Appendix A). The question of the effect of speaker credibility demanded for its 30 31 investigation only the use of a speaker unknown to the audience and a post-test index of the degree of credibility assigned to him by the subjects in the different experimental treatments. These requirements plus those dictated by the three hypotheses are in- corporated into the emerimental design. A pre-test-post-test design made available data on initial attitudes and attitudes following exposure to an experimental message (or, for the control groups, after a comparable period of time). The test instrument to measure retention was administered during the pre-test to index ability differences with the instru- ment and to provide equal warm-up for all subjects. It consisted of a passage from a speech unrelated to the experimental message. Re- tention was, of course, indexed following exposure to an experimental message and an identical measure was administered to the control groups. In addition, general. demographic information was available through the pre-test forms and an index of saliency and credibility was available through the post-test forms. Four experimental groups received differing orders of the two experimental. messages. There was a pro-con ordered and a con- pro ordered salient message and a pro-con and a con-pro ordered non- saJient message. Four control groups were established and charac- terized by being presented retention forms in the same order and of the same saliency as the four experimental groups. For purposes of analysis, the eight basic groups were sub- divided according to initial attitude of subjects into favorable, neutral, and unfavorable categories. Thus, twenty-four data cells were established. 32 To test Hypothesis I, attitude change scores (post-test attitude score minus pre-test attitude score) were computed for the different experimental groups and a t-test of the difference be- tween attitude change scores was computed for the experimental groups receiving a pro-con ordered message and those receiving a con-pro ordered message. With the breakdown made according to initial attitude, Corollary A to Hypothesis I was tested with the favorable group and Corollary B was tested with the unfavorable group. An identical analysis was made for the neutral group. To test Hypothesis II, retention scores for both compati- ble and incompatible material were analyzed by the use of t-tests of the difference between the retention of compatible and incom- patible material. To test Hypothesis III, retention difference scores were computed for the compatible-incompatible ordered messages and the incompatible-compatible ordered messages using standard scores for both compatible and incompatible material. Using these difference scores, t—tests of the difference between the retention of the com- patible-incompatible and the incompatible —compatible treatments were computed. To determine the effect of message saliency on attitudes and retention, t-tests of the difference between attitude change scores, regardless of order of presentation, were computed for the groups receivingathe salient experimental message and the groups re— ceiving the non-salient experimental message. Similarly, retention scores were tested for the different saliency treatments. In addi- tion, t-tests of the difference between perceived salience (as 33 indexed in the post-test fer’the experimental groups) were computed for the groups receiving the salient and.non-salient messages. To determine whether the order of presentation affected.the assignment of credibility to the speaker by the experimental sub- jects, two indices of credibility were incorporated into the post- test. T-tests of the difference between.perceived."knowledgeable- ness" and."bias" were computed for the groups receiving the comp patible-incompatible ordered message and the groups receiving the incompatible-compatible ordered.message. For the completion of this study it was necessary to conp struct suitable experimental messages, indices of attitude and re- tention, and indices of saliency and credibility. These materials need to be described and.explained in detail. Materials gaperimentallMesggggg The experimental messages were used.in the study (see ‘AppendixeA). The tepic, "Social.Fraternities and Sororities at KSU," was selected.and a I'salienrt" message constructed. The follow- ing comparisons note the changes made in the messages so as to pro- vide a.”non-salient“ message on the same topic. The right hand column below contains the words substituted.for the corresponding words in the left hand.oolumn.in.order’to change the saliency of the message. Salient Nun-salient Here on campus On University campuses IHanhattan University community Your campus Campus members at Kansas State College Students Kansas State College 3h 8:00 o'clocks in Eisenhower Hall Regular class attendance Kansas State College Kansas State His school System at Kansas State System Two local chapters Some chapters I-Orpheum Carnivals and shows "Collegian" Record Except for the addition of the sentence, "This problem is not easily solved within the present situation.2 in the non-salient message, the above pairs of words constitute the differences in the two messages. The additional sentence was necessary to equalize the length of both messages. It can be seen through an inspection of Appendix A that for either message the introduction, transition, and conclusion are con- structed to accommodate the reversing of the pro and con segments of the message. An original tape recording of both the salient and non- salient messages was made in a pro-con order by a faculty member of the radio-television area of the Department of Speech at Kansas State University. Each tape was then used to record, by editing, a second tape in a con-pro order. For the study, then, there were provided four taped messages. They were labeled as follows: Non-salient Pro Con Order Non-salient Con Pre Order Salient Pro Con Order Salient Con Pro Order The speaker was a new faculty member on campus (less than four months) and was not identified by the experimenter or on the tape recordings. It is assumed that recognition of the speaker by subjects was negligible. Attitude and Retention Measures Attitudes toward the experimental tepic were obtained two weeks prior to the presentation of the experimental message and 35 immediately after the presentation (see Appendix B). The measure used employed the semantic differential form as developed by Osgood} Three evaluative scales were selected: good-bad, like- dislike, fair-unfair. In both the pre and post measures these scales were masked by other non-evaluative scales. The pre-test measure was also masked by five other topics. For this study the seven blanks between each pair of words were assigned the numbers 1-7, one being nearest to the "favorable" word and seven nearest to the "unfavorable” word. It