1.. u. _ . > ‘ ‘ . . r . m _ . .. i i. fix. L s :4 3, .. L - 3-..: ‘ .. En . sill... N... : . ....l:t: r . r a. c . u- xivATu.4¢u:>.blumHun.2.. u 11 1.1!. II t t .c . t. g . n . . . .‘ 1 h if .y . . u n”? t - .. p Q? lvlt .! l l l I! , v u \ WI?! 5‘ . .4 . . ~ 5.hu v :1. n f ‘ ‘ n‘ c v o . ‘ ~ u v uh . . Al . t v t . o. u, i I I .1 H fin: . 1.. . . .v “I; \ I . V 0‘9 y . y war“ I ? hafiO J . 1 § 0. c 1 . . k » K _ it -uNfiQ .LI. 1. I. « I . m \It hurt I .. a | ¢ I . . V \ "“h‘ L» I _ . 1.13.1! \- .l. J I ' I A? u 0 at ‘ ' O - t ‘ 7 ’ '1 3.! 1.. qII - 110001? «0 Filippo" H .1 It! . . II . . .(_ 1c . ftfhv‘hfséWkw 1!:1flvuMILfomfir. . .71an 5 5r . Jul .0.I;NF: . .500]! I... cl ‘ 15! .. f . {‘3 .h‘. '. lav .. 1. . ot-t :15.ka .- .IRWH..." . « Irv; . . . a bl} o rt. . II... "I. .uAI‘ .\ u _ tl.‘-0ul 1")!le .3 7: fl} y flying! u n: .11.. ”if." .14.. .07.?” {5' . . P413“!!! bin- ~11qu 9" 1(le «i "an“ ' . $110"... In.- . I unavol .Ic III-IN- v “1’in lil‘lJlHM'llll OH E 01". grtp‘li‘l “I ‘ . .—H..‘I ftdODul' Ln!!!“ . t .. 431.11.1qu . . 73: I .I a]. x . . . «Quinnufnn 14.1.4. .farluxla at»! . {M’Lrlnnfi - H.315; n flit]! \ n. 0v). .our! .v. . «.0 .1 . 3?-.. lgtnflprdl .ufiuuhlfis In... J ‘glwuffi. - :Kwnwfli.-. . . ( L. 1’3. L I..- z. . r}.-- - .. $1. J by...“ .. Jul [Illa .I'o..l I Gilli ”MIMIIIIIIIIIIIII L , .r r ,., 3 1293 00998 7342 a: , . .. a m n a, I" a E. “q 5" J i‘ U ll cl. I D .- 9 . -. g .1 E ' V 2‘: '3 ’ : . if I1‘15ng r ' This is to certify that the dissertation entitled THE EFFECTS OF COGNITIVE ELABORATION, VALID ARGUF-IENTS, EVIDENCE, AND MESSAGE STYLE IN PERSUASIVE DISCOURSE presented by RODNEY ALLEN REYNOLDS has been accepted towards fulfillment of the requirements for Ph . D. degreein COMMUNICATION : Major professor :; Date 6/ 15/ % MS U is an Affirmative Action/Equal Opportunity Institution 0-12771 MSU LIBRARIES a! RETURNING MATERIALS: Place in book drop to remove this checkout from yoyr record. FINES wiII be charged if book is returned after the date stamped below. I??? $51993 \ Ix 1‘ 1 _ ‘mufl ‘"“ 0=A it ; u';fi*;l’ffl' '— Q :~- r n A: . .3.“ THE EFFECI‘S (F C(IENITIVE ELABORATION, VALID AKIJMENI‘S, EVIDM: AND MESSAGE SI'YLE IN PERSUASIVE DISCGJISE BY Rodney Allen Reynolds A DISSEIH‘ATIQJ Submitted to Michigan State University in partial fulfilment of the requirements for the degree of moawmnmm Department of Coummication 1986 ABSTRACI‘ 'IHE moms (F m ELABORATICN, VALID mm, NEW, AND MESPGE STYLE IN PERSUASIVE DISCGJISE By Rodney Allen Reynolds 'Ihe perspectives on reasoning, evidence, and message style are reviewed and held to be inadequate in providing sufficient grounds for predictions about message receivers' responses to persuasive messages which vary on argument validity. evidence strength, and message style. Based upon the thinking and the enpirical research from the perspectives of reasoning, evidence. and message style in social influence a model of belief elaboration and attitudes in persuasion is advanced and tested. The full model was not supported. Several secondary analyses, however, offered insight into the relationships proposed in the model. line most significant finding from the study was that the effect of evidence evaluation in persuasion was mediated by the receivers' evaluation of the message. There was some indication that receivers' Widence evaluations are a function of habitual cognitive efforts as well as actual cognitive elaboration on the nessage. Limited support was Provided for the conclusion that the validity of the argtmmts in a message are detectable by receivers. The results of this study are uninformative about the role of message style in persuasion. mm The opportunity to express appreciation for the contributions to my life and education in a formal and lasting document is rare. These contributions have been many and varied: I am grateful for them all. Those who trade particular contributions towards the coupletion of this document certainly deserve special recognition. My advisor, Dr. Michael Burgoon, has expected nuch from me and I have violated those expectations in both negative and positive ways. His patience, persistence and encouragements, however, caused me to conclude that finishing the doctoral degree was well within his expectations. As a scholar and a friend, Dr. Michael Burgoon continues to inspire me. The numbers of my doctoral cannittee offered me the very best in teaching and research. Dr. G. R. Miller micalpromisingly looks to the best in colleagues and graduate students. I am proud to have had the opportunity to work with a scholar, like Dr. Miller, who has lead the way for those of us interested in connunication science. Dr. Judee Burgoon has offered me detailed feedback on any writing that I turned in to her. 'Ihat feedaack has caused me to have nuch higher expectations for my writing. Moreover, her dedication to research and teaching is an exalplar of the distinguished scholar. Dr. Charles Atkin led me to a mach deeper appreciation for mass coununication scholarship. His genuine respect for graduate students, kindness and gentle honesty are an inmadiate attraction for all who cane to know him. In addition, Dr. Atkin's warm bane has provided me with many happy memories of East Lansing. Dr. Joseph Hanna taught me new ways to look beyond the traditions and assurptions in ii my reading, thinking and research. In particular, it was in Dr. Hanna's classes that I more clearly developed ideals for the worth or worthless- ness of ideas. It has certainly been an honor to work with these five outstanding scholars. Dr. Mark Hamilton and Dr. Jack Hunter were very kind in allowing me to employ their path analysis personal conputer software. Both Mark and Jack were also very giving of their time in helping with difficulties I had with both the software and the calputer hardware. Dorothy Tarvin and Tom Ridley encouraged me to' go on to college. Tom also encouraged the conpletion of each of my degrees. Southern Calif. Gas Co., Hunt Wesson Foods, and the California Department of Agricultural Inspection provided the incentive and the means to go on to college. Dr. John Reinard introduced me to the scientific method and the value of testing ideas via alpirical observations and then sharing the results with others. John provided the foundation for graduate school and an academic career. My wife, Lynn, made the difference in my finishing the Master's degree and the decisions that brought me to Michigan State University. Lynn and our children, Brett, Bryan and Mary Beth, have given the absolute best in loving support and demanding closure at appropriate times. 'lhey have made the role of husband and father more than merely rewarding. More than any single contribution to my life and education, these pages are filled with print because of Dr. Cal Hylton and Dr. Jack Ray. Jack defines the concept of a scholarly colleague. He took an imnediate interest in my work on the role of reasoning in persuasion, read the iii orignial sources that I employ and continues to think through ideas with me. If Jack defines the concept of a scholarly colleague, Cal is the ideal type. Cal offered several suggestions that inproved this project and proofed the entire text of the manuscript. In addition to his practical and efficient scholarly, career and personal assistance, I know that I can count on Cal for a friendly “Hey, you,“ a warm cup of coffee, a ready wit and the best of friendship. iv Table of contents List of figures List of equations List of tables Introduction context ofthe problem Statement of thejproblenu Perspectives on human reasoning in social influence Probabiological equations Theoretic ideas Adequacy of the equations Elaboration Attitudes and beliefs Application to persuasion The theory of reasoned action Theoretic ideas Catparison to probabilogical structure Discrepancy and facilitating factors Attitudes and beliefs Elaboration Application to persuasion Elaboration likelihood model Conceptual ideas Elaboration Peripheral route Application to persuasion Summary of the perspectives on human reasoning in social influence A.model of belief elaboration and attitudes in persuasion Elaboration and logical arguments Evidence and elaboration Message style, evidence and elaboration Test of the model Sample Message construction Selection .Argument validity Evidence Message style Measurement Predisposition to elaborate Elaboration: cognitive effort Probability of belief in a proposition Detection of reasoning errors Evidence evaluation Message evaluation Attitude toward thesis Procedures vii viii "‘ I? pgmmmmhhww Table of contents (cont.) Results Measurement model checks Test of the model SUpplementary analysis of the model Evidence route isolated ‘Discussion Modifications and suggestions for future research References Appendix.Az Justification for the use of traditional logic in the study of social influence Appendix B: Experimental.nessages Appendix Ctheasurement instruments Appendix D: Supplementary tables vi 96 101 118 128 Figure 1: Figure 2: Figure 3: Figure 4: Figure 5: List of figures The elaboration likelihood model of attitude change A model of belief elaboration and attitudes An extended model of belief elaboration and attitudes Results of the test of the proposed model Results of the test of the evidence route vii 24 37 48 71 74 taxation 1: Equation 2: Equation 3: Equation 4: kinetic!) 5: List of equations McGuire's probabilogical equation wyer and Goldberg's probabilogical equation Fishbein and Ajzen's reasoned action equation Fishbein and Ajzen's descrepancy and facilitating factors equation Detection of reasoning error equation viii 13 15 62 List of tables Table 1: Carparison of stylistic indices for the research messages Table 2: Carparisons of scores for the source's belief in the propositions Table 3: Calparisons of reasoning errors for each argtxment and the argmnents canbined Table 4: Factor analyses of the belief and attitude scales Table 5: Correlation matrix ix S7 129 131 132 133 Elaboration, arguments, evidence and style 1 The Effects of Cognitive Elaboration, Valid Arguments, Evidence, and Message Style in Persuasive Discourse Introduction Context of the problen In 1967 Wayne Thomson reviewed the experimental studies on renaming and evidence in discourse and concluded that the “experimental findings . . . provide almost no evidence against using sound evidence and argments but not nuch support for euploying then“ (p. 53) . Given the role that reasoning and evidence has played in both rhetoric and canmnicatim theory (see Arnold & Bowers, 1984), Thonpson's conclusion deserves greater theoretic and enpirical attention. Reynolds and Burgoon (1983) provided a recent review of the advances in research on reasoning and evideice. Research conducted in the sixteen years between the Thonpson review and the Reynolds and Burgoon review provides a substantial amount of support for the use of evidence in persuasive messages. Reynolds and Burgoon note, however, that the findings on the use of evidence lack theoretic bases arnd, therefore, offer little in the way of substantial explanations or predictions about the effects of specific types of evidence across or between situations, sources, or receivers. Reynolds arnd Burgoon (1983) also note that the research on reaming is more theoretically sophisticated than the research on evidence but, ironically, offers conparatively little in the way of elpirical findings Elaboration, argtmnents, evidence and style 2 on the use of sound arguments in persuasive discourse. For the most part, reasoning researchers have focused upon the ability to reach valid conclusions for syllogisms or other logical tasks which are typically presented in an objective test format. In short, much of the research on reasoning has been conducted outside of the context of cannmication. In addition to the dearth of enpirical research, extant studies have not adequately dealt with the effects of message style which may add to or interact with the effects of evidence and reasoning in persuasive nnessages. Clearly, stylistic elenents in messages are inportant to the persuasion process. The rhetorical literature has stressed both reasaning and style. For exanple, Kauffman (1981) has argued that “since Aristotle, argument and poetic have been conceived of as interdependent types of discourse" (p. 407). Fran an empirical perspective Burgoon and Miller (1985) have pointed out that "even relatively minor variations in the linguistic and syntatic properties of a message can influence persuasive success.‘ Further, Burgoon (1983) has argued that not accounting for stylistic variables in the study of persuasion can misguide both the conception and interpretation of erpirical studies. The inportance of addressing questions about the inpact of reasoning, evidence and message style in persuasion should not be underestimated. In some communication courses students regularly encounter exhortations to inprove their knowledge of the criteria of logical adequacy because such criteria "constitute standards which [conlmnicators] might well wish to- apply when functioning as discriminating receivers of persuasive communications" (Simona, 1976, p. 192). These same students, however may Elaboration, arguments, evidence and style 3 be informed that the study of logic is “as free from all practical concerns as in sane branch of pure nnathenatics" (Toulmin, 1958, p. 2). In sane courses students may hear pleas on the use of tests of evidence (e.g., McCroskey & Wheeless, 1976, p. 370-372) while also hearing that receivers are unable to apply these sane tests (Harte, 1971) . Similarly, cannunication students may be taught that a clear, concise and expressive linguistic style is inportant (e.g., Wilson 5 Arnold, 1978, p. 254) while also learning that vagueness, anbiguity and ornateness are suasory (Wilson, 1971: Williams & 6088, 1975: Williams, 1980: Brunmett, 1981: Eiseberg, 1984). Sane nay argue that the issues of what we teach and what we know to be effective merely split on what is versus what ought to be persuasive, however, following Henpel (1965) , "to qualify a given action as rational is to put forward an enpirical hypothesis _a_ng a critical appraisal" (p. 463). The sane argument may be advanced for evidential support and message style. In short, before the traditional canons of nnessage construction and reception are retired to a hanne for acadenic pedanticisnns, they deserve greater theoretic and empirical exploration. Fortunately, research on questions about the role of reasoning, evidence and message style in social influence have recently received renewed interest (e.g., Hanple, 1977: 1978; 1979b: Sarndell, 1979: Wyer & Hartwick, 1980: McGuire, 1981a; Fishbein & Ajzen, 1981: Jackson, 1982: Reynolds & Burgoon, 1983: Ray 8 Findley, 1984). These research efforts, however, have not been integrated. Elaboration, arguments, evidence and style 4 The renewed interest in reasaning, evidence and style parallels a resurgent interest in cognitive processes anong the social sciences (see Greene, 1984) annd, in particular, persuasion (see Petty, Ostronn, 8 Brock, 1981) . For exanple, sane "cognitive response" researchers (e.g., Cacioppo & Petty, 1980b) have sought to extend the role of counterarguing in persuasion (for a review see Petty and Brock, 1981) by exanining a broad range of thoughts generated by receivers in response to persuasive messages. Unfortunately, "there exists no 'definitive' theory of cognitive responnses. Inndeed, the cognitive response approach is more a conceptual orientation (or perhaps a statement of faith) toward the role of thought in attitude change and persuasion" (Ostran, 1981, p. 287). Similar theoretic problems emerge with other perspectives on reasoning, evidence annd style in social influence. S_t_:_a1:_g__r_nent of the Problem Given the state of theory and research on reasoning, evidence and style, there is a need for the advancement and testing of a theoretically based model of the effects of persuasive messages which vary on the use of valid arguments, evidence, and message style. Further, such a model mist address how cognitive elaboration (thinking) influences the effects of reasoning, evidence and style on receivers' responses to messages. Wives on human reasoning in social influence There are three major theories or models which address the issue of human reasoning in social influence situations. First, Moanire (1960a: 1981a), Wyer (Wyer, 1970: 1972: 1973; 1974: 1975: Wyer & Carlston, 1979: Wyer & Goldberg, 1970: Wyer & Hartwick, 1980) annd Hanple (1977: 1978: Elaboration, arguments, evidence ad style 5 1979b) have exanined probabilogical models of belief change which McGuire and Wyer claim to represent both subjective probability and syllogistic reasoning processes. Second, Fistbein and Ajzen (1975: 1981: Ajzen & Fishbein, 1980) have advanced a "theory of reasoned action" which contains a presupposition that humans are rational. Third, Petty and Cacioppo (1981) offer an "elaboration likelihood model" of persuasion which is based on the assnmption that hnmnan reasoning is idiosyncratic but that the role of subjective assessments of reasoning is central to persuasion. These perspectives, in general, are conncerned more with the structure of cognitive processes than with the effects of persuasive messages. Lobabilgical eggtions Theoretic ideas. The initial probabilogical equation of reasoning advanced by Wire (1960a) is presented in Equation 1. McGuire claims that this equatian captures both the process of probability assessments of the truth of propositions and the process of syllogistic reasoning. For manire, rationnality amounts to a normative conformity to the standards of traditional logic and the principles of probability. pC = pApB + pk + pD Equation 1 The pC term represents the individual's belief in the conclusion. The pApB term in the equation represents the individual's belief in the two premises of a syllogistic argument. The pK term represents all premises that the individual may consider in addition to pApB in assessing pC. The pD term represents the individual's desirability for pC. Muire has referred to this desirability term as "wishful thinkirng" (mire, 1960a) or as "hedonic consistency" (mire, 1981a) . Elaboration, arguments, evidence and style 6 McGuire's probabilogical equation is an extension of the consistency theories (see Insko, 1967) . Thus the equation is part of a more general theoretic structure in which it is assumed that: 1) People seek to maintain rationnal consistency. 2) Rational inconsistency is aversively motivating. 3) People will behave in ways calculated to restore rational consistency (cf. Miller, Burgoon, & Burgoon, 1984, p. 431). McQuire (1960a) argues that other consistency theories suffer fran either the denial of cognitive processes (by exanininng only behavioral consistency) or by such broad notions of consistency that the possibility of inconsistency is practically denied (see also Fishbein and Ajzen, 1975, pp. 6-8 for a similar discussion) . Wyer and Goldberg's (1970) modification of McGuire's originnal equation is presented in Equation 2: P3 = PA PBIA + 98‘ PBIA‘ Equation 2 The pH term on the left side of the equation represents the conclusion of the two conditional arguments (for consistent references, however, hereafter the conclusion of any argument will be referred to as pC) . The first term in the equation (pA pBlA) represents a conditional argument of the form: A: If A, then B. The second term (pit or pA‘ pBlA‘)in both equations represents another conditional argument of the form: Not A: If not A, then B. The two equations are basically the sane. The first term in both equationns differ only on the assunnption of pi and p8 being independent (mp8) or contingent (pA pBlA). Hanple (1979b) points out that the Elaboration, arguments, evidence and style 7 distinction is not logically or enpirically inportant. The second term (pit or pA‘ pBlA‘) in both equations represents the additional premises other than pA and p8 an individual applies in the assessnent of pC. McGuire (1960a) , however, dropped the pK canponent on the grands that it is probably randomly distributed across any population and inpossible to neasure. Wyer's (see Wyer & Hartwick, 1980) and Hanple's (1977: 1979b) efforts to clarify and enpirically assess the pK term ted to support Mnfluire's (1960a) inplicit conclusion that pK is more error than systenatic variance. Despite wyer's Myer & Hartwick, 1980) assertion that the insignificance of pit may be an artifact of the subject's lack of prior knowledge on the experimental message topics, there are grounds for concern that the pK term might subtract fran our understanding about human reasoning and social influence. Specifically, the inclusion of the pK term may foster the connstruction of ex 2951;, f_a_c_t_o accounts which would deny any explanatory power to the equations. For exanple, Hanple (1977) advances Henle's (1962) argument that people ted to apply additional prenises to connvert conditionals into biconditionals to explain the potential effect of pK. (e.g., to decode "If A, B" as "If B, A" on the bases of knowing K) . Such an assnmed conversion is then used to claim that apparent errors in reasoning are quite logical. However, as Evans (1980) has pointed out, the connversion hypothesis may not be falsifiable since there will always be another possible premise conversion. To paraphrase Henple (1965, p. 471) , such accounts of rationality afford ground for believing that it would have been rational for a person to Elaboration, arguments, evidence and style 8 apply an additionnal prenise, but no grounds for believing that the person did in fact apply such a prennise. Further, such accounts deny the possibility of rational inconnsistency which is ironic in light of McGuire's (1960a) intention to make the notions of consistency and inconsistency more extensional. flag of the Quations. Disregarding the problens with the pK term, the evidence is inpressive that the first term in the equations maps the reasoning of respondents very well. Reviews of the probabilogical research by Wire (1960a: 1981a), Wyer (Wyer, Carlston, & Hartwick, 1979: We & Hartwick, 1980) , Hanple (1977: 1979b) and Fishbein and Ajzen (1975, pp. 169-181) have all reported correlations between predicted and obtained scores for the probability of belief in the conclusion which have ranged between .40 to .96. Cohen and Cohen (1975, p. 56) claim that such correlations are carparatively "large" within the social sciences. In short, this evidence supports the view that people do ted to maintain rational connsistency. _E_1__ab_og_a_t__i_9n_. A major use of the equations has been in the testing of the Socratic effect (Mwuire, 1960a) . That is, the hypothesis that merely asking people to think about their beliefs on a topic will cause these people to examine their cognitive organization and restructure that organization so that rational consistency is increased. There exists substantial enpirical support for the Socratic effect (McGuire, 1981a) . Wyer (Wyer & Carlston, 1979: Wyer & Hartwick, 1980) has argued that the Socratic effect involves not only a nere dwelling uponn the cognitive structure but also an active elaboration about new connections between Elaboration, arguments, evidence and style 9 beliefs. Therefore, in addition to denonstrating that people ted to maintain belief structures which approximate rational connsistency, this research has dennonstrated that cognitive elaboration about one's beliefs is related to the degree of rational consistency. Attitudes and beliefs. One area of difficulty with the link between the probabilogical approach and nnost persuasion research is Wyer's (1974) and McGuire's (1981a) claim that distinctions between attitudes and beliefs are conceptually uninportant. They argue that both beliefs and attitudes represent a person's judgnent of an object's nenbership in a cognitive category. "There is no a mpg; reason to assnme that the factors that affect judgnnents of category menbership depend upon the nature of the categories involved" (Wyer, 1974, p. 25). Hanple (1978) has demonstrated that people are able to reason with affect statenents in the sane way that they are able to reasonn with belief statenents. But, it scene doubtful that belief statements are processed or have the sane inpact that affect statenents have. While nuch could and should be written in criticisnn of the reduction of affect to assessment of truth (i.e., It is true that heroin is bad), it is sufficient for now to point out that such a nnove is inconsistent with McGuire's and Wyer's discussions of the inportance of "wishfull thinking." If we are to take McGuire and Wyer at their words on the similarity of belief and attitude judgments, then we nust connclude that the pD term becanes just another premise or set of premises for pC. Thus there would be no distinction between pit and p0. Such an equivalence, however, is not what they have advocated. Elaboration, arguments, evidence and style 10 Given that pD is included in the discussionns of the probabilogical agaroach and is measured in a manner similar to most attitnde neasures (by asking the respondent to assess how desirable a proposition is), we may conclude that attitudes do play an inportant role in the assessment of argument conclusions. mire (1960a, pp. 76-79: 1960b) explored the pD canponent of his equation and found that the correlation between pC and pApB increased when the effect of pD was partialled out. (In 1960a the increase was fran .48 to .85. In 1960b the increase was from .74 to .96.) This finding is consistent with a host of research which has indicated that attitudes affect people's assessnents of the logical validity of arguments (e.g., Janis & Frick, 1943: Morgan & Morton, 1944: Henle, 1955: Feather, 1964: Bettinghaus, Miller, & Steinfatt, 1970) . Amlication to ggsuasgln. There is a nnajor problen with the application of the probabilogical equations to questions about the use of valid arguments in persuasive discourse. The equations have not been developed sufficiently to offer insight into how people deal with illogical arguments. As McGuire (1960a, p. 101) states: "The tedency to maintain logical consistency among one's cognitions is neither nonexistent nor absolute." Nevertheless, most of the probabilogical research has been limited to exaninations of rationnal consistency rather than the causes and effects of rational inconsistency. Both Mdluire (1960a: 1981a) and Wyer (1977: Wyer, Carlston, & Hartwick, 1979) have discussed the obvious need to inncorporate nonlogical thinking postulates into their equations. The pD term, for exanple, is offered by Mix-Mire (1960a) to represent the nonnlogical thinking brought Elaboration, arguments, evidence and style 11 about by wishful thinking. Unfortunately, the remaining nonlogical thinking postulates that have been proposed are not likely to be incorporated into the equations because they are so broad as to deny explanatory potential. For exanple, Mdsuire (1981b) enploys his other nonlogical thinking postulates not to add to the predictive power of his ideas but rather to argue that "we have to resign ourselves to entering . . . an age of diminished expectations" (p. 51) for the effects of persuasive nessages. Despite Wyer and McGuire's recognition of the need to deal with nonlogical thinking there has been little enpirical exploration of the issue. The lack of extensions of the probabilogical equations to deal with an individual's own illogical reasoning brings into question the