was thus possible to sum the three scales and arrive at a single index of attitude for each subject. For the experimental tepic, a score of 3-10, inclusively, was considered to be a "favorable" attitude. A score of 11-13, inclusively, was considered to be a neutral attitude. A score of 114-21, inclusively, was considered to be an "unfavorable" attitude. The possible significance of attitude strength as a variable in the study was anticipated. Thus, for “favorable" atti- tudes, 3-6, inclusively, was considered a strong attitude md 7-10, inclusively, was consichred a weak attitude. For the "unfavorable" attitudes, 113-17, inclusively, was considered a weak attitude, and 18-21, inclusively, was considered a strong attitude. Retention of the experimental messages was indexed by the "Close procedure" deve10ped by Wilson Taylow.2 This procedure was ICharles E. Osgood, George J. Suci, and Percy Tannenbaum, The Measurement of Meaning (Urbana: University of Illinois Press, Zwilson Taylor, “Application of 'Cloze' and Entrepy Measures to the Study of Contextual Constraint in Samples of Continuous Prose'I (unpublished Ph. D. Dissertation, Dept. of Mass Conmmnications, University of Illinois, 1951:). 36 selected primarily because it permits the use of realistic messages. Traditionally, mltiple-choice or true-false tests are utilized to index the retention of experimental messages. In order to deve10p a reliable number of questions, extensive factual material must be included in the message, or, if that is not possible, series of assertive statements must be included. In the present study, after completion of the salient and non-salient form of th e experimental message, a 250 word passage from both the pro and con segments of each message form was selected. The retention fom contained a mutilated presentation of these passages. That is, every fifth word was deleted and replaced with a standard sized blank (see Appendix B). This method provided four message forms, each con- taining 50 blanks. The number of correct insertions by a subject constituted his retention score. In order to provide a uniform lWarm-mp" with the prece- dure for all. the experimental subjects and to provide an index of "close ability," 3. message independent of the experimental mes- - sage was made part of the pre-test. "Close ability" could signi- ficantly alter retention scores as the ability to restructure a passage in which every fifth word is deleted is, in part, inde- pendent of a specific and recent exposure to that passage. For this reason, a subject obtaining a close score on the pre-test above the mean of all pre-test oloze scores was considered to have "high close ability" and a subject obtaining a close score on the pre-test below the mean of all pre-test close scores was considered to have "low close ability." A passage taken from a collection of speeches by Albert Schweitzer (see Appendix B) was employed to provide a 37 uwarm-mp" with cloze procedure and an index of close ability. Credibility and Saliency Measures Included in the post-test form were additional semantic differential scales to index speaker credibility 31d message sali- ence (see Appendix B). A question concerning membership in a fraternity or sorority was also included. The experimental groups, after exposure to a given experimental message, were asked to rate the speech and the speaker. Two scales, knowledgeable-not know- ledgeable and biased-unbiased, were used to index credibility. Two other scales were used as masking scales. One scale, relevant-to-me- not relevant-to-me, was used to index saliency. Three other scales were used as masking scales. The two credibility. md one saliency scales were scored with the numbers 1-? assigned to the blanks be- tween the paired words. The number 1 was assigned to the blanks nearest the words "knowledgeable," "biased,” and "relevant to me." Data Collection The pre-test form, consisting of an attitude measure and a close ability measure, was administered by each section instructor. Included in the pre-test form was a request for basic biOgraphical data: mane, age, sex, year in school, and major. The students were informed by the instructors that it was to be used in a continuing study undertaken by a member of the Speech Department to investigate the awareness of people concerning various issues facing them. It was made clear that participation in the study would not be related to the assignment of grades. All students were given ten minutes for the completion of the close ability passage. None of the differ- ent instructors reported any difficulty by the students in 38 understanding what was requested of them or that there was a noticeable lack of cooperation from the students. he experimenter presented the tape recorded messages and achinistered the post-test forms. He was introduced by the sec- tions' instructors and then explained that this was the conclusion of a study that they had helped with before. Their cOOperation was again requested and they were asked to listen to a taped speech. Immediately after the completion of the tape recording, the post-test attitude, retention, credibility, and saliency form was adninistered. (See Appendix B). For the control groups, the regu- lar class instructor administered the abbreviated post-test form. This form, of course, did not include the credibility or saliency indices, nor did it include a question on fraternity membership. Eight minutes were allowed for the completion of each of the pro and con segments of the close procedure. After collection of all the data only the data were retained for analysis which consisted of both a pre and a post-test for an experimental or control subject. CHAPTERIV RESULTS AND ANALYSIS This chapter is arranged, generally, according to the order of the hypotheses under test. First, however, there is presented an analysis of the pre-test data and an analysis of the saliency factor. Analysis of the Pre-Test Data Table 1 presents a count of subjects by four factors. The subjects' pre attitude score divided the subjects into favorable, neutral, and unfavorable groups. They were assigned either to an experimental or a control group. The experimental groups were presented with either a pro-con or a con-pro ordered message, and, for each order of message, they were further divided according to the presented treatment of the message in a salient or non-salient manner. The control groups, while they received no experimental message, received attitude measures and "cloze‘I passages ordered and treated identically with those of the experimental groups. Since the subjects, in large majority, were favorable toward the experimental tepic, the pre-test analysis was made on that group. Eight different treatments are noted in Table l for the favorable group. For both the experimental and control groups, salient and non-salient messages are provided in both a pro-con and con-pro order. 39 ho hmm cm as NH mm ms and Haven. nacho sea m pm 0H 3” cm mu Hmmma “mt-2h ma m: meowmeunoz mecHHmm me mad H an as as m a m as NH me as ma EIGOU Mp : Hm um Hmwmaw m ma am an saoHHamgapz. maoaflem coouonm eoaececomcnm Ho nacho Havana .aqo OHflUHO>GH~HD e g .hno Haseeoz e E 0 go .aeuuosae e E assesses one eoaesanomonm oneness ens .oospaees ahocoflase «nacho no unoposu on» on.mnduaooos evenness no assoc huusesn «In .H Manda bl A Chi-square analysis (see Table 2) was made on the vari- ableseof close ability and pre-attitude as they related to the variables of sex and fraternity membership. For purposes of analysis, the close ability score of a subject was categorized as high or low according to its being above or below the total mean of close ability, and his pre-attitude score was categorized as either strong or weak by splitting in half the possible scores on the measure of pre-attitude. TABLE 2. -Chi-square analysis of relationships of close ability and pre-attitude and the variables of sex and fraternity member- ship Cloze Ability Score Pre-Attitude Score High-Low it: Strong—Weak 11: Sex .83 l 3.05 l Fraternity Membership 6.2Kfl 1 21.97” 1 5 *fi .99 (d! . 1) - 6.61], Fraternity membership is related to cloze ability and pre- attitude, although the variable of sex does seem to be independent. As the presentation of all eight treatments was randomly distributed among subjects, the variable of fraternity membership was randomly distributed.1 Thus the effect of this noted relationship tends to lumbership in a fraternity of sorority was indicated only by those subjects who were exposed to the experimental message. How- ever, assignment to both the experimental and control groups was random. h2 increase the variability of obtained data. This variability tends to) produce a more conservative set of findings with respect to the hypotheses under test. - Four Chi-square analyses were run to test for differences among the eight groups on the variables of sex, cloze ability, pre- attitude, and fraternity membership. (See Table 3). TABIE 3. --Distribution of four variables across four experimental and four control groups 23’. * 91-: Sex 6.81; 7 Clone Ability 10.23 7 Pre-Attitude 3.25 7 Fraternity Membership 1.96 3 *12,95 (d: - 7) - use? X2.9S (df ' 3) - 7.82 As the analysis reported in Table 3 revealed no signifi- cant Chi-Square, further analysis was made without sub-dividing experimental or control groups on the bases of any of these four variables. Analysis of the Saliency Factor Three separate tests determined the effect of varying the saliency of the experimental message. The analysis of the saliency factor was made on the "favorable" group of experimental subjects. The four groups of experimental. subjects were asked to judge the relevance of the experimental message presented to them. Judg- ments were grouped according to the order of presentation and h3 according to the saliency of the message. Table 1; presents the re- sult of an F-test among the four groups. The scales on which the reported measures were obtained has 0 as neutral relevance, +1, +2, +3 as increasing relevance, and -l, -2, -3 as decreasing relevance. TABIE h. «Perceived relevance differences among different message orders and saliency differing messages Mean Relevance S .085 S 4’ng P-C C-P NS +.7l NS +.S2 Source of Variance Sum of Sgares 2: Mean Square 3* Between ‘ 2 .99 3 1.00 .27 Within $39.03 IN; 3.71: Total 51:2.02 1147 {F 95 (d-f ' 11411, 3) ' 8055 The different messages and order of presentation did not significantly affect the judgment of message relevance. Attitude change scores for the non-salient and salient groups were analyzed (by t-test) for possible differences. (See Table 5.) The reported mean changes are interpreted as 9 indica- ting no change from pre-test to post-test. A positive number in- dicates a change to a more favorable position and a negative num- ber indicates a change to a less favorable position. The range of the attitude scale is +3, +2, +1, 0, -l, -2, -3. There were no significant differences in attitude scores attributable to the saliency of the message. hh Finally, retention scores of both pro and con message seg- ments for the non-salient and.salient groups were analyzed.(by t-test) for possible differences (see Table 6). TABLE 5. --Attitude change difference between non-salient and salient groups Prquttitude Mean Mean Change Fayorable * .222_~Gr s. a. m. a r P-C +2.16 +2.29 -.h0 -.22 73 .37 0-? +2.08 +2.21 -1.75 -.91 71 1.07 it t.95 (df - 71) - 2.00 TABLE 6. -éRetention differences between non-salient and salient STOBPB Favorable * ___2__Grou a as. 22 2 P 19.00 20.67 1.11 P—c 73 g C 17.26 20.16 2.36 Exp. P 20.78 23.0h 1.39 0-? 71 C 17.39 19.29 1.57 P 13.87 16.2h 1.53 P-C h3 C 33 .92 11:052 053 Cnt e P 17.17 16.12 .73 0-? 28 C 15.25 1h.55 .57 a. 15.95 (d! - 73) " 2.00 145 A significant difference is noted in Table 6 only for the retention of con.material in the pro-con ordered messages between the salient and non-salient groups. It was decided to combine the salient and.non-salient groups fer analyses of the hypotheses. Test Effects To determine whether exposure to the experimental message significantly affected attitudes and retention, a comparison was made between experimental and control groups. The analysis of the effect of exposure to the experimental.message was made on the "favorable" group of experimental subjects. First, an analysis (by t-test) compared attitude change scores (pre-attitude minus post-attitude) for experimental and.control groups (see Table 7). Again, the range of the attitude scale is +3, +2, +1, 0, -l, -2, -3. For the mean change scores a negative number reflects a change to a less favorable position. TABLE 7. --Pre to post-attitude change differences between experimental and control groups Prquttitude Mean, Mean Change Favorable * ____r._Grou Em- smo an- 92.2- 93 2 P-C +2.22 +2.26 "9.331 “1009 118 1073 6"? +2.13 +2001]. '1023 “077 101 1002 a t (d: ’ 101) - 1098 .95 Exposure to the experimental message did not significantly change over-all attitudes. 176 A similar analysis compared retention differences between experimental and control groups (see Table 8). Exposure to the experimental message did result in signi- ficantly higher close scores. Further analysis was made only on the experimental groups. TABLE 8. «Retention differences for both pro and con material between experimental and control groups Mean Retention Favorable H __ Group Experimental Control 9.3 t P-C 19.83 117.98 118 has“ Pro Material fl 0-2? 22 .18 16. 53 101 S .23 *3!- P-C 18.69 111.20 118 5.33 Con Material ** 0-? 