utility of the equationns for exanining the effects of poor reasoning in persuasive messages. In fact, the research conducted with the equationns has not systeIatically exanined the effects of a source's reasoning at all. In most of the studies the research participants provided assessnnents of the probability of belief in the truth of the propositionns. These assessnents were then entered into the equation to test the fit between the equation and cognitive organization. The premises neasured have typically conprised parts of deductively valid arguments only. In those studies where persuasive messages were encountered, the messages were designed to alter only one belief in the syllogistic structure to see if such changes brought about the cognitive reorganization predicted by the equation. Thus the first term of the probabilogical equations only naps logical cognitive processes rather than responses to arguments in nessages. Elaboration, arguments, evidence and style 12 If the equation, and the theory on which it is based is to have much further utility in the study of persuasion, attenpts are needed to account for how people identify, process, and respond to the reasoning of others. While the research participants in these studies to date have denonstrated a tedency towards rational consistency, it is hardly reasonable to assume that all of the messages that they receive are rationally consistent. To nnention one possibility: it would follow from the probabilogical equations that when receivers assess a message and note that the source's pApB term differs substantially fran the pC term in an argument, the receivers should view the argument as rationally inconsistent. Such extension of the probabilogical equations, unfortunately, has not yet been offered. Another major limitation of the probabilogical approach is that there is no indication of how message elements (i.e., style) affect persuasion. Based on the probabilogical literature, it is only intuitively plausible that variables such as message style‘would affect pDu Without a:more explicit account of nessage content, the probabilogical equations are necessarily incanplete models of the persuasion process. Despite the problems with the probabilogical approach to belief clnange, this conceptual and enpirical research garners support for the notions that: (1) People ted to maintain rational consistency. (2) Cognitive elaboration is a key determinate of rational consistency. (3) Affective responses play an important role in the degree of rational consistency. The nnajor limitation to the probabilogical approach is that it has yet to be extended to analyses of the way people respond to the reasoning, evidence use, or stylistic tactics of others. Elaboration, arguments, evidence and style 13 The theory of reasoned action T_h___eo_retic ideas. Fishbein and Ajzen's (1975: 1981: Ajzen & Fishbein, 1980) theory of reasoned action is typically sunnmarized as a single regression equation: B~I=w1AB+wZSN Equation3 The theory is based upon the notion that behaviors (B) are determined by one's behavioral intentions (I) which are in turn determined by the attitude toward the behavior (AB) or the individual's subjective norm (SN). The attitude toward behavior is determined by the sum of the salient belief attributions (b) about the behavior nultiplied by the evaluationns (e) of these attributes (AB = X biei). The subjective norm consists of nornnative beliefs (nb) that significant others think that a behavior should or should not be performed and the motivation to conply (ms) with the significant others (SN -- )3 nbi + moi). @1501} to probabilggical st_rp_c,t_u_r_e. The theory of reasoned action is canplenentary to the probabilogical approach to belief change. In fact, Fishbein and Ajzen (1975: p. 396, p. 486: 1981, p. 349: Ajzen & Fishbein, 1980, p. 239) make explicit links to the probabilogical approach. These links, however, are limited to the claim that to the extent that beliefs can be derived through syllogistic reasoning the probabilogical equationns are informative on the question of belief change. Perhaps because it is too obvious, Fishbein and Ajzen do not enphasizethatboththeABandtheSNcanponentsofthetheoryare syllogistic in structure. The AB canponent breaks down into reasoning such as: Elaboration, arguments, evidence and style 14 pA or b: People who buy Miller beer, work hard. pB or e: flle who work hard are good. pC or A3,: Therefore, people who buy Miller beer are good. The SN canponent breaks down into reasoning such as: pA or MS: My nether wants ne to take birth control pills. pB or MC: I Wivated to do what my nether wants. pC or I: Therefore, I will take birth control pills. Thus, the structure of the first term of the probabilogical equations is preserved in the theory of reasoned action. It is necessary to point out that the SN canponent of the theory of reasoned action is apparently not a potent determinant of intentions (Petty, Ostran, & Brock, 1981). Specifically, an exanination of a nunber of studies reported by Ajzen and Fishbein (1980) reveals that SN and A3 are highly correlated and that SN has consistently added very little to the variannce accounted for in intentions by the AB conponent. anserving that the structure of Fishbein and Ajzen's (1975) fornulation adnnits syllogistic inferences in the formation of beliefs and is, in itself, probabilogical would seen to indicate that the fornulation is based on Fistbein and Ajzen's contention that humans are rationnal. Fishbein and Ajzen, however, do not offer such an account. Instead, the fornulations for both the AB and SN conponents of the theory are based on Fishbein's (1967) extension of Tolman's (1932) and Edward's (1954) expectancy value theories (see Fisl'bein and Ajzen, 1975, pp. 30-32) . Thus, the criterion for rationality under the theory of reasoned action appears to be the maximization of subjective expected utility rather than Elaboration, arguments, evidennce and style 15 valid reasoning structures. Henpel (1965, pp. 466-469) exanined the explanatory potential of the expected utility criterion of rationality and concluded that such accounts are inadequate on the grounds that logically incanpatible decisions can be held to be rational even if the potential utilities are known. Even so, Fishbein and Ajzen (1975) do not offer criteria for knowing what decision rules for maximizing subjective expected utility would counnt as rational choices. Moreover, the nest recent accounts of the theory (Ajzen & Fistbein 1980: Fishbein 8 Ajzen, 1981) make no nenntion of the notion of subjective expected utility but continue to state that the theory is based on the presupposition that people are rational. The role and character of reasoninng within the theory of reasoned action is, therefore, not exactly clear. Discrepancy and facilitath factor . Fishbein and Ajzen's (1975) discussion of discrepancy and facilitating factors in persuasion canpounds the difficulty of identifying the role of reasoning within the theory of reasoned action but may aid in isolating when the reasoning presented in a nessage is nest likely to influence a person. Fishbein and Ajzen claim that the acceptance of a belief (pA) advocated in a nessage is influenced by the discrepancy (D) between the receiver's own belief and the perceived belief of the connnunicator and facilitatinng factors (f) in the cannunication event (pp. 462-509) . The relationship between the variables is expressed in the following equation offered by Fislnbein and Ajzen (1975, p. 464): 9A = (1 - nn)1/f Equation 4 Elaboration, arguments, evidence and style 16 Fisl'bein and Ajzen (1975) are very critical of discrepancy research on the grounds that traditional (social judgment theory) attitudinal discrepancy confuses belief discrepancy with facilitating factors (pp. 486-487). In other words, the connflicting reports (p. 469) on the relationship between attitudinal discrepancy and attitude change resulted fran the conflicting forces of belief discrepancy and facilitation: As belief discrepancy increases, the acceptance of an advocate's claims decrease: as facilitation increases, the acceptance of an advocate's clains increases. Thus, the differing gradients reported in the attitude discrepancy research are not the result of pressure to assimilate (Sherif, Sherif, & Nebergall, 1965) or changes in the size of the latitudes of acceptance or rejection (Hinnelfarb & Eagly, 1974) but, rather, between differences in uncontrolled facilitating factors or unclear specification of the target dependent variable (belief change). Fishbein and Ajzen (1975) claim that the effect of a facilitating factor is an increase (or decrease) in "the person's confidence in his own belief" (p. 462) . The facilitating factors are: source variables such as "credibility . . . attractiveness, sincerity, status, etc." (p. 462): receiver variables such as "persuasibiltiy, chronic anxiety, self esteem, sex, intelligence, etc. , as well as situation and topic-related factors, such as acute anxiety, involvenent, extremity of own position, uncertainty, and the receiver's information about or knowledge of the topic" (p. 463): and nessage variables such as "order of presentation, validity of supportive arguments, and enetional versus rational ameals" (p. 463) . Unfortunately this list of facilitating factors is incanplete Elaboration, arguments, evidence and style 17 and does not reflect Fislbein and Ajzen‘s (1975, p. 463) claim that many of these factors interact with one another in the formation of overall facilitation. Apparently, the facilitatinng factors exponent decreases the parsineny of the theory of reasoned action exponentially. It is unclear what process the facilitating factors represent. They nay be incentives which, within a nessage learning approach, reinforce (or inhibit) the adoption of the advocated arguments for a nessage conclusion. Or, the facilitating factors may represent wishful thinking (pD) which, fran a probabilogical approach, represents nonlogical thinking. Or, these facilitating factors may nerely be subpropositions of belief which lead to the adoption of premises which are, in turn, subpropositions of belief for conclusions. Fishbein and Ajzen suggest all three of these interpretations at different points in their discussion of facilitating factors. Fishbein and Ajzen (1975) are also unclear on what process is elicited by belief discrepancy. The inplicit assunption, however, appears to be that belief discrepancy provokes a defensive posture in which the nessage recipient defeds his or her own belief. Presumably, the defensive posture gives way to persuasion when sane facilitation factor(s) brings about a recognition by the persuadee that his of her own belief is indefensible or causes the persuadee to susped judgment on or otherwise ignore the discrepancy and adopt the advocated position. In short, it would appear that discrepancy increases the degree of cognitive elaboration. an the other hand, it appears that facilitating factors can Elaboration, arguments, evidence and style 18 be split into two categories: Those that aid the persuadee in resolving the discrepancy (presumably via learning or reasoning) and those that inhibit the degree of cognitive elaboration and evince conpliance. Attitudes and beliefs. Little doubt is left after reading Fishbein and Ajzen's (1975) first chapter that their theory is ained at preserving and refininng the distinctions between beliefs, attitudes, intentions, and behaviors that have nuddled and confused other efforts (e.g., Rosenberg & Hovland, 1960) at studying mediating responses to messages. In specific, Fishbein and Ajzen oppose the view that attitudes consist of three canponents (affective, cognitive, and conative) on the gronnds that such a nulticonponent view masks the relationships between these variables and serves only to advance the idea that the attitude-behavior discrepancy problen results fran the conplexity of the attitude construct (Fishbein & Ajzen, 1975, p. 340). Inplicitly, Fistbein and Ajzen have cogently acknowledged that clainne of conplexity nerely beg the question rather than adding to explanation and understanding. (For other criticisms of the nulticanponential view of attitudes see Miller, Burgoon, and Burgoon, 1984, and Zajonc, 1980). Fran a probabilogical perspective, We and Goldberg (1970) used the nuddle and confusion about the nulticonponential view of the attitude construct to argue that the distinction between attitudes ad beliefs is inconsequential. In contrast, Fislnbein and Ajzen have argued that beliefs, attitudes, and intentions are unique variables with specific relationships to each other and to behavior rather than mere conponents of a snmmative variable (see Dubin, 1978, p. 66-68). Elaboration, argmnents, evidece and style 19 It is these specific relationships that are inportant. Fishbein and Ajzen (1975, pp. 457-461) argue against the nessage learning hypothesis (Janis 8 Hovland, 1959: Mcfluire, 1968) that persuasion proceeds fron cognition to affect to conation. (McGuire, 1968, proposed that the nessage learninng process is a five step process involving attention, conprehension, yielding, retention, and action. In 1981b, McGuire expanded this five step sequece to twelve steps). Fistbein and Ajzen (1975) argue that cognition does not lead to affect but that cognition and affect are conbined together in the formation of specific attitudes towards action. Zajonc (1980) and others (e.g., Wheeless, 1971: Love & Greenwald, 1978: Petty a Cacioppo, 1981) have noted that changes in cognition do not necessarily lead to changes in affect but that changes in affect tend to lead to changes in cognition. A close exannination of the AB canponent of equation 3 reveals nere of the difficulty with the link Fishbein and Ajzen draw between attitudes and beliefs. Their fornulation states that an overall evaluative response towards a behavior is the sum of the result of specific attributions about the behavior and separate evaluation judgments of each of these attributes. Thus, the placennent of an object (behavior) along an evaluative dinension is determined by the placenent of each of the attributes of the object along that sane dinnension weighted by the person's strength of belief assessments that the object has each of those attributes. The resulting dimension, therefore, represents a collapsing or nerging of all cognitive responses to one attitudinal dimension. This, while Wyer collapses all attitudes to belief judgments, Fishbein and Ajzen Elaboration, argnnnents, evidence and style 20 collapse all belief judgments onto a single evaluative dinension. Indeed, Fishbein and Ajzen do not typically corpute AB fron measures of beliefs and attribute evaluations but sinply measure AB with a set of traditionnal attitude neasures (see Ajzen & Fishbein, 1980, appedix A and B) . A resolution of these conpeting approaches to dealing with beliefs and attitudes seens to have been offered by McGuire's (1960a) separation of rational consistency and wishful thinking (pD) . That is, rather than atterpting to reduce affect to just another form of belief or attempting to sninsume all beliefs under affective expressions, McGuire (1960a) keeps the two variables as separate determinants of inference conclusions. meg. Fishbein and Ajzen equivocate on the inportance of cognitive elaboration in human information processing. They frequently write that "the theory is based on the assnmption that human beings are usually quite rationnal and make systenatic use of the infomation available to than" (Ajzen & Fistbein, 1980, p. 5) and that they "do n_o_t_ subscribe to the view that hunan social behavior is controlled by unconscious netives or overpowering desires, nor . . . that it can be characterized as capricions or thonghtless" (Ajzen & Fishbein, 1980, p. 5) . But, in the sane publications, they write that they "do not nean to inply that prior to performing each and every action, people systematically scrutinize the determinants of their behavior. Rather, [they] view the processes involved as largely autonnatic or implicit and only in rare cases do we becone fully aware of these processes" (Ajzen & Fistbein, 1980, p. 245) . Such conflicting statenents nay be the result of Elaboration, arguments, evidence and style 21 theoretic inconpleteness with regard to reasoning and elaboration. Nevertheless, these conflicting statements and the discussion of discrepancy and facilitating factors indicates that under the theory of reasoned action, it may be presumed that the degree of elaboration engaged in varies across persuadees and persuasive messages. mugging; to gersuasion. Fishbein and Ajzen (1981, pp. 347-349) explicitly argue that effective persuasive messages trust contain rational arguments and evidence in support of a specific "primary belief underlying the target variable" (p. 348). (For exanple, see their discussion of McCroskey's, 1969, review of the evidence research.) In contrast, Wyer (e.g., Wer & Hartwick, 1980) and McGuire (1981a) do not deal with message construction beyond the indication that a message which changes the belief in a premise will affect beliefs in a conclusion that receivers derive from that premise. Fishbein and Ajzen (1981) are convinced that "one of the problems in research on comunication and persuasion is that arguments are usually selected, not on the basis of a systematic and erpirically validated theory, but quite arbitrarily on the basis of intuition and often fallacious assurptions [such as the idea that a global attitude should determine a specific action)" (p. 349). Unfortunately, Fishbein and Ajzen are not very explicit on how their theory would provide a basis for the selection of arguments or evidence beyond the contention that the arguments and the evidence nust be related to a "primary belief." For instance, Fishbein, Ajzen and McArdle (1980) report the results of McArdle's dissertation on the effects of fear appeals to illustrate the application of the theory of reasoned action for Elaboration, arguments, evidence and style 22 the selection of arguments. The results of the study were that a reasoned action fear appeal was most effective in changing attitudes and behavior, followed in effectiveness by a reward appeal, while a traditional appeal resulted in a boomerang effect on beliefs, attitudes and behaviors concerning the recoumended action. Since the appeals all offered the sannne arguments, the differences between the traditional appeal and the reasoned action appeal are attributable to the manipulation of the specificity (and frequency) of references to the recomnended action rather than the argnmnents per; 533. 'Ihus, while Fishbein and Ajzen argue for messages with reasoning and evidence which have been derived from theory, it is not clear how the arguments selected for this study were based on a theoretic rationale. While there 'renain several conceptual difficulties with the theory of reasoned action, the theory tends to support the ideas that: (1) People enploy rational thinking processes when processing messages (albeit, rationality is yet undefined within the theorY) , (2) Cognitive elaboration is positively related to the belief discrepancy between a source and a receiver, (3) Many individual differences and nessage elements influence a receiver's confidence in his or her own beliefs which, in turn, holds an inpact for how the receiver processes the argmnents in the message (presumably, confidence in one's own beliefs is negatively related to message elaboration). In addition, Fishbein and Ajzen (1975; 1981; Ajzen & Fishbein, 1980) certainly encourage further exploration of the effects of arguments , evidence and message style in persuasive messages . Elaboration, arguments, evidence and style 23 _E_:_l_a_b9ration likghg'm Mud id§_a__s. Petty and Cacioppo (1981) offer an "elaboration likelihood model" (see Figure l) of the persuasion process which they claim to be an extension of the cognitive response approach to persuasion (see Petty, Ostrom, a Brock, 1981) and an integration of previous theories of persuasion and attitude change. 'Ihe major assunption underlying the model is that: line more a person elaborates on the issues presented in a persuasive message, the more likely persuasion will occur and the more enduring the resulting response will be. The model represents two routes to persuasion: A "central route" and a "peripheral route" (Petty & Cacioppo, 1981, chapter 9). The central route has two lanes in the same direction: The receivers' elaboration on the message arguments or the receivers' elaboration on their own arguments related to the issue of concern. In either case, elaboration is determined by the person's motivation (e.g., involvement) and ability (e.g., prior knowledge, self awareness) to elaborate on the issue. 'nne peripheral route involves persuasive factors (e.g., speaker credibility), not innediately relevant to the issue, which "cue" a tenporary attitude shift. Elaboration. While Petty and Cacioppo (1981) place a great deal of enphasis on the concept of elaboration, they never seen to define it. At some points in their discussion of the elaboration likelihood model they use phrases such as "elaborated upon (thought about)" (p. 263) . At other points in their discussion, they use the phrase "the process of Elaboration, arguments, evidence and style 24 Figure l Elaboration Likelihood Model I Persuasive Conrmnication I I Tenporary I I attitude e I F I Is the person motivated to process the cormnnication? Yes fi connunication? Yes Yes I Is a persuas ion I cue resent? Does the person have the abilities to process the Is the person motivated and able No to think about the issues under 1 Yes NO What is the nature of the W I Subjectively I Subjectively I Subjectively strong I weak I anbiguous (Favorable I (Unfavorable I thoughts I thoughts I ggearsed) I rehearsed) J Q What is the nature of the advocacy? I I Counter- I Pro- INeutral —> Attitudinal I Attitudinal I (Favorable I (Unfavorable I thoughts , I thoughts I rehearsed) | rehearsed) i L . Im Boonnerang Retain or regain initial attt’tude I Persuasion I (enduring) Elaboration, argnmnents, evidence and style 25 elaboration (cognitive responses)" (p. 265) . Unfortunately, there is also sone confusion in the cognnitive response literature over what the term cognitive response means. As originally defined by Greenwald (1968) , a cognitive response to a persuasive message consists of " (a) [the] degree of acceptance versus rejection of the position advocated in the communication, and (b) [the] intensity, or vigor, of [the] response" (p. 20). Later articles on the subject, however, tend to broaden the concept substantially. For example, Cacioppo, Harkins, and Petty (1981) claim that "a cognitive response refers to a unit of information pertaining to an object or issue that is the result of cognitive processing" (p. 37). In the same volume of articles, Petty, Ostrcm, and Brock (1981) claim that "a cognitive response includes all of the thoughts that pass through a person's mind while he or she anticipates a communication, listens to a comunication, or reflects on a comrunication" (p. 7). And, in the same volume, Greenwald (1981) states that "cognitive responses constitute a silent internal comminication on the part of an audience menber . . . [which] is assnmned to mediate the effect of the (external) communication on subsequent opinion" (p. 128). Petty and Cacioppo's (e.g., 1977) early operationalization of elaboration also substantially confused the definition of elaboration. Drawing upon previous usage of the thought listing technique for measuring counterarguing (see Miller & Baron, 1973) , Petty and Cacioppo Operationalized elaboration as the nunber of favorable or the number of unfavorable thoughts listed in response to a nessage. The splitting of Elaboration, arguments, evidence and style 26 listed thoughts into favorable or unfavorable categories further confounds elaboration with attitudes. Greenwald (1981) points out that correlations between favorable and unfavorable thought listing counts and attitude measures are so extremely high as to create suspicion that the same construct is being measured. Indeed, Greenwald points out, some attitude researchers have used thought listing as an alternative measure of attitudes. Similarly, Fishbein and Ajzen (1981) have observed that "counterarguing and derogation may sonnetimes be correlates or consequences, rather than antecedents, of acceptance and yielding" (p. 354) . Clearly, the splitting of listed thoughts into favorable and unfavorable categories goes beyond indicating the amount of thonght engaged in by message receivers. Moreover, studies which examine the relationships between favorable and unfavorable thoughts with attitudes may be guaranteed significant (albeit uninterpretable) results. Fortunately, Petty and Cacioppo's more recent research has moved away from tying elaboration to thought listing by examining individual differences such as issue involvement (Petty, Cacioppo, & Schumann, 1983) or "need for cognition" (Cacioppo, Petty, & Morris, 1983) as predictors of message elaboration and enploying nanipulation checks on the amonnt of cognitive effort extended while processing the persuasive messages. Indeed, Cacioppo, Petty, and Morris (1983) discuss message elaboration not in terms of the number of thoughts but, rather, in terms of "greater scrutiny of the information," "effortful thinking" (p. 806) and deliberation. Elaboration, arguments, evidence and style 27 Petty and Cacioppo's (1981) ennphasis on elaborative responses to persuasive messages is based upon both empirical research and scholarly conceptions of cognitive processes. The conceptual bases innclude, for example, Kelman's (1961) distinction between internalization, identification, and conpliance. Petty and Cacioppo (1981, p. 268) claim that since internalization involves the acceptance and integration of new information into the person's belief and attitude structure, internalization represents an elaborative response and, therefore, is descriptive of the central route to persuasion. Petty and Cacioppo claim that identification and conpliance are descriptive of the peripheral route to persuasion since these processes do not involve a consideration of arguments on the issue or a cognitive reorganization and are only effective when the source of the nessage is present or salient (i.e., the attitude change is tenporary) . Similarly, Petty and Cacioppo (1981, p. 268) claim support for their central and peripheral routes by referencing a list of scholars who have proposed two types or levels of human thought. Additionally, Cacioppo and Petty (1982: Cacioppo, Petty, & Morris, 1983: Cacioppo, Petty, & Kao, 1984) have recently included Cohen's (1957: Cohen, Stotland, and Wolfe, 1953) work on need for cognition (i.e., the disposition to "tend to engage in and enjoy effortful analytic activity," Cacioppo, Petty, & Morris, 1983, p. 806) as conceptual support for the notion of elaboration. anviously, while there are a nunber of conceptual perspectives on cognitive processes which ted to support the view that there are different types or levels of cognitive responses to messages, all of these conceptions do not reduce to elaboration vs. nonelaboration Elaboration, arguments, evidence and style 28 as Petty and Cacioppo seen to suggest. Nevertheless, Petty and Cacioppo have offered evidence that there is sonne conceptual support for the exploration of elaborative responses. There is sonne empirical support for the proposition that elaboration is a potent factor in persuasion. Petty and Cacioppo (1981) are quick to point to the research on distraction and resistannce (e.g., Petty, Wells, Brock, 1976) , comterattitudinal advocacy or role playing (see Miller & Burgoon, 1973: Janis a Dunn, 1977), and forewarning (see Cialdini & Petty, 1981) to support their emphasis on elaboration. Specific studies of elaboration, however, are sparse. Love and Greemvald (1978) asked subjects to read a persuasive message, write out one sentence reactions to the three main arguments in the message, and fill out recall and attitude measures. On both immediate and delayed posttests, the snbjects' recall of their own reactions was correlated with their new attitudes while recall of the message arguments was not correlated with their new attitudes (pretest attitudes were partialled) . Petty (1977) demonstrated that subjects' recall of their own thoughts to arguments on raising the driving age to twenty-one predicted attitude change on both inmediate and delayed posttests while recall of the "strong" (vs. "weak" or "neutral") message arguments predicted attitude change on only the immediate posttest. 