18.56 117.83 101 h.05 as- t - 0 - 2.6 .99 “If 1 1) 3 hypothesis I According to Hypothesis I there should be a significant difference between the attitude change scores for those subjects ex- posed to the pro-con order of presentation and those subjects ex- posed to the con-pro order of presentation. Specifically, for the subjects favorable toward the experimental tepic (Corollary A), there should be less change from pre- to post-attitude scores (or it should be a positive number, reflecting a movement on the 3-21 scale toward the low numbered or favorable and of the scale) when those subjects are exposed to the pro-con order. Similarly, for those subjects unfavorable to the experimental topic (Corollary B), h? there should be less change from pre to post-attitude scores (or it should be a negative number, reflecting a.movement on the 3-21 scale toward the high numbered.or unfavorable end.of the scale) when those subjects are exposed to the con-pro order. Attitude change scores for the pro-con and.con-pro groups were analyzed (by t-test) in order to test Hypothesis I (see Table 9). Incidental analysis was made of the neutral group and it also is reported in Table 9. The reported mean.changes are in- terpreted as a 0 indicating no change from pre-test to postdtest. For the favorable group, a negative number’reflects a change to a less favorable position. For the unfavorable group, a.positive number reflects a change to a less unfavorable position. For the neutral group, a positive number reflects a change to a favorable position. The range of the attitude scale is +3, +2, +1, 0,'-1, -2, .3, TABLE 9. -Pre to post-attitude change differences between pro-con and conepro ordered treatment groups arranged according to initial attitude Prquttitude Mean Mean Change Initial * Attitude 2:6. 9:: 21-2 2:2. 22 2 Favorable +2.22 +2.13 -.31 -1.23 rho 2.17* Neutral -.05 +.11 +.53 +1.87 31 1.77 Unfavorable-1.62 “1078 +2 .25 +2 067 L12 035 * - 6 '108 t.95 (df 1h ) 9 he For the favorable group, exposure to the pro-con ordered message did result in significantly less shifting away from an initially favorable position. For the unfavorable group, exposure to the con-pro ordered message did not result in significantly less shifting away from an initially unfavorable position. Thus, support is obtained for Corollary A of Hypothesis I, but there is no support obtained for Corollary B. Hypothesis II According to Hypothesis II, the retention of compatible material should be greater than the retention of incompatible material regardless of order of presentation. Specifically, sub- jects favorable toward the experimental topic should retain greater pro material than con material. The reverse should occur for ini- tially unfavorable subjects. Retention scores of both pro and con material for initially favorable and unfavorable subjects were analyzed (by t-test, cor- rected for correlated means) in order to test Hypothesis II (see Table 10 ). Incidental malysis was also made of retention differ- ences between pro and con material for those control subjects favor- able toward the experimental topic. Since the retention form was, in fact, a framentcd message, the control subjects might be meted to complete more close blanks of the compatible segment. This analysis is also reported in Table 10. For the favorable group, the retention of pro material was significantly greater than retention of con material. The same kind of relationship held for the control group. For the unfavorable group, the difference in retention was not significant. 1&9 TABLE 10. -éRetention differences between pro and con message seg- ments Mean.Retention Initial Pro Con ‘* Attitude Material Material 51!; t ** Favorable 20.99 18.63 1h? 5.9h Favorable * Control 15.60 1h.SB 7h 2.32 Unfavorable 18.59 17.20 h3 1.65 *t (df- 7h) - 2.00 “visas-1h?) - 2.61 Hypothesis III According to Hypothesis III, the difference between.the re- tention of compatible and.incompatible material should be greater when the compatible material is ordered first than when it is ordered last. Analysis was made of the group of experimental subjects ini- tially favorable toward the experimental tOpic. For.favorable sub- Jects, there should be a significant difference between the means of the difference scores for the pro-con treatments and.the conapro treatments. Pro-con difference scores require that both pro and con scores be translated to standard scores. This was done using a mean of SO and a standard deviation.of 10. Pro retention scores were then subtracted from con retention scores. Retention difference scores for pro-con ordered.treatment groups and conepro ordered treatment groups were analyzed.(by t-test) in order to test Hypothesis III (see Table 11). 50 TABLE 11. --Retention differences between pro-con ordered treatments and con-pro ordered treatments Pro-Con Difference Mean P-C Order C-P Order 9; 3* .03 .05 1146 .02 it- t.” (df 1h6) 1.98 The different orders of message presentation did not sig- nificantly affect the difference between the retention of pro and con material. §peaker Credibility The question of speaker credibility was included in the present study because of the availability of data. Analysis was made of the factors of "knowledgeableness" and "bias." Mean scores for the two scales revealed a general attributing to the speaker of "knowledgeableness" and "objectivity." However, neither the degree of message saliency nor the order of presentation had any appre- ciable effect on the two factors. No further analysis was made. 1|».- CHAPTER V DISCUSSION The present study investigated the variable of message order. It developed from consistent research findings on the re- lationship of attitudes and retention within the context of con- troversial material. The purpose of this chapter is to discuss and smumarize the results of the investigation in terms of the hypotheses advanced. The experimental message (in both its salient and non- salient forms) was constructed to treat equally the pro and the con of the subject of fraternities and sororities at Kansas State University. However, for the experimental pOpulation, drawn from Freshman and SOphomore students at the University, it did not seem to do this. In addition, the message tOpic was selected because it was believed that an approximately equal number of students would react favorably and unfavorably to it. This did not occur. 0f the 33? experimental and control subjects, 223 expressed an ini- tially favorable attitude toward the experimental tapic. The re- sponses of these subjects were compatible with the predictions of Hypotheses I and II. An examination of the responses of the unfavor- able group of subjects does not reveal such compatibility. Initially Unfavorable Subjects With respect to Hypothesis I, order seemed to make no 51 52 difference in the change of attitude from pre to post-test. Con- siderable attitude change took place, however, and it was to the favorable side of the attitude scales. This unexpected occurance requires analysis in light of the apparent bias of the experimental message. From the response of the initially unfavorable subjects, it can be argued that the experimental message did not treat equally both sides of the tOpic. For this group, at least, the message more strongly presented the favorable side of the topic. In support of the above line of argument is the finding with respect to retention of the two segments of the experimental message by the initially unfavorable subjects. Four earlier studies revealed that retention of compatible material is greater than is retention of incompatible material. Edwards found this relation to hold for an initially favorable group of subjects who remained favorable and for an initially unfavorable group of subjects who remained unfavor- able. In the present study, there was no significant difference in the retention of pro and con material by the unfavorable subjects. It would seem from this lack of difference in retention that the noted attitude shift to a favorable position reflects a persuasive bias in the message, and also a lack of an unfavorable group of sub- jects for the test of Corollary B of Hypothesis I. The rationale for Hypothesis I links the concept of objec- tivity with the concept of compatible material. From this view, an attitude prediction is made based on the ordering of message seg- ments. Crucial to the rationale is the following of compatible material with incompatible material. Here, there should be little 53 or no change in attitude. Since the initially unfavorable group changed their attitude position, it is reasonable to argue that for them, compatible material was followed by compatible material. Under this condition, none of the three hypotheses advanced can be applied to the group. Because of this observation, and because of the small number of subjects who expressed an initially unfavorable attitude, the testing of the hypotheses utilized the favorable group of sub- jects. Initially Neutral Subjects The probability of a persuasive bias within the experimental message receives further support from the attitude response of the initially neutral subjects. The experimental message was selected, in part, to minimize neutral responses. However, 33 subjects exp posed.to the experimental message had.initially neutral attitudes. For this group of subjects, as for the initially unfavorable group, exposure to the message resulted in a shift to a favorable attitude position. While the effect of order was not significant, it approached significance. 'With the bias, at least in part, making the message a persuasive message in favor of the topic, the present finding can be interpreted as suggestive of a recency effect. It is interesting to recall from.the review of the literature that, while recency tends to dominate, measures employed have generally failed.to obtain signifi- cant differences in its favor. Initially Favorable Subjects The basic analysis of data in the present study was made from the initially favorable group of subjects. Three preliminary at 5h observations are in order before moving to a consideration of con- clusions warrented from the test of the experimental hypotheses. First, the response of the initially unfavorable and neutral subjects affects the generalizability of this study. Only to the extent that the view of message orders as compatible—incompatible ordelswill permit a generalization to audiences unfavorable toward a topic, will the present findings so generalize. Of course, fur- ther research is needed to clarify the use of the terms compatible- incompatible in place of the sets of terms Pro-con and Con-pro. Second, the favorable group of subjects remained favorable. In the present stuchr, order exerted an influence while the group as a whole remained favorable. Also, they clearly retained more pro material than con material. This fact is relevant to Hypothesis II and will be explored shortly; but it leads to a third observation. The cloze procedure did index retention differences between pro and con material. The noted significant difference between the retention of pro and con cloze passages for the initially favorable control group is understandable. The retention forms amounted to fragmentary written messages on both the pro and the con of the ex- perimental tapic. In addition to indexing what it was employed to index, the cloze procedure was particularly adapted to the present study. Of ;. :. most help was the freedom it permitted in message construction. The speech was constructed without concern for the inclusion of a great deal of factual information adaptable to the construction of a true- false or multiple choice retention test. The message, good or bad, was, in a real sense, a speech on the selected topic. SS Hypothesis I Support was obtained for Corollary A of Hypothesis I. The ordering of compatible material. first instead of ordering it last in the two-sided experimental message did significantly reduce the shifting toward a less favorable position following exposure to the message. It seems from this finding, subject to the generalizability of the study, that advice to the speaker who wants to appear "objec- tive" while favoring the side of an issue incompatible with the attitude of the audience is easier to offer than is advice to the "objective" speaker only wanting to explore an issue without chang- ing existing audience attitudes. The order of presentation of message segments in a both- sidee presentation of controversial material seems to affect the attitudes or an audience favorable toward the message topic. Hypothesis II Very significant differences were obtained in the present study between the retention of compatible and incompatible material for the favorable group of experimental subjects. This finding is in support of Hypothesis II. The test of this hypothesis was a replication of earlier studies of controversial material and is consistent with them. As was noted earlier, the consistency of results in terms of the earlier studies, lends support to the con- clusion that for this study only the favorable group of subjects perceived the message as a controversial message. For this reason alone the inclusion of Hypothesis II in the present study was use- ful. Basically, however, it served as a link between Hypotheses I 56 md III. Hypothesis III No support was obtained for Hypothesis III. From the present stucbr it cannot be said that the order of presentation in a contro- versial message affects the retention of compatible material. A question arises in light of the lack of support for Hypothesis III. It is possible that the cloze procedure is not sensitive c2: enough to tap existing order differences. If they do exist, it is clear from the obtained 3 of .02 that the procedure employed is totally inadequate. Of course, the existence of order differences is not dependent on support or lack of support for Hypothesis III. A "real" order difference might exist for retention on such a small scale that the close procedure will not tap it. Yet, the cloze pro- cedure served quite well in distinguishing between the retention of compatible and incompatible material. Certainly, in the present study, order was not found to be a significant factor in the reten- tion of controversial material. Related Questions The factor of message saliency was investigated through the use of a non-salient and salient form of the experimental message. No significant difference was detected. The definition of saliency involved varying the number of audience reference terms. Here, the number of terms employed may not have been sufficient to affect atti- tudes andretention differentially. In retrospect, this seems quite possible. It certainly does not seem reasonable to conclude from the present investigation that the qrestion of saliency is not worth-- continued study. 