'Ihese studies provide empirical evidence that elaboration is important to persuasion. Research on factors that predispose elaborative responses provide the most convincing evidence that elaborative responses to messages are a potent determinate of persuasion. Cohen (1957) demonstrated that subjects Elaboration, arguments, evidence and style 29 who scored higher on need for cognition were less affected by order effects in persuasive messages. Petty and Cacioppo's examinations of "argument quality" and elaboration determirnants such as issue involvenent (e.g., Cacioppo, Petty, & Schuman, 1983), need for cognition (Petty, Cacioppo, & Morris, 1983) or elaboration enhancement such as message repetition (Cacioppo & Petty, 1980a) and elaboration inhibitors such as distraction (Petty, Wells, 5. Brock, 1976) have all resulted in an elaboration by argument quality interaction such that the greater the elaboration likelihood, the more positive the attitudes for messages with "strong arguments" and the less positive the attitudes for messages with "weak arguments." In other words, argument quality influenced attitudes only when elaboration likelihood was high. Unfortunately, these studies tell us nothing about the types of arguments that are persuasive. Petty, Wells, and Brock (1976) conceptually define argument quality as "difficult to counterargue." Cperationally argnment quality is typically defined "such that 'strong' arguments elicit more favorable than unfavorable statenents, as assessed by the thought-listing technique . . . and 'weak' arguments elicit more unfavorable than favorable thoughts" (Cacioppo, Petty, & Morris, 1983) . In addition, the specific strong vs. weak arguments employed in the studies have varied on content, structure, support, recency, relevance, conplexity and absurdity. A typical example can be found in CacioPpop Petty, and Morris (1983) where the strong version of an argument for a tuition increase was that classrooms were overcrowded. The weak version of the argument was that the road to the new campus sports arena was Elaboration, arguments, evidence and style 30 overcrowded. This lack of control of argument relevant variables casts doubt on what can be interpreted frcmn these studies about the effects of argument quality. Further, the conceptual and operational definitions of argument quality presuppose prior knowledge of what would persuade an endience and are, therefore, not theoretically useful for initial audience annalysis, message design (i.e., prediction) or explanation of persuasion. Still, if viewed in the light of formative exploratory research, these studies suggest that argument quality and message elaboration are important to the persuasion process. Perifingral route. Petty and Cacioppo (1981) are not very clear about the peripheral route and how it operates. 'Iheir discussion of the theories of persuasion which they claim to be linked to the peripheral route, indicates that multiple processes may be involved and that a peripheral cue may be shifted to central route elaborative processinng (see Petty, Cacioppo, & Schunmann, 1983, p. 143). (bviously, without an explicit account of the peripheral route, many persuasive effects can be cast teleologically into either the central or peripheral routes on the basis of the degree of cognitive effort reported or how enduring the response is. 'llnere are relatively few studies conducted by the Cacioppo and Petty research team which address the nature of the peripheral route. Petty, Cacioppo, and Schumann (1983) examnined issue involvement (operationalized as probability of being inmnediately affected), argument quality, and celebrity vs. noncelebrity endorsements ("professional athletes" vs. "Bakersfield California") and found that highly involved subjects' (i.e., predisposed to elaborate) postmessage attitudes were mnost affected by Elaboration, argunnents, evidence and style 31 strong arguments while low involved subjects' postmessage attitudes were most affected by celebrity endorsennents. In a study of information processing strategies, Cacioppo, Petty, and Sidera (1982) suggest, but do not test, the idea that persons predisposed to elaborate ted to veridically evaluate message content and structure while persons nnot predisposed to elaborate merely attenpt to place message content within their preexisting cognitive schemata. 'I'ne conparisons between Petty and Cacioppo's (1981) central and peripheral routes with Mohaire's (1960a) rational and wishful thinkinng seem obvious. The main distinction seems to be that while McGuire (1960a: 1981a) assumes that rational and wishful thinking simultaneously contribute to the generation of a person's response to a message, Petty and Cacioppo (1981) assume that each of the processes are unique to individuals, persuasive messages, and stages of the persuasive process (see also Stiff, 1986) . . Aggication t9_g_r_s_p_a_sip_p. The elaboration likelihood mnodel has some intuitive appeal and, at first glance, is even exciting, given Petty and Cacioppo's (1981) claim of integrating much of the established theory and research on persuasion. At second glance, however, the model gives rise to manny more questions than it answers. Nevertheless, the research coducted under the elaboration likelihood model supports the conclusion that what persuadees think (or do not think) about a persuasive message, particularly their reactions to the arguments in the messages (Love & Greenwald, 1978) , is more important to persuasion than what they remember Elaboration, argunents , evidence and style 32 about the message itself. The elaboration likelihood model, however, does not specify what reactions are important or how an advocate might elicit such reactions. Despite theoretic annd enpirical problems in the research, the consistent reports by Petty, Cacioppo and their associates that the influence of argument quality depeds on the annnount of cognitive effort extended while processing a persuasive message affords grounds for a reexamination of Thonpson's (1967) concern about the effects of the traditional standards of argument quality (i.e., sound argunents and evidence). Similarly, the thinking and research on the peripheral route to persuasion suggests that message elennents not directly related to the message argunents (e.g. , message style) may produce effects which are more epheneral and are mediated more by the receiver '3 affective response to the message than by a thoughtful consideration of the message arguments. __.Snma.nr_._.of thepespeetiyemhw. nwimnusgcwial “influence- While there are many differences in both conceptual assumptions and empirical bases for the probabilogical, reasoned action and elaboration likelihood perspectives on hunnnan reasoning in social influence, the similarities seenn much more important. All three perspectives contain the assumption that human reasoning is central to the effects of persuasion. Two of the three perspectives (probabilogical and reasoned action) contain the assunption that hunnnan reasoning is rational. Two of the perspectives (reasoned action and elaboration likelihood) contain the assunption that sound argnmnents and evidence should be more persuasive than unsound arguments and faulty evidence. Two of the perspectives (probabilogical Elaboration, argunents, evidence and style 33 and elaboration likelihood) advance the notion that cognitive elaboration is important to the effects of reasoning in persuasion. All three perspectives contain sone account of how nonargument or extra-argument relevant variables may affect persuasion. Unfortunately, all three perspectives offer inconplete accounts of what types of , or how, reasoning, evidence and message style affects persuasion. Nevertheless, the general themes and ideas from the perspectives on hunan reasoning in social influence appear to be integrable into a theoretic model that may account for reasoning, evidence and message style. A model of belief elaboratipngdittitydg in pgrsuasio . Despite the unnwieldy nunnber of variables discussed in the writing on the probabilogical, reasoned action and elaboration models of persuasion, there are a relatively few nunber of key concepts and relationships that emerge as significant to advanncing a speculative model to aid in the understanding of the role of elaboration, reasoning, evidence and style in persuasion. Among the key concepts, the notion of cognitive elaboration seenns to be fundamental. Elaboration has been referred to mainly as "thinking" (Petty 8 Cacioppo, 1981) or "effortful thinking" (Cacioppo, Petty, & Morris, 1983) and operationally defined in terms of a self perception nmeasure of cognitive effort (Cacioppo, Petty, & Morris, 1983) while processing a persuasive message. It is implicitly assured that when a message recipient engages in elaboration about a message, the recipient enploys sore- degree of discriminative skills (cf. Siegel, Miller, & Wotring, 1969) in the "scrutinny of information" (Cacioppo, Petty, & Morris, 1983). Thus Elaboration, argunents, evidence and style 34 a plausible assunption is that when persuadees elaborate on a message they engage in sonne assessnnent of the arguments and evidence in the message. If so, then the greater the degree of elaboration on a message, the greater the likelihood of detecting reasoning or evidence errors in the message. There are undoubtedly many potential antecedents to the degree of elaboration engaged in while processinng a message. Petty and Cacioppo's work has identified several individual differences (e.g., need for cognition, issue involvenent) , message variables (e.g., distraction, rhetorical questions), and source credibility as potential antecedents of elaboration. The review of Fishbein and Ajzen's (1975) work leads to the speculation that belief discrepancy and confidence in one's own beliefs are antecedent to message elaboration. In addition, the evidence presented by McGuire (1960a: 1960b) on the reduction of rational consistency pronpted by the desirability of an argunent conclusion suggests that a respondent 's attitude toward the advocated belief is antecedent to message elaboration. An attenpt at synthesizing across the potential antecedents of message elaboration seenns to be an overly ambitions task, and perhaps prenature, in light of the formative stage of research and thinking on message elaboration. For now, it seems sufficient to conclude that: (1) There are individual differences which predispose message elaboration; and, (2) the attitude held for the advocated belief tends to be negatively associated with message elaboration. In pursuing these links, future research should be directed at message differences which might affect message elaboration. Elaboration, argunents, evidence and style 35 The link between elaboration and belief change implied by Petty and Caciogno's (1981) elaboration likelihood model is mediated by the persuadee's evaluation of the message argnments. The nature of this evaluation, however, has not been specified. At some points in their discussions of the model, Petty and Cacioppo (e.g., 1981) indicate that it is the reactions specific to the message argunents that are important. At other points in their discussions (see, in particular, Cacioppo, Petty, & Morris, 1983) they indicate that is is the more global impression of message effectiveness that is innportant to persuasion. Sinnce it is assuned that people differ on the degree of message elaboration, it would seem untenable to expect that those who elaborate less are likely to form reactions specific to the message arguments. On the other hand, message receivers are likely to form global impressions of message effectiveness regardless of the degree of message elaboration. 'Ihus, it seenns consistent with much of Petty and Cacioppo's work to suspect that those who ted to elaborate more on messages base their evaluations of messages on specific reactions to the message argunents (i.e., detection of reasoning or evidence errors). For those who ted to elaborate less on messages, the research on the effect of attitudes on reasoning judgments (e.g., Janis a Frick, 1943) would suggest that their evaluation of a message is likely to be based on their attitude toward the persuasive thesis. On the basis of the thinking and research reviewed above, the follov- ing propositions energe as the basis for a model of belief elaboration and attitudes in persuasion: Elaboration, arguments, evidence and style 36 Proposition 1: The greater the predisposition to elaborate, the greater the likelihood of elaboration on a message. Proposition 2: The more positive the attitude tovards a message, the lover the likelihood of elaboration on the message. Proposition 3: The greater the degree of elaboration on a message, the greater the detection reasoning errors in the message. Proposition 4: The the greater the detection of reasoninng errors, the lower the message evaluation. Proposition 5: The more positive the attitude towards a message, the higher the message evaluation. Proposition 6: The higher the message evaluation, the greater the likelihood of belief channge. While these initial theoretic ideas are admittedly more annalytic than synthetic (enpirically grounded), they appear to be consistent with much of the thinking on hunan reasoning in social influence. The relationships between these variables is sunmarized in the path model presented in Figure 2. Employing this model as a base, potential predictions for the effects of message style, reasoning and evidence may be developed in a more systematic manner than has been previously afforded by the research on hunan reasoning in social influence. glibpratmuandrlpggalyagggmgnts. As noted earlier, there is a paucity of empirical research which conpares valid and invalid argunents in persuasive messages. Indeed, an extensive search has failed to discover a study which has directly compared the persuasive effects of Elaboration, argunents , evidence and style 37 Figure 2 A Model of Belief elaboration and attitudes in Persuasion '-- - ‘— _ — -—-"””- Where: mm-- --- --“ A = attitude. DRE = detection of reasoning errors. E = elaboration. ME = message evaluation. pC = probability of belief in the conclusion of the message. PE = predisposition to elaborate. ---—-mw--“’"- ‘A- - -‘ Elaboration, argunents , evidence and style 38 valid argunents and invalid argunents in messages. There are studies of 'logical' and 'enotional' messages but, as Becker (1963) has noted, the conception of 'logic' vs. 'enotion' in these studies is flawed and the operationalizations of logic are typically not what a logician would classify as logical. There is also a growing body of research and theory on hunan reasoning ability (for reviews see Miller, 1969; Wason & Johnson-Laird, 1972: Sample, 1979a, Evans, 1980: Reynolds & Burgoon, 1983) which is variously interpreted as showing that hunnan reasoning is or is not logical (cf. mmple, 1979a, with Evans, 1980). In addition, McGuire (1960a: 1981a), Wyer (see Wyer & Hartwick, 1980) and Bample (1977: 1978: 1979b) report a nunber of studies in which logical consistency predictions for persuasive messages were tested and supported. The findinngs are, however, limited since logical argunents were usually not manipulated or the control groups received "no-communication" rather than messages with invalid argunents. Clearly, many questions about the effects of valid arguments in persuasive discourse renain unanswered. For the model of belief elaboration and attitudes in persuasion it is presunned that the elaboration process involves some form of reasoning and that the reasoning involved approximetes logical forms of thought (see Appedix A). If peOple tend to reason in rational ways when, in fact, they engage in elaboration, then it may be assumed that they are likely to detect errors in their own reasoning as well as in the reasoning of others. Thus, if a message recipient does indeed elaborate on the message and the message contains errors in reasoning, then the recipient is likely to detect the errors and discount the message argunents when evaluating the persuasive message. Elaboration, argunents, evidence and style 39 Evidence. and elabprgtipp. Given the enpirical support for expecting that the use of evidence in persuasive messages will aid persuasion (see Reynolds & Burgoon, 1983) and the need for a theoretical explanation of that affect, it is appropriate to examine how the model of belief elaboration and attitudes in persuasion might account for the effects of evidence. Miller (1966) defines evidence as "those data that are intended to induce a sense of belief in the propositions which the data purportedly support" (p. 25). McCroskey (1982) has proposed that there are three types of data: First-order, second-order, and third-order. First-order data are those beliefs which are held in comnon by both communicator and receiver. Secod-order data are those beliefs held by the commnnicator but not necessarily held by the receiver (in other words, "source assertions"). Third-order data are those beliefs which are obtained from a source indepedent of the conmunicator or the receiver. McCroskey limits the use of the term 'evidence' to third-order data. Most of the research on evidence in persuasion tends to examine third-order data. There are a few studies (Harte, 1977: Kline, 1971a: 1971b; Fleshler, Ilardo, & Denoretcky, 1974: Bradac, Sandell, a Wanner, 1979) that have examined the ability to detect violations of evidential standards. While the conclusions are generally negative, each of the studies report that people are better at detecting sone violations more than others. It is worth note, however, that low dognnatics (Kline, 1971b) and field-dependent persons (Fleshler et a1, 1974) are better at enploying evidential standards than high dognnatics and field-indepedent persons. Moreorer, Elaboration, arguments, evidence and style 40 Luchok and McCroskey (1978) report that when evidence is attributed to a source who is not qualified to comnent on an issue, and when the actual content of the evidence is irrelevant, an advocate (particularly one with low-to-nnoderate credibility) will find his or her audience channging in a direction opposite from that advocated. Perhaps dogmatisnm and field depedence are related to message elaboration. Petty, Cacioppo, and Schunan (1983, p. 143) have previously proposed that speaker credibility influences message elaboration. Thus it appears that the effects of evidence in persuasion are more likely to occur when persuadee engage in message elaboration. ___Messme _s_tyle, attitudes and glgbpgation. Style is typically defined as a characteristic way of making linguistic choice. Sandell (1977) attenpts to limit style to nonsenantic linguistic choice but finds that qualification difficult to maintain when specific stylistic tactics such as message intensity or metaphors are considered. Indeed, style is often discussed in broad termns. Norton (1978) , for example, advance the concept of "commnnicator style" which enconpasse individual dispositional difference and associated verbal and nonverbal commnnication patterns (e.g., an "attentiveness style," Norton & Pettigrew, 1979) . Even Sandell conceptualized style as similar to the personality concept in that a partionlar style consists of a set of stylistic elements or a "composite of traits" (p. 15). While an examination of clusters or dimensions of stylistic choices may well yield worthwhile insight, any effort at studying style should profit from the limitations of past efforts at studying style. Miller, Elaboration, argunents, evidence and style 41 Burgoon and Burgoon (1984) note that, historically, the rhetorical scholarship on style has failed to contribute significantly to our uderstanding of persuasion. This lack of contribution is due to a "major preoccupation . . . [with] sorting and labeling various tropes ad figure . . . [rather] than with positing functional relationships between variable" (p. 403) . The net reult has been the production of a "vast and ponderous" glossary of stylistic tactics rather than a theoretically meaningful and testable taxononny (see Kaplan, 1964, p. 111-112, and Benpel, 1965. For a sanple of the various tropes and figures see Corbitt, 1971, who is also critical of the excesiveness of the Tudor figurists) . Given the lack of knowledge about stylistic factors or traits, "it is probably wise to susped judgnnents about what factors of style there 'are', restinng content with the indications of the abundance of variable amenable to quantification" (Sandell, 1977, p. 46). While the Tudor rhetoricians were amiss, there have been sone efforts to link stylistic variable to other conmunication relevant variable. Certainly, Gorgias (Plato, c. 387 B.C./ 1952), Longinus (c. 100/ 1965), the elocutionists, and Blair (1783/ 1965) have suggeted that style adds ornnateness or beauty to discourse which should enhannce sonrce credibility. Longinus (c. 100/ 1965) points out that stylistic excese should decrease source credibility. Joseph (1947) attempted to classify the various figures of speech under the four categorie of grammar, files, 10398, and gangs. Perelman and Olbrechts—Tyteca (1969) claim that style can be "argumentative" in the sense that the form and the content of a "rhetorical figure" is used to advocate "a change in perspective" (p. Elaboration, argnments, evidence and style 42 169) . Wilson and Arnold (1978) argue that the primary function of style is to increase the clarity of expresion as well as to foster attention, conprehension, and retention of the message (see also Thomas, 1956; Bowers & Osborn, 1966; Zillman, 1972; Zillman and Cantor, 1974). Burke (1969) and Sandell (1977) have proposed that style affects receivers' impresion of homophily with the source. Each of these effects has received, at best, equivocal support (for reviews see Sandell, 1977: Bradac, Bowers, & Courtwright, 1979: Frandsen & Clenent, 1984). By far the most docunented effect of style involve channges in attitude towards the issues in message. The enpirical research on the affective implications of style has examined a nunber of stylistic variable such as language intensity, opinionated language, obscene language (For reviews, see Burgoon & Miller, 1985, and Bradac, Bowers, & Courtwright, 1979) , metaphors (Bowers a Osborn, 1966: Reinsch, 1971: 1974: Jordon, 1972: Siltanen, 1981), analogy (PbCrosky & Conbs, 1969), and rhetorical questions (Zillman, 1972; Zillman a Cantor, 1974: Petty, Cacioppo, & Beeacker, 1981) . The effects found in these studies, however, have frequently energed within interactions with source characteristics (Bostronnn, Basehart, & Rossiter, 1973: Burgoon, Jones, & Stewart, 1975) , and receiver characteristics (Burgoon et al., 1975: Burgoon & Chase, 1973: Basehart, 1971: Petty, Cacioppo, & Beeacker, 1981) . The conplexity of these interactions on affect along with the uncertainty about the other potential effects of style have fostered skepticisnm and complaints abolt theory develognent on message style Elaboration, argunents, evidence and style 43 (Sandell, 1977: Bradac et a1, 1979: Frandsen & Clenent, 1984; Miller, Burgoon, & Burgoon, 1984). Burgoon and his associate (see, in particular, Burgoon & Miller, 1985) , however, are actively advancing a theory of linguistic expectancy violations which has denonstrated potency for accounting for the interaction effects of stylistic variations on attitude. The central proposition of the theory of linguistic expectancy violations is that when a source violate normative expectations about appropriate communication behavior, receivers' attitude will be affected. If the violation exceeds what is expected, the attitude will change in a positive direction. If the violation falls short of what is expected, the attitude will either not channge or change in a negative direction. Thee interaction effects may also be, in part, a reflection of the degree of message elaboration engaged in by the message recipients. In sonne cases it may have been that stylistic tactics produced an affect towards the message thesis that reulted in a reduction of elaboration. In other cases the stylistic tactic may have contained infornnation relevant to both affective reactions and elaborative efforts. For example, language intensity may both violate a receiver's notion of appropriate language on sonne statenents while supplying information on the probable truth of other statenents. (See the operationalization of language intensity by Miller, 1978.) Unfortunately, such a conclusion is at best speculative since elaboration was not a variable of cocern in the studies . Elaboration, argunents, evidence and style 44 A study that reveals the ccmplexitie of studying the relationship between style and elaboration is offered by Petty, Cacioppo and Heesacker (1981) who report three-way interactions between argunent quality, style (the use of rhetorical questions) and predisposition to elaborate ("issue involvenent" or propensity to be affected by the proposal) on attitude change and listed thoughts. Petty et a1. interpreted their data as showing that rhetorical quetions enhanced elaboration for those not predisposed to elaborate and distracted those who were predisposed to elaborate. A closer exannination of their data on thoughts listed and a self perception measure of distraction, however, does not support the enhancement-distraction explanation. Recalling that argunent quality was operationalized in terms of preexisting reponse to the arguments, Petty et al. '5 data actually indicate that for the no rhetorical question coditions, the subjects' elaboration (tholghts listed) was determined by the predisposition to elaborate and the preexisting reponse to the argunents. In the rhetorical question conditions, the subjects predisposed to elaborate gennerated more counterargunents and fewer favorable arguments than other groups while those not disposed to elaborate generated thoughts consistent with the preexisting reponse to the argunnnents. Apparently rhetorical quetions enhanced elaboration regardles of predisposition to elaborate. It was the fornnn of elaboration (consistent with preexisting responses vs. responses unique to the message) that was affected. There were actually two groups that reported being distracted: 1) Those with low issue involvement (i.e. , low predisposition to elaborate) that did not econnter rhetorical quetions and, 2) those with high issue Elaboration, argunents, evidece and style 45 involvement that did encounter rhetorical quetions. In short, the self reports of distraction were apparently not related to the thoughts listed. (Perhaps the distraction score was more of an indication of actual elaboration or elaboration avoidance.) Unfortunately, it is difficult to offer an alternative interpretation of the three-way interaction reported by Petty et al. in light of the operationalizations of argunent quality, predisposition to elaborate, attitudes (Petty et a1. averaged standardized score fromn a traditional attitude measure with standardized score from a sinngle agree-disagree belief measure without reporting standard deviations or reliabilitie) and the lack of information about the subjects' preexisting attitude on the topic. Nevertheles, the interactions seem to be more a function of the conflicting tedecies to engage in elaborative thought than the effect of style as a persuasive cue. Indeed, the rhetorical quetions invited thonght (e.g., "Don't yon agree that . . ." and "Isn't it true that . . " p. 436) as did the task given to the experimental subjects (to evaluate editorials for broadcast on the college radio station). In addition, the use of propensity to be affected (issue involvement) as an operationalization of predisposition to elaborate does not seem to capture the notion of habitual cognnitive effort. At any rate, it is clear that enhancenent or distraction from elaboration do not account for the results. One study that teds to support the speculation abont the relationship between style and elaboration is Basehart's (1971) invetigation of opinionated language and receivers' need for approval . Basehart reports Elaboration, argunents, evidence and style 46 an interaction between need for approval and opinionated language such that those with high need for approval were influenced by the opinionated statenents of the advocate. (Low need for approval and high need for approval without opinionated language groups were not significantly influeced by the message.) The suspicion that the effect of opinionated language on high need for approval receivers is due in part to the receivers' failure to elaborate on the message is partially corroborated by Basehart's report that the high need for approval snbjects compreheded the message les than did the low need for approval subjects. The expectancy violation effect on attitude is cocomnitantly viable in the study since the high need for approval subjects may well have been pleasantly surprised with the directed information provided by the opinionated language. Thee ideas on mesage style, attitude, and elaboration offer insight and a caveat for the role of message style in social influece. The caveat is that stylistic tactics must be examined, in light of normative expectations, for the potential impact on affect as well as the potential impact on message elaboration. Given that this study has made an effort to avoid potential message variable which may directly affect to anount of elaboration it is necessary to offer only the prediction that stylistic tactics will influence receivers' attitude on the persuasive thesis. The considerations of argunent validity, evidence, and message style above suggest that the followinng propositions should be incorprated into the an extended model of belief elaboration and attitude in persuasion: Elaboration, argunents, evidece and style 47 Proposition 7: The greater the degree of elaboration on a message, the higher the evidece evaluation for a message. Proposition 8: The higher the evidece evaluation for a message, the higher the message evaluation will be. Proposition 9: The stronger the evidece in a message, the higher the message evaluation will be. Proposition 10: The use of a stylistic tactic in a message will reult in message receivers being more likely to rate the language in the message as unexpected. Proposition 11: The higher the ratings that the language in a message is unexpected, the higher the attitude toward the message will be. (Assuming a positive violation of expectations.) Proposition 12: The use of valid (vs. invalid) argunents in a mesage will reult in les likelihood of detecting reasoning errors in the message. The extended model of belief elaboration and attitude in persuasion is preented in Figure 3. The exteded model incorporate the links between logical arguments and evidece in a persuasive message to the detection of reasoning errors and incorporate the links between message style, expectancy, and attitude towards the persuasive theis. Elaboration, argunents, evidece and style 48 Figure 3 An exteded Model of Belief elaboration and attitude in Persuasion mm pa 33/ \ ME——fipC Where: pC :- probability of belief in the coclusion of the message ME - message evaluation DRE = detection of reasoning errors = evidece evaluation -= validity of message argunennts -= strength of the evidence in a message = elaboration - predisposition to elaborate = attitude = expectation a message style Sfi’fimifl‘Wfi Elaboration, arguments, evidence and style 49 Methods In order to explore and tet the relationships proposed in the extended model of belief elaboration and attitudes in persuasion, measure or operationalizations for each of the variable were selected or developed and applied in an experimental setting in which subjects were asked to respond to a persuasive message. The persuasive message was on the topic of the legalization of heroin and the subjects were requeted to repond to the message as if it were an editorial or magazine article. The experimental manipulations and measure necesary for teting the model were contained in a single quetionnaire booklet. Following conpletion of the quetionnaire booklet, the subjects were debriefed and thanked for their participation. Following a check upon the measurement model, path analytic procedure were applied to test the model. '_I'_G:§J:_9f the Jmpdel The proposed model was teted with path analysis (see Nie, Hull, Jenkins, Steinnbrenner, & Bent, 1975, pp. 383-397, and Kenny, 1979, chapter 3). Specifically, the path analysis conputer program for personal conputers written by Hunter and Hamilton (1986) was enployed to calculate path coefficients and the sum of squared errors (or etimated unspecified correlations based upon the path coefficients). A chi-square goodnes of fit tet was applied to tet if the data differed significantly from what the model predicts. Elaboration, argunents, evidece and style 50 @242 Subjects were recruited from basic communication course at a large Western state university with a multicultural student population. Since the probabilogical approach reearch has obtained correlations between beliefs in premise and beliefs in conclusions at .40 or better and given the uncertainty about the strength of the relationships proposed in the model under tet, caution would dictate that a population r or .30 should be used in a power analysis (Cohen 5 Cohen, 1975. pp. 117-118) to deternmine an appropriate sanple size. With gong set at .05 and a deired power of .80 or better, the mninimal sanple size for the path analysis equations in this study is 216. The actual sanple size was 240. The subjects participated in the study for extra credit in their courses. we construction glggtjgn. The health related persuasive message on the legalization of heroin constructed by Burgoon, Cohen, Montgomery and Miller (1978) was enployed in the reearch effort. There were five reasons for selection of this message. First, the message was selected because the topic is known to be associated with strong attitude and was, therefore, regarded as a topic likely to be involving to the reearch participants. With such strong attitude and potential involvenent, the message was expected to maximize the potential for elaboration. Secod, the message was selected because the topic was one which was likely to be unfamiliar to most participants. McCroskey (1969) has deonstrated that prior knowledge on a topic reduce the effects of Widence. This reduction is, preunably, because the receivers rely on Elaboration, argunents, evidece and style 51 their own knowledge rather than that provided in the message. A simnilar effect might be expected for the arguments in a message. Thus, the message selection aids in controlling for the influece of prior knowledge on the processing of the message arguments and evidence. Third, the message constructed by Burgoon et a1. (1978) has been shown to be persuasive (see also Miller, 1978) . The necesity of denonstrated persuasivenes of the experimental message for this invetigation is obvious. Fourth, the message was selected because it allowed for manipulations of logic and evidece with a minimum of reconstruction. The Burgoon et a1. (1978) message contains four argunents which are deductively valid. (Since the fonrth argument was enthymematic, an additional premise was inserted to conplete the logical structure. Since the seed argunent advanced a premise within the evidece for the premise, a formal statenent of the premise was added.) In addition, each of the argnmnents in the message was supported with evidece. Fifth, Burgoon (Burgoon et al., 1978: Burgoon 5 Miller, 1985) has argued that the lack of control of message variable in commlnication reearch has lead to conflicting and misleading findings. Thus the Burgoon et a1. message was selected because it had been constructed to control for contingency (Becker, Bavelas, & Braden, 1961) , subject- predicate conpatibility (Clark & Begun, 1971) , language intensity (Burgoon, 1970) , sentence length, and total nunber of words. Given such controls in the initial construction of the message, the difficulty of reestabl ishing thee controls after the manipulations of logic, evidence and message style were mninimnized. Elaboration, argunents, evidece and style 52 ngiyiity. The four arguments in the original message are in the form of deductively valid universal affirmative (Barbara) syllogisnms. The invalid versions of the argunents were created by rearranging the terms and the propositions such that no valid alternative conclusions could be reached by the premise conbinations. The validity manipulations were independently verified by a logic and reasoning reearcher. The universal affirmative form of the argunents was maintained. (See Table 2, Appendix D, for the argunent validity manipulations). mace. The evidence included in the original version of the message was not accompanied with source citation or source qualification statenents. Since reearch on the use of evidece indicate that providing source citations and qualifications enhance the effects of evidence (Bostrom 5. Tucker, 1969) , source citations and qualifications were added to the messages for the high quality evidence versions. For the poor quality evidence coditions the evidece was altered so as to violate two or more of the standard tests of evidece. In each instance, the source citations were deleted and the qualifications were altered so that the source of the evidence appeared to be marginally qualified on the issue to provide the evidece. In addition, each item of the evidece was altered so as to not directly support the premise being advanced. The specific evidece manipulations were: Argument 1 Strong version of evidence: For exanple, in 1979, a team of medical reearchers from the Federal Drug Ennforcenent Administration reported in the £6;de Journal of Megigipg that over 900 addicts died in New York City from tetanus and hepatitis. Weak version of evidece: For exanple, in 1969, a teann of prenedical students from Slippery Rock State College reported that a significant nunber of addicts died in New York City from secodary illnesse reulting from the use of the drug. Elaboration, argunents, evidence and style 53 Argument 2 Strong version of evidence: According to the New York Tynes, reearch by Dr. Al Jordan, of the Harvard Medical School, shows that "when heroin users know the actual strength of the drug they're using, deaths and medical problems caused by overdosing are virtually eliminated . " Weak version of evidece: According to Al Jordan, secretary of the Seattle Drug Rehabilitation Council; "It seems that when herOln users know the actual strength of the drug they're using, deaths and medical problems caused by overdosing are reduced." Argunent 3 Strong version of evidece: Profesor William D. Shintig of the UCLA Center for Drug Addiction Research recently wrote in the UCLA @ R_e_v_i;e1w that "etinnnate by several indepedent reearchers indicate thatmhegoin addicts spend eighty-five percept or more of their income on e rug." Weak version of evidence: Bill "Fetch" Shintig of the National Coalition for the legalization of Marijuana claims that "heroin addicts depend on others to supply them with food and clothe because they can barely find the reources to pay the high costs of the drug." Argument 4 Strong version of evidece: Medical authoritie from the World Health Organization now agree that heroin cause little physical danage. In addition, a team of researcher from Johns Hopkins University reported at a recent congresional hearinng that the synnptonns of heroin withdrawal are much les dangerous than those associated with the withdrawal from alcohol. Weak version of evidence: It is now conmnonnly agreed that heroin causes little physical damage. In addition, a health and nutrition researcher from the University of Arkansas claims that the symptons of heroin withdrawal are about as bad as those associated with the withdrawal fronn extrene case of alcoholism. gimme. Manny stylistic tactics may influece judgments of the probability of truth or alter the logical form of an argument. Further, they might invite elaboration on the argunents. Therefore, it was important that the stylistic variable enployed be one that wonld not have these effects. Nevertheles, the stylistic variable selected should be positively associated with affective response to the persuasive effort. Elaboration, argunents, evidence and style 54 The first step in the selection of a stylistic variable was to elimninate all figure of speech that directly deal with the general thenes or content of a message. This step reulted in the elimination of most of the tropes (changes in the meaning of words or phrases) and all of the figure of thought and amplification. The secod step was the elimination of all schenes (channges in the pattern or arrangenent of sounds or words) that involve hyperbaton (unusual or inverted word order) or change in the spellinng or solnds of words. These schenes were eliminated in order to avoid conplicating the reception of the message. The third step was to eliminate all schenes that involve repetition of words or phrases. These schenes were eliminated on the grounds that such repetitions might enhance learning and recall of the arguments and thereby confuse affective response with elaboration and elaboration with message learning. The remaining list of stylistic variable consisted of a few tropes (mostly word substitutions) and schenes of sound repetition. From this list, alliteration (repetition of initial or medial sonnds) was selected as the one stylistic tactic most amenable to manipulation and most likely to be sufficiently conspicuols to influece reponse to a message. Onnly one tactic was selected in order to maximize the probability of isolating the effect of message style. The specific stylistic manipulations were: Paragraph one, Style: . . . the health and hopefulnes of hnmanity. No style: . . . the health and welfare of its citizens. Paragraph two Style: . . . many addicts face dramatically dreary deaths from disease. Elaboration, argunents, evidence and style 55 Paragraph two, No style: . . . many heroin addicts are dying needlessly from disease. Paragraph three, Style: A secod health hazard issue that we cannot ignnore is that illicit heroin is impure and, thus, intrinsically injurious. No style: A second health hazard faced b the user 15 the parchase of impure hero n. Paragraph fonr, Style Since the cost of illegal heroin is so excesive, users are often unable to afford expediture esential for ecapinng poor health. No style Since the cost of illegal heroin is so excesive, users are often unable to afford itens esential for maintaining good health. Paragraph five, Style A fonrth dimension that must be discussed is whether heroin is actually a dangerous or damaging drug. No style A fonrth issue that must be considered is whether the use of heroin constitute a real health problenn. Paragraph six, Style . . . renoving secodary infection reducing the risk of overdose, raising the availability of nutrition and medication, and recognizing . . . No style . . . eliminating secodary infection, reducing the nunber of overdose, and increasing the availability of nutrition and medication, coupled with . . . Following the insertion of the logic, evidence and style manipulations the experinnental message were compared on productivity (total nunber of words. The message were constructed so that each contained exactly 29 sentence.) , average sentece length, lexical diversity (type-token ratio), redundancy, pausality, enotivenes, conplexity, readability (Flesch score), and contingency. (For a review on all of thee measure, Elaboration, argunents, evidece and style 56 except for contingency, see Lynch, 1970. The measure of contingecy, which indicate the reconstructability or predictability of a message, was developed by Becker, Bavelas, and Braden, 1961) . Since these measure are sensitive to difference in the encoding and decoding of messages and are generally representative of the varions dimensions of message cmposition (see Lynch, 1970) , consistecy of value across the message should assure a similarity in composition and reception. Such controls were enployed to aid in isolating the effects of the specific manipulations in the study. The initial evidece manipulation reulted in a noticeable difference in contingency score between the weak and the strong evidence versions of the message. This differece appeared to have been the reult of the deletion of the source citations. (The continngency index is a conplex ratio of total nouns, total words, and repeated nouns weighted by appearance in the sane or succesive sentence.) In order to correct for this differece the weak versions of the evidece manipulations were reconstructed to include more nonns. Following the reconstruction the stylistic measures were reconputed. The score on each of the stylistic measure are preented in Table 1. The score across the message do not differ substantially. It is worth note that while the contingency index for each of the messages is fairly high, the readability index is low. Since both of thee indices are supposed to be positively related to conprehension the differece appear to preent a conflict. After reexannininng the indices however, it may be argued that the difference are offsettinng rather than Elaboration, argunents, evidence and style 57 Table l Conparison of stylistic indice for the reearch message Logic Valid Valid v'a'l'id""v‘a'lld””‘lhvalid Inva‘l‘i'd Invalidnlnvalid Eyidence Strong Strong Weak Weak Strong St rong Weak Weak Style Ye No Ye No Yes No Ye No Total words 720 716 716 712 725 721 721 717 Average sentence 24.80 24.70 24.70 24.60 25.00 24.90 24.90 24.70 length lexical Diversity .44 .45 .43 .43 .43 .43 .42 .43 Redundancy 1.30 1.40 1.40 1.50 1.60 1.70 1.80 1.80 Pausality 2.40 2.30 2.30 2.20 2.40 2.40 2.40 2.30 Enmotivenes .30 .29 .29 .28 .28 .27 .27 .26 Conplexity 1.71 1.70 1.68 1.67 1.68 1.68 1.66 1.65 Readability 36.99 37.95 39.64 40.58 39.33 39.44 41.13 42.17 Contingency .79 .77 .82 .80 .76 .75 .78 .77 -mu - Elaboration, argunents, evidence and style 58 upsetting. That is, while the readability score indicate that the messages are fairly difficult to comprehend (the score indicate college level writing) because of the multisylabic words and long sentece, the contingency scores indicate that the long words and sentece are offset by the predictable patterns of nouns. Thus, it is reasonable to coclude that the message were moderately difficult to conprehend. Becker et a1. (1961) sugget that moderate conprelnension difficulty is optimal for commnnication effectivenes since low or high levels of difficulty might frustrate the message recipient. film: Wignfi. .e_l_gb9_r‘at_e. There are several measure which might tap the predisposition to elaborate (e.g. , measure of discriminative ability, ability to organize ideas, reasoning skills, scholarly aptitude). For the sake of consistecy, and potential conparisons, however, the need for cognition instrument employed by the Petty and Cacioppo reearch teans was selected as an indicant of the predisposition to elaborate. Since Cohen's (Cohen et al, 1955: Cohen, 1957) original measure of need for cognition is apparently no longer available, Cacioppo and Petty (1982: Cacioppo, Petty, & Kao, 1984) developed new long and short version measure of need for cognnition. The long version measure has 34 Likert-type scales headed by statenents decriptive of habitual cognitive effort (e.g., "I would prefer complex to simple problens" and "thinking is not my idea of fun"). A principal-conponents factor analysis of the scale revealed one primary factor which had an eigenvalue of 10.22 and Elaboration, argunents, evidece and style 59 acconnted for 30.2% of the variance in reponse. The reliability coefficients (Theta, which is a corrected Cronbach's Alpha) for the measure across fonr studie, with sanple from a variety of populations, averaged to .90. The discriminant validity for the instrument was denonstrated for college profesors vs. assenbly line workers and enjoyment of tasks with simple vs. conplex rule. The measure is not correlated with trait anxiety or social deirability but is correlated with measures of field independence, dognnatism (negatively), and subjects' recall of their college entrance aptitude exam score. The short version of the need for cognition scale (Cacioppo et a1, 1984) consists of 18 of the highet loading scale itens fromn the long version. The reliability (Theta) of the short version for a sanple of college students was .91. The correlation between the long version and the short version is quite high (r s .95) . The short version of the need for cognition measure was used in this study as an etimate of the predisposition to elaborate (see Appendix C). Elaboration: mutive effort. Cacioppo, Petty and Morris (1983) reported the measurenent of cognnitive effort in message procesing with two 9 point scale one of which was specific to the experimental task ("To what extent were yon working hard to evaluate the student's performance as a journalist?" and "How much effort did yon put into evaluating the conmunication?"). The reliability of the measure was not reported. In order to innprove upon the Cacioppo et a1. (1983) measure of cognitive effort in message procesing, measurenent items were constructed by consulting a theaurus to generate a list of eighteen phrase decriptive Elaboration, argnments, evidence and style 60 of cognitive efforts or a lack thereof. Six of the decriptive phrases were deleted due to redundancy, excesive ambiguity, or words which might be unfamiliar to the research participants. The renaining twelve phrase consisted of six phrases descriptive of cognitive effort and six phrases decriptive of a lack of cognitive effort. The phrase that were selected are: Attenpting to analyze the issue in the mesage; Not very attentive to the ideas; Deep in thought about the mesage; Uncocerned with the ideas: Extending a good deal of cognitive effort: Distracted by other thonghts not related to the message: Not really exerting yonr mind: Doing your best to think about what was written: Reflecting on the implications of the arguments: Resting yonr mind: Searchinng yonr mnind in reponse to the ideas; Taking it easy. Seven point agree-disagree Likert-type scale were placed after each of the phrase. The twelve phrases were randomnly selected for order of placement in the reearch questionnaire. The instrunent was conpleted with a heading which reads: "While reading the message were you:" (see Appedix C). Probapility of belief in a prmitipp. McGuire (e.g., 1960a) and Wyer (e.g., 1975) have typically measured the probability of belief in a proposition by preenting subjects with an eleven interval scale bonnded by expresions such as true-false, very probable-very improbable, or not at all likely-extenely likely. Wyer (e.g., 1975) base the assunption that the probability of belief scale is sound and "approximately interval with an origin at zero" on "the quantitative accuracy of [the probabilogical] equation . . . in previons studie" (p. 312). Neverthe- less, the reliability and validity of the measurenent technique are not typically discnnssed or reported in the literature. Elaboration, arguments, evidece and style 61 Fishbein (Fishbein & Raven, 1962: Fishbein & Ajzen, 1975: Ajzen & Fislnbein, 1980) has developed a set of senantic differential-type scale to measure belief (defined as "the probability dimension of a concept," Fistbein 8 Raven, 1962, p. 35) that have been denonstrated to be reliable (above .85) and valid (See Fishbein & Ajzen, 1975, p. 108, and Reynolds, 1978) . In order to capitalize on the reliability and to maintain consistecy in the quetionnaire structure, the Fisl'bein belief scale were enployed. The measure consists of four scale bonnded by the bipolar adjectives true—false, probable-improbable, likely-unlikely, correct- incorrect (see Appendix C). Detection of rw (DRE). Unlike the host of rescuing studie (see Reynolds & Burgoon, 1983) it is untenable in this invetigation to preent subjects with the arguments being studied in an objective tet fornnat. Such an effort might pronpt the subjects to detect errors that were not detected while procesing the persuasive message. The work coducted in testing the probabilogical equations, however, sugget a viable approach to the measurenent of reasoninng error detection. If the first term of the probabilogical equations is an adequate measure of the rationality of an argunnent held by a person, which it appears to be, then the measurenent of the constituent parts of the equation can be altered in order to assess recipients' perceptions of the rationality of an argunent preented in a message. That is, the multiplication of the etimate of the sonrce's belief in the propositions of an invalid message argument (pgApsB) should not equal the etimate for the sonrce's belief in the coclusion (pSC) . Thus, it follows Elaboration, arguments, evidence and style 62 theoretically that the sum of the absolute value of the difference between (psApsB) and (pSC) shonld be an index of the detection of reasoning errors. Thus: DRE =2 HpsApsB) - (PsCH Equation 5 The probability of belief scale were used to measure the relevant conponents of the index of rescuing errors for each message (see Appedix C for the measurenent and Table 2, Appedix D, for the specific argnment premise) . Evidence eyalpa. ion. The measu renent of evidece evaluation was accomplished by means of a twelve item instrument with seven point Likert- type scale with stimulus statenents derived from discussions of the traditional tests of evidence (e.g., Miller, 1966: McCroskey & Wheeles, 1976) . The instrunent was headed by the phrase "Midge preented in the message was:" The evidence evaluation measurenent itens were: Inconsistent with the arguments beinng advanced: Current to the issue today: anfficient to prove the points being supported: Not typical of what is known abont this problem: Irrelevant to the coclusions drawn in the message; Not clear and understandable: Documented well enough so that any reader could look it up: Contained clear and understandable statistical information: Taken as a whole, supported the point being made; Came from experts on the topic; Came from persons who were biased on the topic: Came from credible magazine or books. Five of the twelve itens were worded negatively in order to reflect the pole of the measurenent scale (see Appendix C). Elaboration, argunents, evidence and style 63 Message evglpation. The absence of a test of Petty and Cacioppo's (1981: CaciOppo, Petty, & Morris, 1983) contention that persuasion is mediated by the recipients' message evaluations is in part due to their lack of a measure of message evaluation which is informed by an expertise in conmunication. Cacioppo et a1. (1983) report the use of a "message effectiveness" measure which is conposed of 5 nine-interval scale on which the repodents are asked to report how "effective," "convincing," "conpelling," "well written," and "liked" the message (or the reasoning) is. The reliability of the instrument was not reported. Clearly, the conceptual difference between these measurenent itens indicate that the measure is problenatic. Mesage evaluation or message effectiveness is frequently mentioned in the commnnication literature but seldcmn measured. Annong the attenpts to measure message evaluation, most tend to focus more on source perceptions (e.g., Street & Brady, 1982: Johnnson & Buttny, 1982) or on delivery skills (e.g., Clevenger, 1964: Barker, Kibler, & Geter, 1968: Freimuth, 1976). Burgoon (1975) enployed scale to measure perceptions of how well the message was supported and message clarity. Bradac, Desnnond and Mnnrdock (1977) enployed message evaluation scale that asked for perceptions of message organization, difficulty, general effectiveness as well as attitude toward the topic. While these efforts share a conmon cocern for message evaluation, the foci seem to be quite divergent. me effort specific to invetigating message evaluation has been offered by Burgoon (1972) . Burgoon constructed message evaluation scale which were representative of the evaluative dimension reported by Osgood, Elaboration, arguments, evidece and style 64 Suci annd Tannenbaum (1957). Snnbjects' responses to the scale after receiving one of several messages advocating social change were factor analyzed. The reults revealed two factors: Message evaluation and message interet. (A third factor also energed which was specific to militancy which Burgoon had intentionally manipulated in the study.) Since the degree to which a message is intereting was not an immediate concern in this invetigation, the interet factor was not measured. The message evaluation scale developed by Burgoon (1972) are seven interval senantic differential type scale with the bipolar adjective: Accurate-inaccurate; good-bad: believable-unbelievable: wise-foolish: responsible-irreponsible; acceptable-unacceptable. In order to facilitate conparison with other studie that have used measure of message evaluation (partionlarly Cacioppo, Petty and Morris, 1983) , the following bipolar adjective repreentative of thee measure were added to the instrument: Powerful-weak: mild-strong: convincing-unconvincing: effective-ineffective; well written-poorly written; compelling- unconpelling; organized-disorganized: clear-unclear: supported- unsupported: logical-illogical: rational-irrational: sound-unsound: valid-invalid: well reasoned-poorly reasoned. Attitude £01781“ 6 thejg. McGuire (e.g., 1960a) measured pD (the desireility of the argunnent coclusion) by asking subjects to repond to each proposition on a five-interval scale bonnded by the expresions; very deirable-very undesirable. Since it is being argued the pD is conceptually and Operationally similar to standard measure of attitudes and since a standard measure of attitude is more likely to be valid, Elaboration, argunents, evidece and style 65 reliable and, thus, offer a greater opportunity for conparison to other persuasion studie than a single deirability scale, a standard measure of attitude was enployed. The measure consists of six senantic differential-type scale bounded by the bipolar adjectives good-bad, foolish-wise, deirableundesirable, harmful-beneficial, pleasant- unpleasant, and worthles-valuable. Miller (1978) enployed a similar set of scale for reponses to message similar to those in this study and reports a reliability coefficient of .93. W. In order to measure the subjects' expectancy reactions to the message, four of the five senantic differential-type scale enployed by Miller (1978) were adopted. (The fifth scale, likely-unlikely, was deleted to avoid repetition with the belief measure.) The scale are bonded by the bipolar adjective predictable-unpredictable, expected- unnexpected, surprising-not surprising, anticipated—unanticipated. Miller (1978) reports a reliability coefficient of .76 for the expectancy scale. Leagues Snbject recruitment for participation in the study was acconplished by a single page handout that explained the need for participants in the study, described the study as requiring the reading of a message and repodinng to a questionnaire, mentioned the possibility of extra credit, and listed the available times for participation. The subjects signed up for participation on the sane sheet and were provided with a reminder slip on which they were instructed to mark down the time and place that they had signed up for. The recruitment sheets were handed out and collected during regular class honrs by the instructors or the reearcher. In addition, a sign up sheet was posted on a bulletin board in a hallway Elaboration, argnments, evidece and style 66 where students could continue to sign up after the initial recruitment process and instructors were asked to frequently remind students of the reearch opportunity. The research periods were scheduled over a two week time period in late Novenber. From 12 to 35 subjects participated in each of the reearch sesions. The research booklets were randomly distributed to the subjects after they were seated in the reearch room. The first page of the research booklet provided an introduction to the study and instructions on participating in the study. The introduction to the study explained that the subjects were participating in a test of several different messages which were beinng considered for subsequent research and that it was, therefore, important that they read the message as if they were reading it in a magazine or the opinion section of a newspaper. The instructions directed the subjects to follow uniform procedure while participating in the study. The subjects then read the version of the message contained in the quetionnaire booklet. After reading the message, the subjects conpleted the quetionnaire which contained measure for belief in the theis of the message, attitude towards the theis in the message, message evaluation, language expectancy violations, assesnnents of the sonrce's belief in each of the argument prOpositions, evidece evaluation, message elaboration, and need for cognition (See Appendix C). Following conpletion of the measurenent instrunents, the subjects were handed a debriefing quetionnaire on their knowledge of the reearch purpose. None of the subjects had an accurate interpretation of the research purpose. Following the debriefing quetionnaire, the snbjects were debriefed on the study and thanked for their participation. Elaboration, argunents, evidece and style 67 RMTS Measurenent model checks Prior to the tet of the proposed model an examination of the measurement model was coducted. The reliabilitie (Crobach's alpha) for each of the measured varionle were: Need for cognition, .86: elaboration, .84; expectancy, .85: attitude, .93: evidece evaluation, .82: detection of reasoning error (DRE), .23: message evaluation, .94: belief, .89. With the exception of the measurenent of DRE, the measurenent reliabilitie were quite high. The low reliability for the measurenent of DRE was proned first by examination of the individual reliabilitie for the measurements of the source's belief in the premise and coclusions of the argunents. The average reliability coefficient for the eight premise and fonr coclusion measurewas .97witharangebetween .94and .98. Itisworthnotethat the relianility analyse of thee measure indicated low scale variance. The secod probe of DRE measurement involved a conparison of DRE scores (as conputed by equation 5) for each of the fonr argunents in the message. The reults oftheprobe ofeachofthearguments (seeTable2 and Table 3, Appedix D). indicated that the first three argunnent validity manipulations produced significant differece in the DRE scores. For each of the first three arguments, the average DRE score for the invalid arguments was nearly twice the score of the average DRE score for the Elaboration, arguments, evidece and style 68 valid argunents. Consistent with the average DRE scores, the objects' probability etimate of the sonrce's belief in the coclusions of each of the first three arguments were higher for the valid arguments (X -= .71) than for the invalid argument (x a .55). For the fonrth argunent, the subjects etimated the sonrce's belief in the coclusion to be higher than logic would permit for both the invalid and the valid versions of the argument and, also, did not differ on the DRE scores. It appears, therefore, with the exception of the fonrth argument, that the manipulation of argument validity was successful and that the DRE equation (5) was sensitive to the manipulation. The third probe of the DRE score innvolved the examination of scattergranns of the DRE scores between each of the first three arguments in the message for the valid and invalid coditions. The scattergrans provided two insights into the low reliability of the DRE measure. First, approximately one-half of the objects consistently had low DRE score across the arguments. (A low DRE score would indicate that the source's belief in the coclusion was logically consistent with the sonrce's beliefs in the argunnnent prennise.) As expected, fewer objects in the invalid argument (47%) tlnann valid argunent (69%) coditions had consistently low DRE score across the argunents. Secod, on each scattergram, approximately one-third of the objects had high DRE scores on one argunent with low DRE score on the other argunent. This pattern was much more pronounced in the invalid (41%) argunent coditions than in the valid (28%) argument coditions. Therefore, it appears that the low reliability for the DRE measure is primarily the reult of inconsistent Elaboration, argunents, evidece and style 69 asseenents of the arguments and partially the reult of innsufficient variation prodnnced by the validity manipulations. It must be stressed, however, that the patterns of DRE score are far from random and that the validity manipulation did produce significant differece in the DRE score. An examination of the strength of evidece manipulation (strong vs. weak) via one-way AMA revealed a significant affect on evidece evaluation (a - 6.473, df - l/239, p -= .0116, onegaz - .03). The size of the effect, however, is somewhat disappointing. An examination of the stylistic manipulation (alliteration vs. no alliteration) via one-way ANNA revealed a nonsignificant affect on language expectancy (F -= 0.361, df - 1/239, 1H3). This lack of success with the style manipulation cast dobt upon the ability to tet the style route of the proposed model with the current data. Given the controversy in the literature over the relationship between beliefs and attitude, the belief and attitude scale were factor analyzed in order to examine the factorial structure of the scale. Fistbein and Raven (1962) codncted a factor analytic study on a similar set of scale for cocepts on which there is some controversy tout the actual existece of the phenonena (e.g., extrasensory perception) and fond two clear orthogonal factors: one for belief and the other for attitude. In this study the belief and attitude cocept of interet (the legalization of heroin) is more of a policy onetion than an existence question and, thus, provide grouds for a reassessment of the factor structure for the belief and attitude scale. Elaboration, arguments, evidence and style 70 A principal conponents factor analysis with varimx rotation revealed that the belief and attitude scale of the legalization of heroin topic fornned a unidimensional structure (see Table 4, Appendix D). A reexamination of the factor structure with an oblique rotation, however, revealed a two factor structure. Three observations are evident concerning the belief factor that emerged in the analysis with the obliqne rotation. First, the eigenvalue for the belief factor is just below 1.0. Sons reearchers wonld consider such an eigenvalue grounds for rejecting the factor. Second, the correlation between the belief factor and the attitude factor is . 73 which, of conrse, inndicate that the factors are far fron the orthogonal factors reported by Fisrbein and Raven (1962) . It shonld be noted, however, that the pattern of factor loadings evinnce sone confidence in a two dimensional structure. 'nnird, the factor loadings for the correct-incorrect and true-false belief scale are split between the belief and attitude factors and, thus, may not conceptually match "the probdnility dimnennsion of a concept' (Fistbein & Raven, 1962, p. 35) definition of belief. an the basis of the factor analysis of the belief and attitude scale, it was concluded that a conparison of beliefs and attitude separated annd combined would be justified. Tet of the & 'lhe path coefficients, sun of squared errors annd chi-square tet for the full model are preented in Figure 4. As wonld be expected given the measurenent problems discussed above, the chi-square tet inndicate that the pattern of correlations are different from those predicted by the model. Inshort, thefullmnodelwasnotsupportedbythedata. An Elaboration, arguments, evidence annd style 71 Figur Resultsoftheetetofthe preposed model" V -0 29 \61) -00‘ (-004) 015 (017) Sun of squared errors - .819 x2-198.9,d£-45,p<.on. Indicatethatthemodeldoenotfitthe Where: A - attitude ax Inexpectation B - belief -message evaluation DRE - Detection of reasoning error Ncog' -Need for cognnition E - elaboration -mesage style - evidence evaluation V - Validity e - evidence strength *Coefficients corrected for attemtion appear in parentheses. Elaboration, arguments, evidence annd style 72 examination of the error matrix for this analysis revealed that a large portion (61%) of the error was due to the path specifications for attitude. In particular, much of this error stems from the attitude-belief relationship. There are mmmerons possible explanations for this error (such as the lack of strength for the style manipulation or the unidimensional structure of belief annd attitude scale) but: ultimately, this finding evinnce the conclusion that style-expectancy- attitude ronte in the proposed model must be deleted prior to further exploration of the model. An examination of the path coefficients in Figure 4 shaws that the paths between the strength of evidence manipulation and evidence evaluation (.15), mesage elaboration and DRE (-.12), annd between massage elaboration annd evidence evaluation (.10) were significant bnnt not particularly strong. ‘nne path between DRE annd message evaluation (.04) was not significant. Correction for attenuation due to measurenent error improve the path coefficients for the message elaboration-DRE annd DRE-message evaluation paths but doe not improve the path coefficients for the message elaboration-evidence evaluation or for the strength of evidence manipulation-evidence evaluation paths. The improvements are largely dnne to the low reliability for the measurenent of DRE (note that the path from the manipulation of argument validity to DRE jumps from -.29 to -.61 when the correlations are corrected for attenuation). An exanminnation of the directions of the path coefficients revealed that the massage elaboration-DRE annd the DRE-mesage evaluation paths were in directions opposite of that predicted. Given the low reliability of Elaboration, arguments, evidence annd style 73 DRE annd the small magnitude of thee two path coefficients, interpretation of thee two reverse directional coefficients in certainly problenatic and may, perhaps, be meaningles. All of the other paths were in the direction predicted. the most obvions explanation for the negative path coefficient between message elaboration annd DRE is that there may be an interaction between argunent validity and message elaboration on DRE such that for invalid arguments as message elaboration increase, DRE inncrease while for valid arguments, as message elaboration increase, DRE decrease. Il‘his potential interaction effect was teted by means of a 2 x 4 validity by mesage elaboration (quartile split) MINA for DRE. The main effect for validity on man was signnificant (g; - 22.86, or = 1/232, p - .001, omegaz . .095). The main effect for message elaboration on DRE (2 . 1.38, df - 3/232, nan) annd the interaction effect for validity annd message elaboration on DRE (1.: - 1.76, df 8 3/232, M) were not significant. 'lhe pattern of mneans, however, inndicate that for valid arguments, DRE slight decrease as eldnoration increase while for invalid arguments, the message elaboration-DRE relationship forms an S shaped curve where DRE decrease, increase annd then decrease again as elaboration increase. Perhaps with a more potent argunent validity manipulation, greater power and a more reliable DRE inndex, the interaction may have been signnificant. For this data set, unfortunately, the mesage elaboration-DRE path appears to be uninterpretable. Similarly, the nonsignificant path between DRE annd message evaluation is also uninterpretable. Elaboration, arguments, evidence annd style 74 Figure 5 Reults of the tet*of the evidence ronte .31 (.36) .10 (.11) .70 (.80) .51 (.56) M .. E -EE-—————..HE fl -v---- _ -43 (.17) Sum of squared errors a .041 x2 - 7.73, df - 10, p < .01. Indicate that the data do not differ from the model's predictions. Where: 8 - belief - evidence strength E - elaboration -mesage evaluation £3 a evidence evaluation Ncog' -Need for cognition *When belief annd attitude are conbined, the coefficient between ME and Belief/Attitude is .57 (.61). All other coefficients remain approxintely thesanmeandtheoombinationdoenotsnbstantiallyimprovethefit between the model annd the data. Coefficients corrected for attenuation appear in parenthee. Elaboration, argunents, evidence and style 75 Elm annalm of the model Evidence ronte isolated. Due to the caution dictated by the error reulting from low reliability of the measurement of detection of reasoninng error and the connter—to-expected mnesage elaboration-DRE and DRE-mesage evaluation paths, the evidence ronte of the model was examined in isolation. The path coefficients, sum of squared errors, and chi-square tet for the ronte are preented in Figure 5. With the exception of the path from message elaboration to evidence evaluation, the paths along the ronte are all fairly strong. 'nne chi-square tets (for raw and corrected correlation matrice) indicate that the data do not differ from those the model would predict for the ronte. Indeed, the sunm of squared errors of . 041 (. 063 followinng correction for attenuation) is impresively low. DISCIJSSICN The stronget and most significant finding from this study is that the effect of evidence evaluation in persuasion is mediated by the receivers ' evaluation of the mesage. 'Ihis finding, if supported by subsequent teting, may well provide insight into a host of null effect and small effect evidence studie (see Reynolds & Burgoon, 1983). In addition, there is sonne indication that receivers ' evidence evaluations are a funnction of habitual cognnitive efforts as well as actual cognnitive elworation on the message. Limited support was garnered from this study for the conclusion that the validity of the arguments in a message are detectable by receivers ' assessments of the sonrce's belief in the argument premise and Elaboration, argunments, evidence and style 76 conclusions. The predictions that message elaboration results in the inncreased detections of reasoning errors ad that the detection of reasoninng errors will reduce message evaluation were not supported. The reults of this stndy are uninformative wont the role of message style (as it was conceptualized) in persuasion. There was no support from this study for the idea that message style which positively violate expectations will foster more favorable attitude toward the message theis. There was also no support for the notion that a message style innduced attitude towards a message theis will affect the elaboration on the message or the evaluation of the message. Taken together, thee findings support two major tenets of the cognitive reponse approach to persuasion (e.g., Greenwald, 1968: Petty, (btrom, & Brock, 1981: Petty & Cacioppo, 1981). First, the impact of message variable on persuasion is mediated by the receivers' evaluation of the persuasive message. Secod, receivers' message evaluations are affected by their cognnitive hanits and their cognnitive eldnoration of the messagecontent. 'l'hesupportfortheetmtenetsmnnst, however, be qualified by the strong critical comentary and suggestions for necesary revisions in the elaboration likelihood model of cognnitive reponse provided earlier (see also Stiff, 1986) . Perhaps more important, the conclusion concerninng the role of message evaluations in persuasion was reached here withont the pitfalls of the thought-listing technique (see Miller a Baron, 1973, and Greenwald, 1981) . 'lhe most glarinng difficulty with this study is the failure to satisfactorily manipulate message style. apparently, the alliteration Elaboration, arguments, evidence ad style 77 effects attributed to fanous speeches (e.g., King's 'I have a drenmnn‘I speech) are either much more obvionsly a part of the message: depedent on other message, preentational or sonrce variable: or simply better example of alliteration than those enployed in this study. Then again, the effects of alliteration in famous speeche may simply have been over etimated by rhetorical critics. The onetion remains an enpirical issue. A secod difficulty with this study occurred with the relationship between attitude and beliefs. Previons reearch led to the speculation that message variable might affect a receiver's attitude towards the message thesis which would, in turn, affect the procesinng of the message. Of conrse, such a statement wonld suggest a longitudinal study with the capacity to measure attitude derived from the message as separate fromn the receivers ' initial attitude or their post-message attitude on the message theis. Perhaps a better etimate of the expected effect for attitude wonld have been obtained if the stylistic manipulation had been succesful ad premesage attitude were measured and partialled ont of the derived attitude. an the other had, the evidence provided by this study teds to support the conclusion that attitnde ad beliefs are probably not conceptually distinnct or, at a minimum, belief is not yet accnnrately operationalized. It may also be that the relationship between attitude ad beliefs varies with the cognitive concept under consideration. Since Fislbein and Raven (1962) report an orthogonal relationship between belief ad attitude for quetions of existence (or fact) while factor analysis reported herein indicate a unidemensional (or, at bet an. oblique) relationship for Elaboration, arguments, evidence ad style 78 questions of policy (shonld heroin be legalized), than perhaps there is a variation in the relationship between beliefs ad attitude between concepts that wonld be classified as quetions of fact, policy or value. A third difficulty with this study concerns the measurenent of DRE. While the adaptation of the McGuire (1960a) probabilogical formla (conpare equations 1 and 5) to asses recipients' perceptions of the rationality of an argument preented in a message did discriminate between valid and invalid argunent manipulations in this study, the value for each argument apparently do not snmm together efficiently into a single DRE score. This difficulty with the measurement of DRE seems to be largely thereultofinconsistentasseenentsoftheargnmnentsinthemesage. Thisinnconsistencymaybetheconsequenceofageneral inanilitytoasses multiple arguments. Ray, Carranza ad Reynolds (1986) nnote a similar lack of consistent argument decoding with disjunctive permitted choice utterance and attribute the effect to contextual difference in the arguments. Sincethearguments inthisstudyarecontextuallysimilar, however, thelackofconsistentargnmmentdecodinginbothstudiemay perhaps be a funnction of some other information procesing pattern. Two alternative explanations for the argnmnent decoding inconsistencie are: (1) An equity principle; where, after message repodents unfavorably asses the validity of one argunent, there is a conpulsion to liberally assess the snbsequent argument or: (2) a reponse bias where tet-wise studentparticipantsareunlikelytoaoceptalloftheargunentsina message as being valid or invalid (i.e., all 'true' or all 'falBe'). Since the previous research with the probabilogical equations (e.g. , Elaboration, arguments, evidence ad style 79 moire, 1960a: Wyer & Golanerg, 1970: Rample, 1979b) has focused upon only one argument at a time, that literature offers little insight into this problem. To conplicate matters further, given the positive, yet small correlation between the DRE score and message evaluation ad the negative correlation between message elaboration ad DRE, there is some doubt that the DRE score actually represents mesage recipients ' perceptions that the argunents are valid or invalid. anvionsly, a great dealofwork renainetobedoneontheetdnlislnentofaDREmeasurethat is both reliable and valid. The unnpredicted negative direction of the message elanoration-DRE path renains perplexing. While the potential for an argument validity by message elaboration interaction on DRE renains plausible, there are nanny other rival explanations for this relationship. For example, it may be that those who elaborate more on message are more likely to I'correct" errors in others' reasoning. Perhaps, also, those who elanorate more are distracted by their focus upon the entire message ad, thus, are less likely to detect reasoning errors. Ewen more plausible, those who elaborate are more likely to recall the oconrrence of statements in a message and, therefore, are more likely to overestimate the sonrce's belief in the statements which would reult in an artificially low DRE score. (The higher the etimate for all three propositions in an argument, the lower the DRE score.) The low positive path coefficient between DRE and message evaluation renains sonewhat disturbing. While the low reliability for the DRE measure might mitigate any concern here until a better DRE measure is Elaboration, argunents , evidence ad style 80 arrived at, the strength of the path from the validity manipulation to DRE and the positive coefficient for DRE to message evaluation (the relationship was expected to be negative) suggets that more is amiss with the DRE-message evaluation path than just the low reliability of DRE. A scattergram analysis of DRE with message evaluation revealed an apparent curvilinear relationship such that when DRE was either low or extrenely high, message evaluation was low. At moderately high levels of DRE, message evaluation was high. If this curvilinear relationship actually exists, then it would indicate that a number of the message receivers assessed the argument premise and conclusions in a manner consistent with logical validity, but still evaluated the message negatively. Modifications and flies for future reearch The most evident modification necesary before this program of reearch can proceed is to etdnlish a relianle ad valid measure of the detection of reasoning errors in message. This improvement may be acconplished with a simple modification of the DRE formula that take innto account the equity principle or reponse bias that appears to be operating in the assessments of multiple arguments. Further consideration and refth of Wyer's (e.g., Wyer n. Hartwick, 1980) and Hample's (1978) work with the pK canponent of the McGuire formula mnight help with the measurenent of the detection of reasoning errors. It may also be that a simple direct measure of message recipient's perceptions of the logical validity of message arguments conld be developed. The secod modification needed for this reearch program to continue is the need to replicate this study with an improved manipulation of Elaboration, arguments, evidence ad style 81 message style. As discussed under the selection of the mannipulation enployed in this invetigation, the issue of the relevance of stylistic variable to message content procesing vs. 