57 The factor of speaker credibility, likewise, received no clarification in the present study. The factor of credibility is known to be more complex than the assigning to an unknown speaker the qualities of "lowwledgeableness" and "objectivity" (or lack of "bias"). The addition of this factor in a limited fashion to the present study does seem to point up the need for concerted effort toward indexing any possible effects of order on the assignment of credibility by an audience to a speaker. Research Implications The present study used college students as experimental sub- jects. The findings in relation to Hypothesis II for the favorable group of subjects are consistent with other studies employing college students as subjects. The clearest defense of this studyis generali- zability beyond college papulations is in the use of an experimental topic adapted to the interests of college students. It might be assumed that the present findings should apply to any speech situa- tion with a cormarable adaptation of message content to audience interests. Of course, this assumption needs testing before any findings similar to those of the present study can be fairly gener- alized to other kinds of audiences. The failure to support the experimental hypotheses for the unfavorable group of subjects suggests the need for additional re- search. Specifically, within the context of the present stuchr, a comparable audience which would view the experimental message as controversial, while remaining unfavorable toward it, is surely available. An organized campus group of independent students might be located which could serve as an experimental group. The 58 non-salient form of the message would be applicable to any college campus; and, if necessary, it could be altered to remove or balance the persuasive bias. Such a replication of the present study should provide clarification of corollary B of Hypothesis I and Hypothesis. II. A large University pOpulation, with a corresponding low number of Greek letter societies, might provide a sufficient num- ber of neutral subjects to explore the effect of message order on suhh subjects. The literature on primacy-recency and climax-anti- climax message order would have to be first re-explored and a rationale developed to predict a favoring of either initial or final position for a given side of the message presented. The lack of support for Hypothesis III keeps open the question of the effect of order on the retention of message seg- ments within a two-sided presentation of controversial material. An analysis of the present rationale suggests a possible basis for a different approach. The following line of argument was contained in the ration- ale for Hypothesis III: For controversial material, attitudes vary directly with retention; since it is predicted that order affects attitudes, it is predicted that order affects retention. This has, at least, a surface consistency. It was noted that a molar-molecu- lar relationship would have to hold for either prediction to be justified. The variation of attitudes with retention is on a molar level: that is, if you favor one side of an issue, you will retain more of that side when it is presented to you than you will the other side when it is presented; and you will continue to favor the 59 same side. The influence of order on attitudes and retention is on a molecular level: that is, within this favored range of atti- tudes and retention, you may shift around as to attitude strength and extent of retention. In the present study, the molar-molecular relationship seemed to hold for attitudes but not for retention. There appears to be a need for an independent basis of prediction of the effect of message order on retention. It is possible that Hypothesis III should not be advanced. in independent analysis of order effects on retention might reveal no reason to predict such variation in retention. On the other hand, it might call for a related pre- diction under a different kind of experimental design. In summary, the present stucb' undertook to provide a rationale, based on an audience definition of controversial material, to predict the effect of message order on audience attitudes and retention. This was accomplished and a test was made of the result- ing hypotheses. With respect to the variable of attitude, it seems justified. Further research is needed to clarify the operation of the retention variable, but, perhaps, the present stuck has helped to define the direction such research could take. BIBLIOGRAPHY Bartlett, F. C. Remembe : A Study in Experimental and Social m. r 6: Cambridge UniversityfiPress, I932. Cromwell, Harvey. "Relative Effect on Audience Attitude of the ‘ First Versus the Second Argumentative Speech of a Series," Speech Monographs, XVII (June, 1950), 105-122. Ecru-6s, A. L. "Political Frames of Reference as a Factor Influ- encing Recognition," Journal of Abnormal and Social Psycholog, mm ( 191:1). 314-50. Ehrensberger, Ray. I'An Experimental Study of the Relative Effec- tiveness of Certain Forms of Emphasis in Public Speaking," §neech Mariam III (19:45), 9h-111. Gilkinson, Howard, Panleon, S. F., and Sikkink, D. E. I'Contro-- versial Statements in Connected Discourse, Their Form and Pglitécs," Speech Monogaphs, XX (November, 1953), 2 3-2 0. . ''Effects of Order and Authority in an Argumentative “Speech," Qiarterly Journal of Speech, IL (April, 1951;), 1 3-192. Culley, Halbert E. and Berlo, David K. "Effect of Intercellular and Intracellular Speech Structure on Attitude Change and Learning,a Speech Mom, XXIII (August, 1956), 288-297. Hovland, C. I., Inmsdaine, A. A., and Sheffield, F. D. Experiments on Mass Commnication Studies in Social Ps hoIo iii or ar . o . . Princeton: Wagon Uiversity 83, o Jersild, A. T. “Primacy, Recency, Frequency, and Vividness,n Journal of Experimental Psycholog, III (1929), 58-70. Knower, Franklin H. uIi'drzperimental Studies of Changes in Attitude: II. A Stucw of the Effect of Printed Argument on Changes in Attitude," Journal of Abnormal and Social Psycholog, XXX (1935), 522-532. Levine, J. M. and Murphy, G. "The learning and Forgetting of Con- troversial Material," Journal of Abnormal and Social P cholo , mvnr (Em-517. 