'peripheral ronte“ (Petty ad Cacioppo, 1981) procesing renains to be explored in depth. The need to clarify the relationship between belief and attitude in message procesing is certainly a object for further research that this study points to. Related to the belief ad attitude issue is the need for refinenent in the measurenent of beliefs. Clearly, much of the probabilogical, reasoned action, and elaboration likelihood literature turns on the belief and attitude relationship. If the relationship between belief and attitude varie with the type of proposition being assesed, that findingwillhavewideoonsequenceforalargebodyof persuasion theory. An invetigation of how the relationship between beliefs and attitude might vary with the cognitive concepts nnnder consideration is currently being coducted by this author. Stiff 's (1986) analysis of the elaboration likelihood model ad advanncement of Kahnenan's (1973) elastic capacity model of information procesing might offer fertile gronnds for extensions ad refinements of the message procesing model proposed in this study. In particular, Stiff 's work with the elastic capacity model raise a number of gnestions concerning the effects of need for cognition upon elastic capacity and information procesing. If people who are high in the need for cognition are operating at high capacity, are they then operating in a single mode of procesing (as opposed to dual mode procesing) , or do they simply have a higher capacity to call upon? Donbtlesly, the probeilogical, reasoned Elaboration, arguments, evidence and style 82 action, elaboration likelihood ad the message procesing model explored in this study conld all benefit from a syntheis within an information procesing model such as the Kahnenman's (1973) elastic capacity model. Following and concurrently with the modifications ad extension of the reearch reported in this report, there is a need to replicate and exted the findings of this study with a variety of message, persuasive contextsandsnbjects. Inparticular there is aneedtomove ont ofthe 'editorial" issue franne enployed in this study and into more conmon interpersonal influence situation (see Miller a Burgoon, 1978) . Finally, a nunber of new and revised measurenent instruments were enployed in this invetigation. Aside fromn the need to reolve diffiolltie with the measurement of the detection of reasoning error, the evidence evaluation and message elaboration scale require validation. In addition, the message evaluation scale derived largely from Burgoon (1972) require further exploration and teting. Elaboration, argnmments, evidence, and style 83 Reference Ajzen, I., & Fisl'bein, M. (1980). Urder_s_t_a_nn§_ilg attitude and predictyg mm. Bnglewood Cliffs, nu: Prentice-Ball. Arnold, C., & Bowers, J. W. (Eds.). (1984). Bandnook of rhetorical and W- 3081:0313 Allyn and 3300!!- Barker, 1.. L., Kibler, R. J., & Geter, R. W. (1968). Two invetigation of the relationhip among selected ratings of speech effectiveness and carprehensim- W. .05.. 400-406. Barrett, W. (.1958) . Irrational min: A M lg existential mum. Garden City, NY: Donbleday Anchor Books. Basehart, J. R. (1971). Message opinionation ad approval-depedence as determinants of receiver attitude change and recall. m m. .33: 302-310- Becker, S. L. (1963). Research on emotional and logical proofs. Southern W: 2.8: 198-207. Becker, S. W., Bavelas, A., a Braden, M. (1961). An index to measure contingency of eglish sentence. W, _4_, 139-145. Bettinghaus, E., Miller, G. R. , & Steinfatt, T'. (1970). Sonrce evaluation, syllogistic content, and judgments of logical validity by high- and low-dogmatic persons. Journal of Personality and Socig m, 2.0.: 150-159. Blair, B. (1965). Lecture m rhetoric 9g _b_e_lle lettges. (B. F. Harding, Ed.). Carbodale, 111.: Southern Illinois University Pres. Boshenskie, I. 14. (1956/1961). A giggly of fog lgglc (I. Thomas, Tran.) . Notre Dame: university of Notre Dame Pres. First published as Formale lgg . Bostron, R. N., Basehart, J. R., a Rossiter, C. 14., Jr. (1973). The effect of three type of profane language in persuasive communication. gongngl of Conmunication, 23, 461-475. Bostronn, R., 5 Tucker, R. (1969). Evidence, personality, and attitude dunge- W: 3.6.: 22-27- Bowers, J. W., & Osborn, M. (1966). The attitudinal effects of selected type of concluding metaphors and persuasive speeche. M W: 23: 147-155- Elaboration, arguments, evidence, ad style 84 Bradac, J., Bowers, J., a Conrtwright, J. (1979). Three language variable in connunication reearch: Intenity, immediacy, ad diversity. amnan Comrunication mm, 1, 257-269. Bradac, J., Desmond, R., a mrdock, J. (1977). Diversity and density: Lexically determined evaluative and informational consequence of linguistic conplexity. Communication Mating, 14, 273-283. Bradac, J., Sandell, R., a L. Wenner, (1979). The phenonenology of evidence: Information-sonrce utility in decision making. Communnicatlon my: 2.7: 3545: Brumett, B. (1981). Gastronomic reference, synnedoche, and political muses: WM: .61: 138-145- Burgoon, J. K. (1975). Conflicting information, attitude and message variable as predictors of learning and persuasion. m W: l: 133-144. Burgoon, H. (1970) . Prior attitude and language intenity as predictors of message style and attitude change following connnterattitudinal commnication behavior. Unpublished doctoral dissertation, Michigan State University. Burgoon, M. (1972) . A factor-analytic examination of message advocating 806151 change. W: .32: 290-295- Burgoon, M. (1983). Argument from Aristotle to analysis of variance: A modet reinterpretation. Paper preented at the International Conference on social psychology and laguage, university of Bristol, England. Burgoon, 14., & Bnnrgoon, J. K. (1975). Mesage strategie in influence attenpts. InG. Bannenan EW. Mcawen (Eds.), W ad M12! (pp. 149-165). Reading, MA: Addison-Wely. Burgoon, 14., a Chase, 1.. J. (1973). The effects of differential linguistic patternn in message attenpting to induce reistannce to persuasion. W: 3.0.: 14- Burgoon, 14., Cohen, 11., Montgomery, C., 5 Miller, M. (1978). An enpirical tet of a model of resistance to persuasion. W mglp é, 27.39. Burgoon, 11., Jones, S. B., n. Stewart, D. (1975). Toward a message-centered theory of persuasion: Three enpirical invetigation of language intensity: W: l: 240-256- Burgoon, 11., 8 Miller, G. R. (1971). Prior attitude and laguage intensity as predictors of message style ad attitude change following counterattitudinnal advocacy. gonrnal of Personality and figitl 25191391931: 29: 246-253- Elaboration, arguments, evidence, and style 85 Burgoon, M. & Miller, G. R. (1985). Ann expectanncy interpretation of laguage and persuasion. In B. Gile & R. St.Clair (Eds.), Recent advances in 1mm connunication fl social Mlggy. New York: Erlbanmm. Burke, K. (1969). A grammar of motive and a rhetoric 9g motlves. Berkeley: The University of California Pres. Cacioppo, J. T'., Barkins, S. G., 8: Petty, R. E. (1981). The nature of attitude ad cognitive reponse and their relationships to behavior. In R. Petty: 1'. Ostrom. s T. Brock (368.): W Persuasion (pp. 31-54). Billsdale, NJ: Erlbaunm. CaciOPPO: J . T., 8: Petty, R. E. (1979). Attitude and cognitive repone: An electrophysiological approach- W MM! .12: 2181’2199- Cacioppo, J. T‘., 8: Petty, R. E. (1980a). Persuasivenes of commercials is affected by exposure frequency ad conmunication cogency: A theoretical and enpirical analysis. In J. H. Leigh, C. R. Martin (Eds.), Gnrrent lssue and teeagct in manta (pp. 97-122). Ann Arbor : University of Michigan. Cacioppo, J. T. , 5: Petty, R. E. (1980b). Sex difference in influenceanility: Toward wecifying the underlying proceses. Personality and Soci_a_l may Elletin, _6_, 651-656. CaciOPpO: J. T., 8 Petty, R. E. (1981). Social psychological procedure for cognitive reponse assessment: The thonght-listingtechnique. In T.V. Merluzzi, C. R. Glass, a 14. Genest (Eds.), Mtive Assesment (pp. 309-342) . New York: Guilford Pres. Cacioppo, J. T., 5 Petty, R. E. (1982). The need for cognition. 1%; Per i 1 ,4__2_, 116-131. Cacioppo, J. T'., Petty, R. E., a Kao, C. F. (1984). The efficient easement of need for cognition: MW: e, 306-307. Cacioppo, J. T., Petty, R. E., a Morris, K. J. (1983). Effects of need for cognition on message evaluation, recall, and persuasion. Meal 9_f Personali and 1 , _4_5, 805-818. Cacioppo, J. T'., Petty, R. E., a Sidera, J. (1982). The effects of a salient self-schema on the evaluation of proattitudinal editorials: Top-down versus bottom-up message procesing. W aggrimental §o_gig_ mmlgy, lg, 324-338. Cappella, J. N. (1980). Structural equation modeling: An introduction. In 1’: Mange & J- N- Cappella (Eds-1: WW Met; on reeatgt (pp. 57-109) . New York: Academic Pres. Elaboration, argunents, evidence, and style 86 Cialdini, R. B., 5 Petty, R. E. (1981). Anticipatory opinion effects. In R. E. Petty, T'. 14. Ostrom, a T. C. Brock (Eds.), W in grsuasion (pp. 217-236). Billsdale, DU: Erlbaum. Clnapnan, L. J., & Chapman, J. P. (1959). Atmosphere effect re-examined. MW: L8: 220-226. Clark, H. B., & Begun, J. S. (1971). The semantics of sentence objects. W: 1.4: 34-36: Cleveger, T'. (1964) . Influence of scale conplexity on the reliablitiy of ratings of general effectiveness in public speaking. m 33293218.: 11.: 153-156- Cohen, A. R. (1957) . Need for cognition and order of communication as determinants of opinion change. In C. I. Bovland (Ed.), W geentation in granaslon (pp. 79-97) . New Haven: Yale University Pres. Cohen, A. R., Stotland, E., a Wolfe, D. 11. (1955). An experimental investigatim of need for commitm- W Engineer: él: 291-294: Cohen, J., 8: Cohen, P. (1975). lied ti 1e r resi corr on W: Hilledale: NJ: Brlbatm- Corbett, E. (1971). Classical rhetoric tor the firm gm (2nd ed.). New York : Oxford Unniversity Pres. Dubin, R. (1978). W (rev. ed.). New York: The Free Pres. Edwards, W. (1954). The theory of decision making. W M! 19: 72-83- Eiseberg, E. 14. (1984) . Ambiguity as strategy in organizational communication- W: .51: 227-242. Evan, J. St. B. T. (1980). olrrent issue in the psychology of reasoning. WM: 11: 227-239- Feather, N. T. (1964). Acceptance and rejection of arguments in relation to attitude stregth, critical ability, and intolerance of inconsistency. mg of Abnofl and Socitl My, _6_9, 27-1360 Fishbein, 14. (1967) . A behavior theory approach to the relation between beliefs abont an object and the attitude toward the object. In 14. Fishbein (Ed.) , Readlms in attitude Mg and W (pp. 389—400) . New York: Wiley. Fislbein, 14., a Ajzen, I. (1975). Belief, attitude, intentlon and ion WW- Readim: HA: Addison Wesley. Elaboration, arguments, evidence, and style 87 Fishbein, 14., a Ajzen, I. (1981). Acceptannce, yielding, and impact: Cognitive procese in persuasion. In R. E. Petty, T. 14. Ostrom, & T. C. Brock (Eds.), C_O_Qltive rm in grsuasion (pp. 339—359). Hillsdale, NJ: Erlbaum. . Fishbein, 14., Ajzen, I., a McArdle, J. (1980). Chaging the behavior of alchoholics: Effect of persuasive conmunication. In I. Ajzen a 14. Fishbein (1980) . Understandll'g attitude and predicth socitl _b_e1_nay_i_o_r (pp. 217-242). Englewood Cliffs, NJ: Prentice-Hall. Fishbein, 14., 1: Raven, B. E. (1962). The AB scale: An operational definition of belief and attitude. W, l5, 35-44. Flehler, B., Ilardo, J., a Denoretcky, J. (1974). The influence of field depedence, speaker credibility set, and message documentation on evaluation of speaker and message credibility. W W: 32: 389-402- Fradsen, K. D., a Clenent, D. A. (1984). The fnunctions of hnman communication in informing: Communicating and procesing information. In C. Arnnold n. J. W. Bowers (Eds.), (1984). Banndnook of cal and W (pp. 338-399). Boston: Allyn and Bacon. Freimuth, V. (1976) . The effects of communication apprehension on communication effectivenes. WW, 2, 289-298. Greene, J. O. (1984) . Evaluating cognitive explanation of communication intramu- WM: 12, 241-254. Greenwald, A. G. (1968). Cognitive learning, cognitive rspone to persuasion, and attitude change. In A. G. Greenwald, T. C. Brock, ad T. 14. (ntrom (Eds.), l i oudat on t (pp. 361-388). New York: Academic Pres. Greenwald, A. G. (1981). In R. E. Petty, T. 14. Ostrom, & T. C. Brock (11:13.): We (pp- 127-1341- Hilladale: nu: Erlbanmm. Bample, D. (1977). Testing a model of value argumennt and evidence. Wicatim gmrggp fl, 106.120. mmple, D. (1978) . Predicting immediate change ad adherence to argument claims. Connmnnicatlon Mrm, 5;, 219-228. ample, D. (1979a) . A review of enpirical literature on logical procese. Paper preented at the annnual Speech Communication Association Convention, San Antonio, Texas, Novenber. mmple, D. (1972)) . Predicting belief and belief change using a cognitive theory of argunent and evidence Waste: 56.: 142-151. Elaboration, arguments, evidence, and style 88 Harte, T. B. (1971) Audience anility to apply tets of evidence. gonrnal gt t3 Agrican Fotenlc Associatlon, t, 109-115. Hawes, L. C. (1976). How writing is used in talk: A study of communicative logic in use- W: 3: 350-360- Henpel, C. G. (1965). am of scientific gplanation and 9_tb_e;_r esajg in inlgsfly of science. New York: The Free Pres. Henle, 14. (1955). Sone effects of motivational procese on cognition. W: Q: 423-432- Henle, 14. (1962) . an the relation between logic and thinking. Mtgical Review, §_9_, 366-378. Himmnelfarb, S., a Eagly, A. H. (1974). Orientation to the study of attitude and their change. In S. Hemmelfarb a A. H. Fagly (Eds.), Readllg' s in attitude ttn_a_nge (pp. 2-49). New York: Wiley. mater, J. E., a Hamilton, 14. A. (1986). A least Square static path analysis program [Conputer program]. East Lanning, 14I: Dept. of Psychology: Michigan State university. Insko, C. A. (1967). Theotie of attltude gm . Eglewood Cliffs, N.J.: Prentice-Hall. Jackson, S. (1982). Two models of syllogistic reasoning: Ann enpirical conparism- W: 5.9.: 205-213. Janis, I. L., & Frick, F. (1943). The relationhip between attitnnde toward conclusions and errors in judging logical validity of syllogisme. golgn_a_l of ggrimntal gang-Iggy, 3;, 73-77. Janis, I. L., a Hovland, C. I. (1959). An overview of persuasibility reearch. In C. I. Hovland s: I. L. Janis (Eds.), ggrsonallty a ptrsuasibility (pp. 1-26) . New Haven: Yale University Pres. Janis: I. L., 5 Mann, L. (1977). MW MW- New York: The Free Press: Johnson, D. 14. (1968). Reasoning and logic. In D. L. Sills (Ed. ), - , _- ‘5 (V01. 3' mo 344-350). New York : MacMillan. J Johnson, F., & Buttnny, R. (1982). White listeners' reponse to 'souding Black" and 'sonnding White“: The effects of message content on judgments wont langnage. Communicatlon germ: 42, 33-49. Jordon, W. J. (1972). A reinforcement model of metaphor. m M: 3.2: 223-236- Joseph, 14. (1947). Shakme's use of the 3g of 1m . New York: Columbia University Pres . Elaboration, argnments, evidence, annd style 89 Kahnenan, D. (1973). Attention and effort. Bglewcod Cliffs, m: Prentice-Hall. Kaplan, A. (1964). The conduct of i - fo ioral 591m. Scranton, PA: Channdler Publishing Co. Kauffiran. c. (1981). Poetic as argmmt. WW: g2, 407-415. Kelman, H. C. (1961). Procese of opinion change. mblic minim W! E! 57-78- Kenny, D. A. (1979). Correlation and causality. New York: Wiley. Kline, J. A. (1971a). A Q—analysis of encoding behavior in the selection of evidence. Comuunication Mrm, 3.8, 190-197. Kline, J. A. (1971b) . chmatismn of the speaker and selector: fo evidence. Conmunication mrm, 38, 354-356. Lingle, H. H. & Ostron, '1'. H. (1981). Principle of memory annd cognniticn in attitude formation. In R. E. Petty, '1‘. M. “tron, 8 '1'. C. Brock was.) , Mtive pm in greasion (pp. 399-418) . Hillsdale, NJ: Erlbaum. Donginus (c. 100/ 1965). an the snblimne. In T'. S. Dorsch (Trans.), Wsnm. New York: Penguin Booka- Love, R. B., & Greenwald, A. C. (1978). Cognitive reponse to persuasion and mediators of opinion chame- W: 10.41.: 231.2410 Luchok, J., & McCroskey, J. C. (1978). The effect of quality of evidence an attitude and source credibilitY- W m. .12. 371-383- Lynch, M. D. (1970). Stylistic analysis. In P. Enmert a W. D. Brooks (Eds.): Wage; (pp. 315-342). Boston: Houghton Nifflin Co. McCroskey, J. C. (1967) . The effects of evidence in persuasive commi- cation. Wetern gong; of m Communigigg, 33, 190-199. McCroskey, J. C. (1969). A summary of experimental reearch on the effect of evidence in persuasive conmunicaticn. WM, 5.5, 169-176. McCroskey, J. C. (1970). The effects of evidence as an inhibitor of counter-persuasion. W, 31, 188-194. McCroskey, J. C. (1982). An rcduct on to r rical on: (4th ed.). Bglewcod Cliffs, NJ: Prentice—Hall. Elaboration, argunents, evidence, and style 90 McCroskey, J. C., & Conbs, W. H. (1969). The effects of the use of analogy on attitude change annd sonrce credibility. W, .12: 333-339. McCroskey, J. C., Young, '1‘. J., & Scott, M. D. (1972). Special reports: The effects of nessage sidednes and evidence on inoculation against connterpersuasion in small group comunication. Miss; m: 3&1 205-212. McCroskey, J. C., & Wheeles, L. C. (1976). W W. Boston: Allyn and Bacon. mire, W. J. (1960a) . A syllogistic analysis of cognitive relationships. In C. I. Hovlannd & H. J. Rosenberg (Eds.), W M (pp. 65-111). New Haven, 01': Yale University Pres. McGuire, W. J. (196(1)). Direct annd inndirect persuasive effects of dissonance practicing mnesage. W m £9. 354-358. McGuire, W. J. (1968). Personality and susceptibility to social influence. In E. F. Gorgatta a w. s. Lambert (Eds.), m and reearQ (pp. 1130-1187). Chicago: Rand NcNally. McGuire, W. J. (1969). The nature of attitude annd attitude change. In G. Lhflzey & IE- Arcnsan (Eds.), W (2nd- ed., Vol 3., pp. 136-314). Reading, MA: Addison Weley. McGuire, W. J. (1981a) . The probabilogical model of cognitive structure annd attitude change. In R. E. Petty, T'. N. Ostron, & T'. C. Brock (Eds.), mtive rw in greasion (pp. 291-307). Hillsdale, nu: Erlbanmn. McGuire, W. J. (1981b) . Theoretical foundations of canpaigns. In R. E. Rice and W- J- Paisley (Eds.), WM (99- 41-70) . Beverly Hills, CA: Sage. Miller, M. D. (1978) . The relationship of violations of receiver expectations and reistance to persuasion. Unpublished doctoral dissertation. University of Florida. Miller, M. D., & Burgoon, H. (1979). The relationship between violations of expectations annd the infliction of reistance to persuasion. amen Conuunication Reearc_h, _5, 301-313. Miller, G. R. (1966). Evidence annd argument. In G. R. Miller a T'. R. Nilsen (Eds.), WW (pp. 24-49). oucagos Scott, Foreman. Miller, G. R. (1969). Sons factors influencing judgments of logical validity of argunents: A reearch review. W M' g, 276-286. Elaboration, arguments , evidence, annd style 91 Miller, G. R., a Burgoon, N. (1973). New t9_cEn_ig_ne of mrsuasion. New York: Harper and Row. Miller, G. R., & Burgoon, H. (1978). Persuasion reearch: Review annd W- In B. D. lube-1mm). W1; (99- 29- 47). New Brunswick, no: Transactions books. Miller, G. R., Burgoon, N., & Burgoon, J. K. (1984). Comunicaticn and conpliance: An overview of the reearch on the proces of persuasion. In C. Arnold & J. W. Bowers (Eds.), Handbook of rhetorical and W (11). 400-474). Boston: Allyn annd Bacon. Miller, N., a Baron, R. S. (1973). an measuring counterarguing. M W: 3.: 101-118. Hinneo, P. (1983) A psycholinnguistic perspective on reasoning and A argumentation. Paper preented at International conference on social psychology and language, University of Bristol, England. Morgan, W. J., & Morton, J. T'. (1944). The distortion of syllogistic reasoning produced by personal convictions. W m. L0: 39-59- Nie, N. B., Hill, C. H., Jenkins, J. G., Steinbrenner, R., a Bent, D. H. (1975). Statistical e or soc science (2nd ed.). New York: Maw-Hill. Norton, R. W. (1978) . Foundations of a communicator style construct. gm Meade Reseagth .5, 99-112. Norton, R. W., & Pettegrew, L. S. (1979). Attentivenes as a style of cmnication: A structural analysis. MW. .26.: 13-26. Osgood, C. B., Suci, G. J., & Tannenbaum, P. H. (1957). The measurement W. Urbana, IL: University of Illinois Pres, 1957. Ostron, T‘. C. (1981) . Theoretical perspective in the analysis of cognitive reponse. In R. E. Petty, '1‘. ll. Ostron, & T'. C. Brock (Eds.), Mtive pm in grsuasion (pp. 283 to 290). Hillsdale, m: Erlbaunn. Perelman, Co: a Olbrechts-Tyteca: L. (1969)- W on arm (J. Wilkinson 5. P. Weaver, T'rans.). Notre Dame, IN: university of Notre Dame Pres. Petty, R. E. (1977). A cognitive reponse analysis of the tenporal persistence of attitude change innduced by persuasive conmunications. Unpublished doctoral dissertation, Ohio State University. Elaboration, argunents, evidence, and style 92 Petty, R. E., 8 Brock, T'. (1981). Thought diruption and persuasion: Assesing the validity of attitude change experiments. In R. E. Petty, 1‘. n. Ostran & 'r. c. Brock (Eds.)n MW (pp. 55-80). Hillsdale, NJ: Erlbaum. Petty, R. E., a Cacioppon J. T‘. (1977). Forwarning, cognitive reponding, and reistannce to persuasion. Journal of Personality fl Social m, 3_5, 645-655. Petty, R. E., & CaciOPpOn J. T‘. (1979). Issue involvenent can increee or decrease persuasion by enhancing mesage-relevant cognitive reponse. Jonrnal of Personality and Eial Malay, 31, l-19. Petty, R. B., & Cacioppo, J. '1'. (1981). Attitude rsuas on: C1 ic W- We: IA: m. C. Brown. Petty, R. E., CaciOppoy J. T'., 8 Goldnnan, R. (1981). Personal invcvennent as a determinant of argunent-based persuasion. W and Socia_l MM: .9; 847-855. Petty, R. E., Cacioppo, J. T'., & Heeacker, I4. (1981). The use of rhetorical quetions in persuasion: A cognnitive reponse annalysis. WW_ 43-2'440 Petty, R. E., CaciOPPO' J. T'., a Schuman, D. (1983). Central and peripheral ronted to advertising effectivenes: The moderating role of involveumt- W: 19, 135-146. Petty, R. E., Ostron, T'. N., & Brock, T'. C. (1981). Historical foundations of the cognnitive reponse approach to attitude and persuasion. In R. E. Petty, T‘. N. Ostron, a T'. C. Brock (Eds.), M W (pp. 5-29). Hillsdale, N. J.: Erlbaum. Petty, R. E., Wells, G. L., 8 Brock, T'. C. (1976). Distraction can enhannce or reduce yielding to propaganda: Thonght disruption versus effort iustifieatim- WW. :1, 874-884. Plato (c. 357/1952) Gcmias. (W. C. Helnbold, T'rans.) Indianapolis: Library of Liberal Arts. Ray, J. L., Caranza, E., & Reynolds, R. A. (1986). Decodeing permitted choice utterannce . Paper preented at the Wetern Communication Association Convention, Tucson, Arizona, February. Ray, J. L., & Findley, P. (1984). The effects of open and retricted context sentence on decoding and reasoning with conditional propositions. Communication Wang, 3;, 243-252. Reinsch, N. L., Jr. (1971). An invetigation of the effect of the metaphor and simile in persuasive discourse. W: 3.6: 142-145. Elaboration, argunents, evidence, and style 93 Reinsch, N. L., Jr. (1974) . Figurative language and source credibility: A preliminary invetigation and reconceptualization. W Mr .1: 75'30- Reynolds, R. A. (1978). The impact of defedant race, defedant characteristics, and inadmissible tetimony upon results of a simulated jury trial. Paper preented at the Wetern Speech Conmunicaticn Association Convention, Phoenix, Az. November. Reynolds, R. A., a Burgoon, M. (1983). Belief procesing, r,easoning and evidence. In R. Bostrom (Ed. ), Con_gin_nication Yearbook VII (pp. 83-104). Beverly Hills: Sage. Reynolds, R. A. , & Mineo, P. (1984). Hetascientific problems in the study of argument. Paper preented at the aneeh Communication Association convention, Chicago, Novenber. Rcberge, J. J. (1971). Some effects of negation on adults' conditional reasoning abilitie. W: 22: 838-844- Roseberg, N. J., & Hovland, C. I. (1960). Cognitive, affective, and behavioral compoenents of attitude. In C. I. Hovland & H. J. Roseberg (Eds.) , Attitude orflzaticn and we (pp. 1-14) . New Haven: Yale University Pres. Salmon, W. C. (1967) . an; fonndatig 9f scientiiic EEGEQ- University of Pittsburgh Pres. Sandell, R. (1977). Lingpistic gtyie and grsuasicn. London: Academic Frag. ”rift Co We, Sherif, Hop & Nabergall, R. E. (1965). Attim 3rd attitude m : ,, ; Philadelphia: Saunders. Siegel, E. R., Miller, G. R., a Wotring, C. E. (1969). Sonrce credibility and credibility proneness: A new relationship. W: :5, 118-125. Siltannen, S. A. (1981) . The persuasivenes of metaphor: A replication and extension. §chern m Communication Journal, 31, 67-83. Simons, a. w. (1976). Personae: W- Reading, MA: Addison-Wesley. Steinnfatt, T'. M. (1971) . Ambiguity as a predictor of syllogistic dificulty. Unpublished doctoral dissertation, Michigan State University. Steinfatt, T'. E., Miller, G. R., & Bettinghaus, E. P. (1974). The concept of logical ambiguity and judgments of syllogistic validity. WW. .1, 317-328. Elaboration, argunents, evidence, and style 94 Stiff, J. B. (1986). Cognitive procesing of persuasive mesage cue: A meta-analytic review of the effects of anpporting information of attitude. Communication Mgrm, 53, 75-89. Street, R., 8 Brady, R. (1982). Speech rate aoceptannce range as a function of evaluative domain, listener speech rate, and oonmunication context. MW: .12: 290-308- Thonas, G. L. (1956). Oral style and speech intelligibility. m WI 23! 