60 61 Lund, F. H. "The Psychology of Belief: IV. The Law of Primacy in Persuasion,"_Journal of Abnormal and Social Psycholog, xx (1925), 153:191. Osgood, Charles E. , Suci, George J ., and Tannenbaum, Percy. The Measurement of Meaning. Urbana: University of Illinois- Press, 1957. Paulson, Stanley F. "The Effects of Prestige of the Speaker and Acknowledgment of Opposing Arguments on Audience Retention and Shift of Opinion," Speech Monographs, XXI (November, 195h), 267-271. Sherif, M. The Psychologny Social Norms. New York: Harper, 1936. Sponberg, Harold. "The Relative Effectiveness of Climax and Anti— Climax Order in an Argumentative Speech," Speech Monographs, XIII (19h6), 35-hh. Taylor, Wilson. "Application of 'Cloze' and Entropy Measures to the Study of Contextual Constraint in Samples of Contimous Prose." Unpublished Ph. D. dissertation, Dapartment of Mass Commmications, University of Illinois, 1951:. Thistlethwaite, Donald L., Kemenetzky, J. and Schmidt, H. l'Fac'tors Influencing Attitude Change Throu Refutative Commiss- tions,‘ Speech Monographs, XXIII March, 1956), 111-25. Watson, H. S. and Hartmann, G. uThe Rigidity of a Basic Attitu- dinal Frame " Journal. of Abnormal and Social Psycholog, mrv (19395, 3334-335. APPENDIX A Hon-Salient Message: Pro-Con Order Introduction Alpha . . . Omega, two words that bring to mind for most people either religion, science or Greek letter societies. You may have referents for these two words in all. three areas. My concern today is the latter reference. Certainly, you recognize the growing demands placed on the United States in this world of daily crises. Denocracy never could exist without an educated citizenry. Today, its survival necessi- tates, in the words of our President, " . . . even greater sacrifices on the part of those citizens who would attempt continued freedom under law." Your willingness to invest four years of your life in higher education is testimomr of your interest in preparing for to- narrow. Is the structure of University life in its totality adapted to the needed sacrifices? Now, totalitiea are difficult to fathom in am given amount of time. In the short time I have I would like to examine the role played by social fraternities and sororities in Universities. An exmnination of thisifacet of education seems fit- ting diring a time when so new are turning to the Universities for needed guidance . 62 63 Pro Material It seems to me that there is clear evidence in at least three areas pointing to the values of the fraternity system on University campuses. In the first place, while they provide the means for fun, they do so while develOping a sense of responsibility. Consider, for instance, the attention paid to academic responsibility within the context of fraternity or sorority life. Most houses have schol- arship chairmen. If needed, tutors are located. Other members, who have had the same course or are even in the same curriculum, are available to offer adiice as are, of course, the general academic leaders of the chapter. Such obvious direction can instill a sense of responsible fun plus study. Responsibility goes beyond the campus. Almost all frater- nities and sororities on a national level have philanthrOpic en- deavors. They are linked with national pregame for the blind, men- tally ill, crippled, and other socially responsible institutions. Closer to home, there is a close liason with the University com- munity providing members with training and experience in fulfilling an individual's role in his commnity. A large part of any social organization's activities, and campus fraternities and sororities are no exception, is the infor- mal interaction that takes place. Hen-parties and bull-sessions range from sex to politics, and while they provide a chance to ex- press views, they also force a person to develOp a legical defense for the position he takes. The fact that these sessions take place. in an organized setting forces a member to come back the next time with 6h better arguments or else admit he would like to join the "clear- headed," "deep-thinking" opposition. As social fraternities and sororities, they provide useful guidance for the develOpment of an individual socially. For most college students, if they join, this is the first participation in a continuing social group. It won't be the last, or likely, even the most important group. Yet, file here that the social graces from table manners to total social adjustment are explored with lots of room for esror. You have all heard the sad tales of embarassment at this or that social function. Fortunately, there were not then any vice-presidents or wives of vice-presidents who could as easily relate them. Such formalised training in social procedures are highly valued by many alums who learned to laugh and profit from . mistakes made while at college. However, no member of a fraternity or sorority that I have talked to view the social funttion they engage in as simply more edu- cation. There is plenty of justification for members in the activity itself. Social interaction with m of their choosing, having common interests is welcomed distraction from regular class attendance. I noted earlier that there are strong local ties with national associations. This leads to an area of value in and of it- self. Social fraternities and sororities, joined while at college, provide long lasting associations far beyond the four years spent on campus. They help bind the graduate closer to his school by encoura- ging allegiance to the University. With Alumni Associations all over the country it is also possible to continue an active interest in the movement after graduation. 65 Transition I mentioned at the outset of this talk, Alpha and Omega, the beginning and the end. As a justification for not appridsing only one view of the role played by fraternities and sororities in the Univer- sity, I offer a brief reflection on your college experiences as both a beginning and an and. Just as your college training marks a new beginning of critical thinking and curiousity it marks an end to formalized direction to that thinking and curiosity... The blanket buying of one view, either favorable or unfavorable, of social fra- ternities and sororities is not in keeping with the beginning or the end of the period for training in critical thinking. With this in mind, permit me to approach the question from the other side. Con Material No one can examine the fraternity-sorority system without honestly asking what are the problems that they present for the basic task of getting a college education? Legitimate answers might fall into three areas. To begin with, they have failed to adjust to the changing University. This is not a new problem but it is a presently criti- cal one. It became clear with the return of the G. I.'s that the concept of hazing had little meaning for war veterans. That the ha- zing ritual concept died slowly is evidence by adherence by some chapters even today. This seeming inability to adapt to Changing conditions is most clearly evidenced, however, by the pushing of too many members into too many extra-curricular activities. In a time when even education for education's sake must be sacrificed, it seems to me that activities' activity for activities' sake is in- tolerable. 