45-54- Thompson, W. (1967). wtitative reearch in 3911c addres _aig W. New York: Ranndon Honse. Tblman, E. C. (1932). Ruggive behavior in animals and men. New York: Appleton-Century-Crofts. mumin, S. (1958). m use of arm. Cambridge: Cambridge University Pres. Wason, P. C., & Johnson—Laird, P. N. (1972). W: W. London: B.T‘. Batsford Ltd. Wheeles, L. (1971). The effects of conprehension loss on persuasion. W: 33, 327-330- Willard, C. A. (1983). Mean and the social ggonnds 9; knowige. University, AL: The University of Alabama Pres. Williens, M. (1980) . The effect of deliberate vaguenes onreceiver recall and agreenmt- WW: 3i: 30-41. William, M. L. , & Goss, B. (1975). Eguivccaticn: Character innsurannce. W: l: 265-270- Wilson, G. B., (1971). Purposeful ambiguity as a persuasive mnesage strategy. Unnpublished masters theis, Michigan State university. Wilson, J. F., 8 Arnold, C. C. (1978). Public fly as a liberal 3:55, (4th ed.). Boston: Allyn and Bacon. Woodworth, R. S., & Sells, S. B. (1935). An aurosphere effect in formal syllogistic reasmmg- MW: 13: 451-460. Wyer, R. S. (1970) . The gnantitative prediction of belief and opinion , change: A further tet of a subjective probability model. m Personaligy and Social Mlggy, 36, 559-571. Wyer, R. S. (1972). Test of a sucjective probability model of social evaluation procese. _ . . -- 32, 279-286. Elaboration, argunents , evidence, and style 95 Wyer, R. S. (1973). Further tet of a subjective probdnility model of social Merenoe- WW: 1: 237-253. War, R. S. (1974) . Some implications of the “Socratic erct' for alternative models of cognitive consistency. WM, _4_2, 399-419. Wyer, R. S. (1975). Effects of previously formed beliefs on syllogistic inference procese. , _ - . a _ - _33, 307-316. wyer, R. B., & Carlston, D. (1979). Social t on erence annd gttribution. Hillsdale, NJ: Erlbaunn. Wirer, R. S., Carlston, D., a Hartwick, J. (1979). The role of syllogistic reasoning in inference based upon nnew and old information. In R. S. We a D. Carlston. Scci tion er and Vlyer, R. S., & Goldberg, L. (1970). A probabilistic annalysis of the realtionships amoung beliefs and attitude. W, _71, 100—120. Vlyer, R. S., & Hartwick, J. (1980). The role of information retrieval and conditional inference procese in belief formation and change. In L. Berkcwitz (Ed.), Advance 19. momenta gig My (Vol. 13, pp. 243-284) . New York: Academic Pres. Zajonc, R. B. (1980). Cognition and social cognition: A historical perspective- In In Festmger (Ed.), Win W. New York: Oxford University Pres. Zillman, D. (1972) . Rhetorical elicitation of agreement in persuasion. W: 2.1.: 159-165. Zillman, D., & Cantor, J. R. (1974). Rhetorical elicitation of concesion 1n persuasion. W: 25: 223-236. Elaboration, arguments, evidence and style 96 Appendix.A Justification for the use of traditional logic in the study of social influence Elaboration, argunents, evidence and style 97 Henpel (1965) in offering an analysis of rationality as an explanatory concept points ont that 'the concept of rationality is by on [sic] means as clear and unequivocal as is sonnetimnes implied in the literature on rational explanation“ (p. 477) . For example, among comnunicaticn scholars, there are vast difference of opinion on the notion of rationality (see Reynolds & Minneo, 1984) . Somne communication scholars discuss rationality as if it amounts to conforming to social normns (e.g., Hawe, 1976) and, therefore, “rationality [is] a rhetorical totem and a field standard but not a hnman characteristic" (Willard, 1983, p. 88) . Somne scholars have examined models of decision making (e.g., see Wyer a Carlston, 1979) as repreenting models of rationality. The prevailing view of rationality, however, involve some connection to formal standards of (primarily deductive) logic (e.g., Miller, 1969: Steinfatt, Miller, & Bettinghaus, 1974: Jackson, 1982: Ray & Findley, 1984) such that the 'concept of a rational agent [is] . . . viewed as an idealized explanatory model conparable to the enplanatory concept of an ideal gas' (Henpel, 1965, p. 477) . It is somewhat in vogue to challenge any view of man as a rational creature. Thee challege, however, seem to emerge from the abstractnes of modern logic, the poverty of the Enlightenment era view of 'rationality [as] God's secular connterpart' (Willard, 1983, p. 89), and 20th century existentialist dread (Barrett, 1958) as well as philosophical inquiry into the ampliativity of deduction and the justification of induction (see Salmon, 1967) . Elaboration, argunents, evidence and style 98 no a leser extent, the challenge to the view of man as a rational creature have been based on enpirical studie of human innference making. Amazingly, reearch reports that document consistent patterns of reasoning errors (e.g., Woodworth & Sells, 1935: Chapmn & (Tnapnan, 1959: Janis & Frick, 1943) seem to be often cited by others (e.g., Jackson, 1982: mans, 1980) insupportoftheclaimthathnmnansarenotrational. Insuch light, McGuire's (1960a) claim that “rational consistency is nneither nonexistent nor absolute" (p. 101) is worth repeating. (bvionsly, if human rationality did not vary, it would not have attracted centurie of study and conmnent. What is important is that the enpirical reearch, in general, is supportive of the conclusion that humans are capable of, and tend towards, reasoning in a mannnner honomorphic with the traditional rule of logic. The early theorizing on reasoning errors, such as Woodworth annd Sell's atmosphere effect hypotheis, was not advannced as an alternative to logical procese but rather as an explanation of the problem inherent in repoding to innconclusive syllogistic premise pairs. An examination of the reports (Woodworth a Sells, 1935: Sells, 1936: Sells & Koop, 1937: (Inapnan and (Tnapnan, 1969: Begg & Dennny, 1969) revels that for premise sets where a conclusion was logically derivable, the objects teded to reach logical conclusions. Similarly, in Jackson's (1982) conparison of the atmosphere hypothesis and the ambiguity hypothesis (Steinfatt, 1970) , the atmosphere hypotheis was superior in acconnting for reasoning errors. But, the prepoderannce of reponse to the reasoning tasks were consistent with logic. Hample (1973)) reviewed the enpirical literature Elaboration, arguments, evidence and style on logical procese and concluded that the weight of evidegxgne favored the view that cognitive procese are logical (of. Evans, 1980) . Wason and Johnson-Laird (1972, p. 44) report that the accuracy of logical conclusions to modus ponens (if p, then q: p: therefore q) is so high that this inference pattern is fundamental and comnon to human thought. an the other hand, there is evidence that people are more likely to commit reasoning errors when the premise contain negative terms (e.g., Roberge, 1971: Wason a Johnson—Laird, 1972), abstract terms (see Wason & Johnson—Laird, 1972) , terms that are retricted to particular contexts (My n. Findley, 1984) , or when the conclusion is discrepant with the individual's preexisting attitude (Janis a Frick, 1943) . In short, violations of the rule of traditional logic can be snbsumed under coditional statements about the nature of the reasoning task or the reasoner. Minneo (1983) and Evans (1980) quetion the use of traditional rule of logic in assesments cf hnmman reasoning. Evans (1980), for example, argue "that systenn of formal logic are not inntended as hypothee abont the nature of thought, but as technique for answering the correctnes of argument. We would not, for example, pose as a general quetion ‘do people think mathematically?" (p. 228) . But, of conrse, the i preupposition that logic or mathematics are social invention doe not deny the utility in studying how people perform thee skills. We would not, for example, conider the same argument as grounds for dispensing with the study of language, politics, busines or, for that matter, persuasion. Moreover, Aristotle (as well as many others) did envision Elaboration, argunents , evidence and style 100 rationality as uniquely hnman and the early beginning of the study of logic began with the examination of refutational technique (Bcchenski, 1956/1961). The contention by some scholars (e.g., mans, 1980: Minneo, 1983) that formal logic is an inadequate model of hnman thonght is indisputable and they shonld be encouraged in their efforts to explore actual inference patterns. For non, hovever, we may use traditional logic (notasachheadorasanirrelevanttool, but) asamodelofhunnan reasoning on the grounds that: 1) 'no other model [of rescuing] yet prcposed decribe the facts with fewer deviation“ (Johnnscn, 1968) , and 2) the model is extensional and, therefore, provide a common ground for analysis and conparison. Elaboration, arguments, evidence and style 101 AppendixB Message Elaboration, arguments, evidence and style 102 VALID W, mm mm, mm meofthehighetconcernsofonrsccietyisthehealthand hopefulnes of humanity. This concern has reulted in legislation which has provided specialized health care programs for the elderly, the poor and the unenployed. While thee programs have benefited many people, several segments of onr society remain in desperate need of adequate health care. In fact, one of the groups which needs care the most, heroin users, has actually been helped the least by current policie. More often than not, the laws regulating the sale of heroin in this conntry have done more harm than good. While many arguments concerning legalization of heroin sale involve issue otherthanhealthcare,wefeelthatonrmajor focusshonldbeon health-related issue. Preent criminal sanction against the sale of heroin were, in part, deigned to protect the American citizen from the medical harmnwhichwereoncethonghtcausedbythedrug. Thereultof those laws, however, is that many addicts face dramatically dreary deaths from diseases caused not by heroin, but from conplications which are pronoted by the drug's continued illegality. For example, in 1979, a team of medical reearchers from the Federal Drug Enforcement Administration reported in the g! M £11.92; g: Medicine that over 900 addicts died in New York City from tetanus and hepatitis. Thee deatnn were not caused by heroin, but were the reult of improper mneans of injection. Sinnce hypodermic syringe cannot be obtained legally, users are forced to reuse and share nneedle, or they improvise with objects not deigned for injecting drugs into the bloodstream. Thus the laws that ban the sale and personal use of heroin actually cause a significant nnumber of heroin-related deaths each year . Asecodhealthhazardissuethatwecannotignoreisthat illicit heroin is impure and, thus, intrinsically injurions. Because the drug is available only on the illicit market, it is haphazardly prepared by street dealers who have little concern m the health of their clients. Sinnce users are never sure of the amount or quality of the heroin they purchase, they are often unable to regulate the dosage of the drugs they take. Consequently, many heroin users accidentally die each year from drug overdose. According to the _Ng York Timnes, reearch by Dr. A1 Jordan, of The Harvard Medical School, "shows that when heroin users know the actual stregth of the drug they're using, deatln annd mnedical problen caused by overdosing are virtually eliminated." If the sale of heroin were legalized, the government wonld be in a position to enforce quality controls on the heroin sold, tins saving many live each year. Sinnce the cost of illegal heroin is so excesive, users are often unable to afford the expediture esential for ecaping poor health. Addicts' failure to get needed nutrition, medication and doctors' care is directly linked to their being forced to use all of their reource to pay inflated black market price for the drug. Profesor William D. Shintig of the UCLA Center for Drug Addiction Elaboration, arguments , evidence and style 103 Research recently wrote in the [ILA Law Review that "etimate by several indepedent reearchers indicate that heroin addicts spend eighty-five percent or more of their income on the drug." Under a system of legalized heroin sale, the price of the drug wonld be greatly reduced and users conld afford other esential health-related products. A fonrth dimension that must be discussed is whether heroin is actually a dangerons or danaging drug. Users of a drug that is not physically damaging can lead normal live. Medical authoritie from the World Health Organization now agree that heroin cause little physical damage. In addition, a team of reearchers from Jchnn Hopkins University reported at a recent cogresional hearing that the symptoms of heroin witndrawal are much les dangerons than time associated with the withdrawal from alcohol. Therefore, when assured of the legal supply of the drug, the heroin user is capable of leading a meaningful and productive life. Varions argunents can be used to support the legalized sale of heroin. We feel, hwever, that even considering only the health care benefits, renoving secondary infection, reducing the risk of overdose, raising the availability of nnutrition and medication, and recognizing the evidence that heroin itself cause little physical damage, justify the legalization of the sale of heroin in the United State. Elaboration, arguments, evidence and style 104 VALID Am, snmr; mm, no “mm One of the major concerns of on: society is the health and welfare of its citizens. This concern has resulted in legislation which has provided specialized health care programs for the elderly, the poor and the unenployed. While thee programs have benefited many people, several segnents of our society remain in deperate need of adequate health care. In fact, one of the gronps which needs care the most, heroin users, has actually been helped the least by current policie. More often than not, the laws regulating the sale of heroin in this conntry have done more harm than good. While many argnments concerning legalization of heroin sale involve issue other than mm care, we feel that our major focus shonld be on health-related a. Preent criminal sanction against the sale of heroin were, in part, deigned to protect the American citizen from the medical harms which were once thoght caused by the drug. The reult of those laws, however, is that many heroin addicts are dying needlesly from diseases caused not by heroin, but from secondary conplication which are promoted by the drug's continued illegality. For example, in 1979, a team of medical reearchers from the Federal Drug Enforcement Administration reported in the _Ne_n_w Q'gland Jonrnal gf Medicine that over 900 addicts died in New York City from tetanus and hepatitis. These deaths were not caused by heroin, but were the reult of improper mean of injection. Since hypodermic syringe cannot be obtained legally, users are forced to reuse and share needle or improvise with objects not deigned for injecting drugs into the bloodstream. Thus the laws that ban the sale and personal use of heroin actually cause a significant nnmber of heroin-related deatln each year. Asecodhealthhazard facedbytheuseristhepurchaseofimpure heroin. Because the drug is available only on the illicit market, it is haphazardly prepared by street dealers who have little concern abont the health of their clients. Sinnce users are never sure of the anount or quality of the heroin they purchase, they are often unable to regulate the dosage of the drngs they take. Consequently, nanny heroin users accidentally die each year from drug overdose. According to the _N_ew_ York Time, reearch by Dr. Al Jordan, of The Harvard Medical School, "shows that when heroin users know the actual stregth of the drug they're using, deatln and medical problen caused by overdosing are virtually elinminated." If the sale of heroin were legalized, the government wonld be in a position to enforce quality controls on the heroin sold, thus saving nanny live each year. Sinnce the cost of illegal heroin is so excesive, users are often unable to afford itens esential for maintaining good health. Addicts' failure to get needed nutrition, medication and doctors' care is directly linked to their being forced to use all of their reonrce to pay inflated black market price for the drug. Profesor William D. Shintig of the [11A Center for Drug Addiction Reearch recently wrote in the [CIA Law Review that "etimate by several indepedent reearchers indicate that heroin addicts sped eighty-five percent or Elaboration, arguments, evidence and style 105 more of their inncome on the drug." Under a system of legalized heroin sale, the price of the drug wonld be greatly reduced and users conld afford other esential health-related products . A fonrth issue that must be conidered is whether the use of heroin constitute a real health problem. Users of a drug that is not physically damaging can lead normal live. Medical authoritie from the World Health Organization non agree that heroin cause little physical danage. In addition, a team of reearchers from Johns Hopkins University reported at a recent cogresional hearing that the symptoms of heroin witldrawal are much les dangerons than those associated with the withdrawal from alcohol. Therefore, when assured of the legal supply of the drug, the heroin user is capable of leading a meaningful and productive life. Varions argnments can be used to support the legalized sale of heroin. We feel, however, that even conidering only the health care benefits, eliminating secodary infection, reducing the number of overdose, and inncreasing the availability of nutrition and medication, conpled with the evidence that heroin itself cause little physical damage, justify the legalization of the sale of heroin in the Unnited State. Elaboration, arguments, evidence and style 106 VALID Am, WEAK EVIDHCE, AILITRATIW aeofthehighetconcernsofonrsocietyisthehealthand hopefulnes of humanity. This concern has resulted in legislation which has provided specialized health care programs for the elderly, the poor and the unenployed. While thee prograns have benefited nanny people, several segments of our society remain in deperate need of adequate health care. In fact, one of the gronps which needs care the most, heroin users, has actually been helped the least by current policie. More often than not, the laws regulating the sale of heroin in this conntry have done more harm than good. While many arguments concerning legalization of heroin sale innvolve issue other than health care, we feel that on major focus shonld be on health-related issue. Preent criminal sanctions against the sale of heroin were, in part, deigned to protect the American citizen from the medical harms which were once thought caused by the drug. The reult of those laws, however, is that many addicts face dramatically dreary deaths from disease caused not by heroin, but from secodary complication which are pronoted by the drug's continued illegality. For exanple, in 1969, a team of prenedical students from Slippery Rock State College reported that a significant number of addicts died in New York City from secodary illnese reulting fromn the use of drug paraphernalia. Thee deaths were not caused by heroin, but were the reult of improper means of injection. Sinnoe hypodermic syringe cannot be obtained legally, users are forced to reuse and share needle or improvise with objects not deigned for injecting drugs into the bloodstream. Thus the laws that ban the sale and personal use of heroin actually cause a significant nnumber of heroin-related deaths each year. A secod health hazard issue that we cannot ignore is that illicit heroin is impure and, thus, intrinsically injurions. Because the drug is available only on the illicit market, it is haphazardly prepared by street dealers who have little concern about the health of their clients. Since users are never sure of the amount or quality of the heroin they purchase, they are often unable to regulate the dosage of the drugs they take. Consequently, many heroin users accidentally die each year fromn drug overdose. According to Al Jordan, Secretary of the Seattle Drug Rehabilitation Council: "It sees that when log term heroin users know the actual stregth of the drug they're using, deaths and medical problems caused by overdosing are reduced." If the sale of heroin were legalized, the government wonld be in a position to enforce quality controls on the heroin sold, thus saving many live each year. Since the cost of illegal heroin is so excesive, users are often unnable to afford the expenditure esential for escaping poor health. Addicts' failure to get needed nutrition, medication and doctors' care is directly linked to their being forced to use all of their reonrce to pay inflated black market price for the drug. Bill "Fetch" Shintig of the National Coalition for the legalization of Marijuana claims that "heroin addicts doped on others to supply Elaboration, argnments, evidence and style 107 than with food and clothes because they can barely find the reonrce to pay the high costs of the drug." Under a systemn of legalized heroin sale, the price of the drug wonld be greatly reduced and users conld afford other esential health-related products. A fourth dimension that must be discussed is whether heroin is actually a dangerons or damaging drug. Users of a drug that is not physically damaging can lead normal live. It is now commonly agreed that heroin cause little physical damage. In addition, a health and nutrition reearcher from the University of Arkansas claims that the symptoms of heroin witldrawal are abont as bad as those associated with the witndrawal from extrene case of alcoholism. Therefore, when assured of the legal supply of the drug, the heroin user is capable of leading a meaningful and productive life. Various arguments can be used to support the legalized sale of heroin. We feel, however, that even considering only the health care benefits, renoving secodary infection, reducing the risk of overdose, raising the availability of nutrition and medication, and recognizing the evidence that heroin itself cause little physical damage, justify the legalization of the sale of heroin in the United State. Elaboration, argunents, evidence and style 108 VALID W, WEAK EVIDBCE, no AILITMICN aneofthemajorconcernsofonr societyisthehealthand welfare of its citizen. This concern has resulted in legislation which has provided specialized health care prograns for the elderly, the poor and the unenployed. While thee prograns have benefited many people, several segments of onr society renain in deperate need of adequate health care. In fact, one of the gronps which needs care the most, heroin users, has actually been helped the least by current policie. More often than not, the laws regulating the sale of heroin in this conntry have done more harm than good. While nanny argunents concerning legalization of heroin sale innvolve issue otherthanhealthcare,wefeelthatonrmajorfccusshonldbeon health-related issue . Preent crinminnal sannction againnt the sale of heroin were, in part, deigned to protect the American citizen from the medical harmswhichwereoncethonghtcausedbythedrug. Thereultof those laws, however, is that manny heroin addicts are dying needlesly fronn disease caused not by heroin, but from secodary conplications which are pronoted by the drug's continued illegality. For example, in 1969, a team of prenedical students from Slippery Rock State College reported that a significant number of addicts died in New York City from secodary illnese reulting from the use of drug paraphernalia. Thee deaths were not caused by heroin, but were the reult of improper mean of injection. Sinnce hypodermic syringe cannot be obtained legally, users are forced to reuse and share needle or improvise with objects not deigned for injecting drugs into the bloodstream. Thus the laws that ban the sale and personal use of heroin actually cause a significant number of heroin-related deaths each year. Asecodhealthhazardfacedbytheuser isthepurchaseof impure heroin. Because the drug is available only on the illicit market, it is haphazardly prepared by street dealers who have little concern abont the health of their clients. Sinnce users are never sure of the anount or quality of the heroin they purchase, they are often unable to regulate the dosage of the drugs they take. Conequently, many heroin users accidentally die each year fronn drng overdose. According to Al Jordan, Secretary of the Seattle Drnng Rehabilitation Council: "It seems that when log term heroin users know the actual stregth of the drug they're using, deatln and medical problems caused by overdosing are reduced." If the sale of heroin were legalized, the government wonld be in a position to enforce quality controls on the heroin sold, thus saving many live each year. Since the cost of illegal heroin is so excesive, users are often unable to afford itens esential for maintaining good health. Addicts' failure to get needed nutrition, medication and doctors' care is directly linked to their being forced to use all of their reonrce to pay inflated black market price for the drug. Bill "Fetch" Shintig of the National Coalition for the legalization of Marijuana claims that "heroin addicts doped on others to supply them Elaboration, arguments, evidence and style 109 withfccdandclothebecausetheycanbarelyfindthereonrceto pay the high costs of the drug." Under a systenn of legalized heroin sale, the price of the drug wonld be greatly reduced and users conld afford other essential health-related products . A fonrth issue that must be conidered is whether the use of heroin contitute a real health problem. Users of a drug that is not physically damaging can lead normal live. It is now connonly agreed that heroin cause little physical damage. In addition, a health and nutrition reearcher from the University of Arkansas claims that the symptoms of heroin witldrawal are abont as bad as those associated with the witldrawal from extrene case of alcoholismn. Therefore, when assured of the legal supply of the drug, the heroin user is capele of leading a meaningful and productive life. Varions argunents can be used to support the legalized sale of heroin. We feel, however, that even conidering only the health care benefits, eliminating secodary infection, reducing the nunber of overdose, and increasing the availability of nutrition and medication, conpled with the evidence that heroin itself cause little physical damage, justify the legalization of the sale of heroin in the United State. Elaboration, arguments, evidence and style 110 INVALID W, M EVIDDCB, no AILITEATIW One of the major concerns of our society is the health and welfare of its citizen. This concern has reulted in legislation which has provided specialized health care programs for the elderly, the poor and the unenployed. While thee programs have benefited manny people, several segments of onr society renain in deperate need of adequate health care. In fact, one of the gronps which needs care the most, heroin users, has actually been helped the least by current policie. More often than not, the laws regulating the sale of heroin in this country have done more harm than good. While many arguments concerning legalization of heroin sale involve issue other than health care, we feel that onr major focus shonld be on health-related issue. Preent criminal sanction against the sale of heroin were, in part, deigned to protect the American citizen fronn the medical harms whichwereoncethoghtcausedbythedrug. Thereultofthoselaws, however, is that many heroin addicts are dying needlesly from diseases caused not by heroin, but from secodary conplication which are promoted by the drnng's continued illegality. For exanple, in 1979, a team of medical reearchers from the Federal Drug Enforcement Administration reported in the _Ng ggland Jonrnal _q_f_ Medicine that over 900 addicts died in New York City from tetannus and hepatitis. These deaths were not caused by heroin, but were the reult of improper means of injection. Because drug users reuse and share needle or improvise with objects not deigned for injecting drugs into the bloodstream, the drug and hypodermic syringe cannot be ontained legally. Thus reducing the nunber of heroin-related deatln