66 Should it be argued that competition for honors other than grades is healthy, then I would reply that true competition seems hardly noticeable over the cheers for Mg for winning's sake. Let us hope that the age we live in need not demand of you full time study. For there surely comes a time when sacrifice in one area be- comes self-defeating. And yet when you add up the man hours that {go into annual Fraternity-Sorority carnivals and shows, it may be possible to say that a healthy balance is not nowastablished. The normal pressures that accompany the desire to accomplish well the present task, when that task is certainly extra-curricular, initiates the habit of low tolerance for study time. It is budgeted into future cram hours and perhaps even then further readjustment may have to be made. This failure to adapt to a changing University points up another major criticism. The Opportunities for development in lead— ership which is not possible within the context of classroom study comes face to face with the closed-corporation, a closed fraternity- sorority corporation in campus activities. The concept of the in- dependent thinker is basic to the American University, and yet, be- cause he is independent in association, the unaffiliated student is pushed into the background. To organize for a specific Opportunity is to require even greater study sacrifices than those demanded of members of an on-going group. The result is virtual ostracism from positions of campus leadership. This problem is not easily solved within the present situation. Now it may be argued that not everyone can hold such posi- tions of campus leadership and the real justification is in providing 67 some individuals with useful training instead of no training for amt- one. If this is so , then there still remains little excuse for a few sure winners to be urged into taking more and more extra jobs be- cause it looks good in the record to have one of "our" group elected or appointed again. Not enough is known, but certainly too much cannot be said of the personal conflict that fraternities and sororities present to a college student when his energies need to be channeled for education. The financial cost balanced against reward is re-ex- amined over and over again. The personal need for acceptance among peers leads to the most serious conflict. The pressure to partici- pate so as to belong is felt with a constant cross-pressure to stuw so as to achieve. These cross-purposes make for long range and long lasting problems. Conclusion I have tried in this talk to examine critically both sides of the question of the role played by social fraternities and soror- ities in the University. It is my hepe that this discussion of one facet of the educational experience each of you are engaged in is in keeping with the contemporary need to critically view the one insti- tution assigned the task of developing educated leaders for an edu- cated democracy. 68 Salient Message: Con-Pro Order Introduction Alpha . . . Omega, two words that bring to mind for most people either religion, science, or Greek letter societies. As college students you may have referents for these two words in all three areas. My concern today is the latter reference. Certainly, as college students, you recognize the growing demands placed on the United States in this world of daily crises. Democracy never could exist without an educated citizenry. Today, its survival necessitates, in the words of our President, " . . . even greater sacrifices on the part of those citizens who would attempt continued freedom under law." Your willingness to invest four years of your life in higher education is testimny of your interest in preparing for tomorrow. Is the structure of University life in its totality adapted to the needed sacrifices? Now, totalities are difficult to fathom in am given amount of time. In the short time I have I would like to examine the role played by social fraternities and sororities at Kansas State. An examination of this facet of education seems fit- ting mring a time when so many are turning to the Universities for needed guidance. Con Material No one can examine the fraternity-sorority system at Kansas State without honestly asking what are the problems that they present for the basic task of getting a college education? Lagitimate ans- wers might fall into three areas. 69 TO begin with they have failed to adjust to the changing University. This is not a new problem but it is a presently criti- cal one. It became clear with the return of the G.I.'s that the con- cept of hazing had little meaning for war veterans. That the hazing ritual. concept died slowly is evidenced by adherence by a least two local chapters even today. This seeming inability to adapt to changing conditions is most clearly evidenced, however, by the push- ing of too many members into too may extra-curricular activities. In a time when even education for education's sake must be sacrificed, it seems to me that activities' activity for activities' sake is in- tolerable. Should it be argued that competition for honors other than grades is healthy, then I would reply that t rue competition seems hardly noticeable over the cheers for winning for winning's sake. Let us hope that the age we live in need not demand of y... run time study. For there surely comes a time when sacrifice in one area be- comes self-defeating. Ind yet, when you add up all the man hours that went into the brief presentation of I-Orpheum it may be possible to say that a healthy balance is not now established at Kansas State University. The normal pressures that accompany the desire to accomplish well the present task, when that task is certainly extra-curricular, initiates the habit of low tolerance for study time. It is budgeted into future cram hours and perhaps even then further readjustment may have to be made. This failure to adppt to a changing University points up another major criticism. The Opportunities for development in leader- 7O ship which is not possible within the context of classroom stuchr comes face to face with the closed corporation, a closed fraternity-sorority corporation in cmpus activities. The concept of the independent thinker is basic to the American University, as it is basic to Kansas State, and yet, because he is independent in association, the unaffiliated student is pushed into the background. TO organize for a specific Opportunity is to require even greater stu