Sfoachheyeirr‘wonldactuallyelimimtetheneedforlawsthatbantlesale ro . Asecodhealthhazardfacedbytheuseristhepurchaseof impure heroin. It is because the drug is haphazardly prepared by street dealers who have little concern abont the health of their clients that the drug is illegal. Sinnce users are never sure of the amount or onality of the heroin they purchase, they are often unable to regulate the dosage of the drugs they take. Conequently, nanny heroin users accidentally die each year fromn drug overdose. According to the fig York Times, reearch by Dr. Al Jordan, of The Harvard Medical School, shows that "when heroin users know the actual stregth of the drug they're using, deaths and medical problens caused by overdosing are virtually eliminated." If we conld reduce the number of live lost eachyear, thenthegovermentwonldnotneedtobeinthepositionof enforcing quality controls of the heroin sold, thus removing the need to keep the drug illegal. Since the cost of illegal heroin is so excesive, users are often unable to afford itens esential for maintaining good health. If, however, addicts' did get needed nutrition, medication and doctors' care they wonld not be forced to use all of their reonrce to pay innflated black market price for the drng. Profesor William D. Shintig of the (ILA Center for Drug Addiction Research recently wrote in the (BIA Law Review that "etimate by several indepedent Elaboration, arguments, evidence and style 111 researchers indicate tiat heroin addicts spend eighty-five percent or more of their income on the drug." Under a system where the price of the drug wonld be greatly reduced, the sale of heroin conld be legalized. Afourthissuethatmustbeconiderediswhethertheuseof heroin contitute a real health problem. If you lead a normal life, the use of a drug is not physically damaging. Medical authoritie from the World Health Organization now agree that heroin cause little physical damage. In addition, a team of reearchers fromn Johns Hopkins University reported at a recent cogresional hearing that the symptoms of heroin withdrawal are much les dangerous than those associated with the withdrawal from alcohol. Therefore, if heroin users lead meaningful and productive live, there is no need to continnue keeping the drug illegal. Various argunents can be used to support the legalized sale of heroin. We feel, however, that even conidering only the health care benefits, eliminating secondary infection, reducing the number of overdoses, arnd inncreasing the availability of nutrition and medication, coupled with the evidence that heroin itself cause little physt1m t_égal damage, justify the legalization of the sale of heroin in the State. Elaboration, arguments, evidence annd style 112 INVALID W, sum; EVIDDCE, AILITERATICN aneofthehighestconcernsofoursocietyisthehealthand hopefulnes of humanity. This concern has resulted in legislation which has provided specialized health care programs for the elderly, the poor annd the unenployed. While thee programs have benefited nanny people, several segments of onr society renain in deperate need of adequate health care. In fact, one of the gronps which needs care the most, heroin users, has actually been helped the least by current policie. More often than not, the laws regulating the sale of heroin in this conntry have done more harm than good. While many argnmnents concerning legalization of heroin sale involve issue other than health care, we feel that onr major foons shonld be on health-related issue. Preent criminal sannction against the sale of heroin were, in part, designed to protect the American citizen from the mnedical harms which were once thoght caused by the drug. The reult of those laws, however, is that many addicts face dramatically dreary deaths from diseases caused not by heroin, but from secondary complications which are promoted by the drug's continued illegality. For example, in 1979, a team of medical reearchers fromn the Federal Drug Enforcement Administration reported in the M England Jonrnal 9_f_ Medicine that over 900 addicts died in New York City from tetannus annd hepatitis. Thee deaths were not caused by heroin, but were the reult of improper means of injection. Because drug users reuse annd share needle or improvise with objects not deigned for injecting drugs into the bloodstream, the drug annd hypodermic syringe cannot be obtained legally. Thus reducing the number of heroin-related deaths :fachneyefiwwldactuallyelimimtetheneedforlwsuatbantlnsale r0 0 A second health hazard issue that we cannot ignnore is that illicit heroin is impure annd, thus, intrinsically injurions. It is because the drug is haplnazardly prepared by street dealers who have little concern about the health of their clients that the drug is illegal. Sinnce users are never sure of the amnount or quality of the heroin they purclnase, they are often unable to regulate the dosage of the drugs they take. Conequently, many heroin users accidentally die each year from drug overdose. According to the _Ng York Times, reearch by Dr. Al Jordan, of The Harvard Medical School, shows that "when heroin users know the actual stregth of the drug they're using, deatln and medical problens caused by overdosing are virtually eliminated." If weconld reducethennmber of live losteachyear, thenthe government wonld nnot need to be in the position of enforcing quality cfintrols of the heroin sold, thus removing the need to keep the drug legal. Since the cost of illegal heroin is soexcesive, users are often unable to afford the expenditure esential for ecaping poor health. If, however, addicts' did get needed nutrition, medication annd doctors' care they would not be forced to use all of their reonrce to pay inflated black market price for the drug. Profesor William D. Shintig of the m Center for Drug Addiction Research recently Elaboration, arguments, evidence annd style 113 wrote in the (ILA Law Review that "etimate by several inndependent reearchers inndicate tlat heroin addicts spend eighty-five percent or more of their income on the drug." Under a system where the price of the drug wonld be greatly reduced, the sale of heroin conld be legalized. A fonrth dimension tlat must be discussed is whether heroin is actually a dangerons or damaging drug. If yon lead a normal life, the use of a drug is not physically damaging. Medical authoritie from the World Health Organization now agree that heroin cause little physical damage. In addition, a teamn of reearchers from Johns Hopkins University reported at a recent cogresioal hearing that the symptoms of heroin withdrawal are much les dangerous than those associated with the withdrawal from alcohol. Therefore, if heroin users lead meaningful annd prodnctive live, there is no need to continue keeping the drug illegal. Various argnmennts can be used to support the legalized sale of heroin. We feel, however, that even conidering only the health care benefits, removing secondary infection, reducing the risk of overdose, raising the availability of nutrition annd medication, annd recognizing the evidence that heroin itself cause little physical damage, justify the legalization of the sale of heroin in the United State. Elaboration, arguments, evidence annd style 114 INVALID W, WEAK mm, m ALITERATICN One of the major concernn of our society is the health annd welfare of its citizen. This concern has resulted in legislation which has provided specialized health care progranns for the elderly, the poor annd the unenployed. While thee progranns have benefited many people, several segnnnennts of onr society renain in deperate need of adequate health care. In fact, one of the groups which needs care the mnost, heroin users, has actually been helped the least by current policie. More often than not, the laws regulating the sale of heroin in this conntryhavedonemsrenarmthangood. Whilemanyargnmnennts concerning legalization of heroin sales innvolve issue other than fith care, we feel that our major focus shonld be on health-related a. Preent criminal sannction against the sale of heroin were, in part, deigned to protect the American citizen fronn the medical harms which were once thoght caused by the drug. The reult of those laws, however, is that many heroin addicts are dying needlesly from disease caused not by heroin, but fromn secondary complication which are promoted by the drug's continued illegality. For example, in 1969, a team of prenedical students fronn Slippery Rock State College reported that a signnificant nunber of addicts died in New York City from secondary illnesse resulting fromn the use of drug paraphernalia. These deaths were not caused by heroin, but were the result of improper means of injection. Because drug users reuse and stare needle, or they improvise with objects not deigned for injecting drugs into the bloodstream, the drug and hypodermic syringe cannnot be obtained legally. Thus reducing the number of heroin-related deatln each year would actually eliminate the need for luvs tint ban the sale of heroin. Asecondhealthhazard facedbytheuser isthepurchaseofimpure heroin. It is because the drug is prepared haptazardly by street dealers who have little concern abont the health of their clients that the drug is illegal. Sinnce users are never sure of the amnount or qnality of the heroin they purchase, they are often unable to regulate the dosage of the drugs they take. Consequently, many heroin users accidentally die each year fronn drug overdose. According to Al Jordan, secretary of the Seattle Drug Rehabilitation Council; "It seens that when log term heroin users know the actual stregth of the drug they're using, deatln and medical problens caused by overdosing are reduced." If we conld reduce the nunber of live lost each year then the government wonld not need to be in the position of enforcing qnality controls on the heroin sold, thus rennoving the need to keep the drug illegal. Sinnce the cost of illegal heroin is so excessive, users are often unable to afford itens esential for maintaining good health. If, however, addicts did get needed nutrition, medication annd doctors' care they wonld not be forced to use all of their reonrce to pay inflated black market price for the drug. Bill "Fetch" Shintig of the Natioal Coalition for the legalization of Marijuana claims that "heroin addicts depend on others to supply them with food annd clothes Elaboration, arguments, evidence annd style 115 becausetheycanbarely finndthereonrcetopaythehighcostsof the drug." Unnder a system where of the price of the drug would be greatly reduced, the sale of heroin conld be legalized. A fonrth issue that must be considered is whether the use of heroin constitute a real health problem. If yon lead a normal life, the use of a drug is nnot physically damaging. It is now commonly agreed tiat heroin cause little physical damage. In addition, a health annd nutrition reearcher from the University of Arkansas claims that the symptoms of heroin withdrawal are abont as bad as those associated with the withdrawal from extrene case of alcoholisnn. Therefore, if the heroin users lead meaningful and productive live, there is no need to continnue keeping the drug illegal. Varions arguments can be used to support the legalized sale of heroin. We feel, however, that even conidering only the health care benefits, eliminating secondary infection, reducing the number of overdose, annd increasing the availability of nnutrition annd medication, conpled with the evidence that heroin itself cause little phys!l gal danage, justify the legalization of the sale of heroin in the Un t State. Elaboration, argnmnents , evidence annd style 116 INVHLID.ARGUMENTS,‘WEHK EVIDENCE,nALITERNTION aneofthehighetconcernsofonrsocietyisthehealthand hopefulnes of hnmanity. This concern has resulted in legislation which has provided specialized health care programs for the elderly, the poor annd the uneployed. While thee programs have benefited nanny people, several segments of on: society resin in deperate need of adequate health care. In fact, one of the gronps which needs care the most, heroin users, has actually been helped the least by current policie. More often than not, the laws regulating the sale of heroin in this comtry have done more harm than good. While many argnments concerning legalization of heroin sales involve issue other than health care, we feel that onr major focus shonld be on health-related issue. Preent criminal sannction against the sale of heroin were, in part, deignned to protect the American citizen from the medical harms which were once thoght caused by the drug. The reult of those laws, however, is flat many addicts face dramatically dreary deaths fromn disease caused not by heroin, but from secondary conplication which are promoted by the drug's continued illegality. For example, in 1969, a team of prenedical students from Slippery Rock State College reported that a significant number of addicts died in New York City fronn secondary illnese reulting fromn the use of drug paraphernalia. These deaths were not caused by heroin, but were the reult of improper means of injection. Because drug users reuse annd share needle, or they improvise with (injects not deignned for injecting drugs into the bloodstream, the drug and hypodermic syringe cannnnot be obtained legally. Thus reducing the nnmber of heroin- related deaths each year wonld actually eliminate the need for laws tint ban the sale of heroin. A second health hazard issue that we cannnnot ignnore is that illicit heroin is impure annd, thus, intrinsically injurious. It is because the drug is prepared taptazardly by street dealers who have little concern abont the health of their clients that the drug is illegal. Since users are never sure of the amnount or quality of the heroin they pnrcnase, they are often unable to regulate the dosage of the drugs they take. Conequently, many heroin users accidentally die each year from drug overdose. According to Al Jordan, secretary of the Seattle Drug Rehabilitation Council; "It sees that when log term heroin users know the actual stregth of the drug they're using, deatln annd medical problen caused by overdosing are reduced." If we conld reduce the number of live lost each year then the government wonld not need to be in the position of enforcing quality controls on the heroin sold, thus renoving the need to keep the drug illegal. Sinnce the cost of illegal heroin is so excesive, users are often unable to afford expenditure esential for ecaping poor health. If, however, addicts did get needed nutrition, medication annd doctors' care they wonld not be forced to use all of their reonrce to pay inflated black market price for the drug. Bill "Fetch" Shintig of the Natioal Coalition for the legalization of Marijuana claims tlat "heroin addicts depend on others to supply then with food annd clothes Elaboration, argnments, evidence and style 117 becausetheycanbarelyfinndthereonrce topaythehighcostsof the drug." Under a system where of the price of the drug wonld be greatly refined, the sale of heroin conld be legalized. A fonrth dimennion trat must be disonssed is whether heroin is actually a dangerons or damaging drug. If yon lead a normal life, the use of a drug is not physically damaging. It is now conmonly agreed that heroin cause little physical damage. In addition, a health annd nutrition reearcher from the University of Arkansas claims that the symptons of heroin withdrawal are about as bad as those associated with the withdrawal from extrene case of alcoholisnm. Therefore, if the heroin users lead meaningful annd productive live, there is no need to continue keeping the drug illegal. Varions argunents can be used to support the legalized sale of heroin. We feel, however, tnat even conidering only the health care benefits, removing secondary infection, reducing the risk of overdose, raising the availability of nutrition annd medication, annd recognizing the evidence that heroin itself cause little physical damage, justify the legalization of the sale of heroin in the united State. Elaboration, arguments, evidence annd style 118 Appendix C Measureent Instrnmnents Elaboration, argnmennts, evidence annd style 119 Conmunication Reearch Fall 1985 Introduction In order to improve on reearch annd education in conmunication, the reearch faculty of the communication studie department must compare several different message. Some of thee message may be selected for future communication experiments. Yon are asked to aid in this effort by reading annd providing yonr reaction to the following message. In order to inncrease the accuracy of the comparison it is important that yon read themessageasyonwonldanyothermessage. PLEASEREADTHEMESSPGEASIFYGJAREREADDGIT'IN AMPGAZINEORT‘HEQINIWSECPIWG‘AWGSPAPER It is very important that everyone follow similar procedure in the conpletion of this booklet of materials. Therefore, please follow the instructions very carefully. Since the other people participating in this effort may be reading annd reponding to different message it is important that you do not talk or distract others while conpleting the materials. If yon have a quetion or need assistannce, please raise yonr hannd annd wait for the reearch assistant to cone to yon. After yon finish reading this page, turntothenextpageanndbegin readingthemessage. Whenyou finish reading the message continue to the reponse section of the When yon lave conpleted the materials please turn this booklet booklet. face down annd await further inntructions. Thank yon very much for yonr cooperation. In order to help us keep track of the booklets, please list the last four digits of yonr social security number in the space provided below. unis information will be used only to identify the booklets annd will not be used to identify you with yonr reponse. IASI'FGJRDIGITSCFYQJRKXIIALSEIIJRITYMER Elaboration, argnmennts , evidence and style 120 EEASEREADTHEWVMGRWLY an the following page yon will find a series of statements and objects followed by several scale. Please mark each scale in the blannk tlat BEST represents how yon feel about the stateennt or object. For example, here isan itemliketheoneyonwill see: The United State shonld withdraw fronn the United Nations. Yonr job is to place a check mark (X) above the line that bet indicate your feeling toward the stateennt. For example, if yon feel that 0.8. withdrawal would be a very good idea, yon would check as follows: If yon feel such a move (withdrawal) would be slightly beneficial, yon would check as follows: Good::X::::Bad If yon feel neutral or indifferent about the proposition, or if yon feel :hat particular scale is irrelevant to the proposition, you wonld check as ollows: Remember: Fill ont every scale and never make more than onemark on a Single scale. Thank yon for yonr cooperation. Elaboration, argnmennts, evidence and style 121 l. The sale of heroin should be legalized in the United State. Inncorrect ___:___:_:_:___:__:_ Correct True __ :_ :_ :_____:___. n__ a“ False Probable _:__ :_ :__:__ n__ :_ Imprwable Unlikely _:___ n__ :_ :_ :_ :___ Likely 2. The legalization of the sale of heroin in the United state to improve addicts' health wonld be. Bad ___:___ :__.,:___ :___:____:__ Good Valuable : n :_ :_ n__ :_ Worthles Foolish _ :_ :_:_ :_ :_ :_ Wise Pleasant _:__:_:__:_:___n__ Unpleasant Desirable : n : : n : Unndeirable 3. Please inndicate your reaction to the we yon have just read. Accurate __ :_ :___ :____ n____ :_ :_ Inaccurate Unbelievable __ :___:__ :__:_ :___:__ Believdnle Valid _=_.3.._=...=_..3... :_ Invalid Good n n z n____:___:___ Bad Wise __ :_ :_ :_:___ :_____ nu Foolish Responsible : : n_: z : Irresponible logical _:_:__:___:___x___:_ Illogical Powerful : n : :___:__:__ Weak Mild : z z n____:____:____ Strong Poorly reasoned __.n___:___x : n : Well reasoned Clear _:__: z : : :_ Unnclear anpported ____:___3___: n : :_ Unnupported Unnconvinncing : z :_ :_ :___:___ Convincing Effective _:__:___:___:___:_:__ Ineffective Unsonnnd : n : :___:___:__ Sonnnd Well written _:_____:___:__:___:__:__ Poorly written Conpelling _:_ :___:___:___:____:____ Unnconpelling Disorganized __ :__:__:_:__:_:__ Organized Rational : n n n n : Irrational ——m—-mm”~ 4. Please indicate yonr reactions to the M in the message you have just read. Predictable ____,:__, :m :____:_____:_____:____ Unpredictable ed _:___:___ :__:__:__.:____ ed Anticipated __:___:__: : :_ Unanticipated : : : Elaboration, arguments, evidece and style 122 5.Wewou1d liketoknowwhatyouthinkthemgofthemessage believe. A. It is because heroin and hypodermic syringe are illegal that addicts use improper meann of injecting the drug. A* It is because heroin addicts use improper mean of injection ttat the drug and hypodermic syringe cannnot be obtained legally. Themofthemesagethinksthatthestateentdnoveisn Unlikely _n___:_n_n__:___:___ Likely Incorrect _n__n__n____n___n__n__ Correct True _x___:___n__n___n___:_ False Prdnable ____:_ n_n__n___:___:__ Imprdnable B. Manyoftheheroin relateddeatineachyeararetheresultof improper means of injecting the drug. Thesonggofthemessagethinksthatthestateentaboveis: Unlikely _:___ n_n___n___ n_x____ Likely Incorrect _n__n__n___n__n___:__ Correct True ___n___n___n_n___n_n__ False Probable __n__n___n_n__n_n___ Improbdnle C.Thelawsthatbanthesaleanduseofheroinactuallycausea signnificant number or heroin related deatln each year. C* Reining the significant nnumber of heroin related deatin wonld eliminate the need for laws that ban the sale and use of heroin. nemofthemesagethinkstlatthestateentaboveis: Unlikely _n___n__n___n__n__n____ Likely Incorrect _:_:__n__n___n__n__ Correct True ___:___n___n___:___:____n____ False Probdnle ___n_:_n__n___n__n___ Imprdnable D. Becausestreet heroin is impure, manyaddicts overdoseonthednug. flamethemessagethinkstlatthestateentaboveis: Unlikely _:__:___:___n___x__:__ Likely Incorrect ___n____n____:__: : : Correct True _n__:___n___:_n__n__ False Probable _n__ n_:__:___ n_:___ Imprdndole *Statements specific to the invalid argnmennts. Only those statements specific to the version of the message read were encountered. Elaboration, arguments, evidece and style 123 E. It is because heroin is illegal that addicts use inure heroin. 8* It isbecausestreet heroin is impure that heroinis illegal. hemofthemessagethinksthatthestateentaboveisn unlikely ___n_n___:___n___n__n__ Likely Incorrect n_n : : n_n Correct True 2 n n n n 3 False Probable _n__ n_n_n__ n___ n__ Imprdndnle F.1tisbecauseheroinisillega1thatmnyaddictsoverheonthe drug. F*Itisbecausemanyaddictsoverdosethatheroinisillegal. fiemofthemesagethinksthatthestatementaboveisx unlikely ___n__n__n n n n___ Likely Incorrect n n n n :___n___ Correct True 3 n n n_n___n___ False Prcbdnle ____n__ n_x___ :_ n_n_ Inprdndnle G. If the use of heroin was legalized, the price of the drug wonld be greatly reoced. 6* If addicts received proper health care, they wonld not be forced to use all of their reonrce to pay for the drug. Thesonrgofthemessagethinkstiatthestatementaboveis: Unlikely _n n n n n : Likely Incorrect __:_n__:__n___n_:__ Correct T'rue _x__n__n__n___n___n_ False Probable _n__n__n_n__n__n__ Imprdndnle B. If the price of heroin was greatly rednced, addicts conld afford other essential health related products. Themofthemessagethinksthatthestatenentaboveis: Unlikely ___n___:___n__n__x__:_ Likely Incorrect _n_n__n___n___n__n__ Correct True 3 n n n 3 False Eldnoration, argnmnents, evidece and style 124 I. If heroin was legalized, addicts conld afford other essential health related products. 1* If the price of heroin was greatly reduced, the sale of heroin conld be legalized. 'nnesongceofthemessagethinksthatthestatementdnoveis: Unlikely _n___:__n__:__n___n__ Likely Incorrect _:___n__n___n__:___n__ Correct True n n n n_:___n__ False Probable _:__n_:_n___n_n__ Improbdnle J. Users of a drug that is nnot physically damaging, can lead normal live. J*Ifyonleadanormallife,thentheuseofadrugneednotbe physicallydmaging. Themofthemessagethinksthatthestatementdnoveis: unlikely _n_n___n___n___n___:___ Likely Incorrect ___n__n___n_n__n__n__ Correct True _n__n__n___n__n__n__ False Probdnle _n_n__n_n__ n_:___ Imprdndnle K. The use of heroin cause little physical damage. Themofthemessagethinksthatthestateentdnoveisn unlikely _n__n__:__n__:__n__ Likely Incorrect _n__n_ n_:__:___:_ Correct True _n__n__n____n___:_:__ False Prdndnle ___:_n___n___n__n__n___ Improbdnle L. Heroin users can lead a normal life. L*Ifheroin users lead nnormal live, thenthereisnoneedto keep theanpplyanduseofthedrugillegal. Themofthemessagethinksthatthestateentdnoveis: Unlikely _n__:_n_n_:__n__ Likely Incorrect _n__n_n__n___:__n___ Correct True __n___n : n n 3 False Probable ___n___n__:___n_ n_n_ Imprdadnle Elaboration, argnments, evidece and style 125 6. Wpreented inthemessagewasn A. Inconsistent with the arguments being advanced. Agree _n_ n__ n_n___:_ :_ Disagree B. Onrrent to the issue today. Agree ___n_n__ n_n__:__:__ Disagree C. Sufficient to prove the points being supported. Agree 2 n ______n :_ Disagree D. Not typical of what is known abont this_— problem. Agree ___n_:___n_n_n_n__ Disagree E. Irrelevant to the coclusions drum in the message. Agree _n__n___x___n_:___n__ Disagree F. Not clear and understandable. Agree _:_n___ n:__n_n_ __ Disag ree G. Doommentedwellenonghsothatanyreaderconldlook itup. Agree__:__nn:n_____n_____ Drisag ree H. Contained clear and understandable statistical information. Agree __ :___:___n___n___:__n____ Disagree I. Taken as a whole, sumorted the point being made. Agree _:___n__n__n_n__n___ Disagree J. Came from experts on the topic. Agree _:_n___n__n_n__n__ Disagree K. Came from persons who were biased on the topic. Agree _n__n___n__:__n__n__ Disagree L. Came from credible magazines or books. Agree _n__n___n__ n_n___ :_.__ Disagree 7. While reading the message were m: A. Attenpting to analyze the issues in the message. Agree _n___n__n____n___n____n___ Disagree B. Not very attentive to the ideas. Agree __ n_n___n___n___n__n_ Disagree C. Deep in thoght abont the message. Agree _n___ n_n___n__n__n___ Disagree D. Uncocerned with the ideas. Agree ___n__n__n___:___n___n___ Disagree E. Extending a good deal of cognnitive effort. Agree __ n__ n___ n_n____:__n___ Disagree F. Distractedbyotherthonghtsnnot relatedtothemessage. Agree __ n_n___n____ n_n___ :_ Disagree G. Not really eerting yonr mind. Agree _n___: n n n_n_ Disagree Elaboration, arguments, evidece and style 126 H. Doingyonrbesttothinkdnontwhatwaswritten. Agree _n__:____:__n___ n_n__ Disagree I. Reflecting on the implications of the argnmnents. Agree _n__n__n__n___n_n___ Disagree J. lasting yonr mind. Agree 3 n ______2 :_ Disagree K.Searchingyourmindinresponetotheideas. Agree ___n_n___:_n_n_n__ Disagree L. Taking it easy. Agree _n__n_:__n__n___n_ Disagree 8.Wewou1dliketokncwhowyondescribeyourselfn A. I wonld prefer complex to simple problems. . Agree 2 n r__ : __:__:___ Disagree B. I like to have the responibility of handling a situation that require a lot of thinking. Agree ___:_:___n___n__n_n__ Disagree C. Thinking is not my idea of fun. Agr ree : n _n___ Disagree D.Iwouldratherdosomethingtiatrequireslittlethonghtthansomething that is sure to challege my thinking abilities. Agree _n___n__:___n___n___n___ Disagree E. I try to annticipate and avoid situations where there is a likely chance I will have to think in depth wont something. Agree _n__: __ n____n__n____n__ Disagree F. I find satisfaction in deliberating hard and for log hours. Agree __ n_n__n___:__n__n___ Disagree G. Ionlythinkashardaslhaveto. Agree __ n_n___:__ n_n___n___ Disagree H. I prefer to think abont small, daily projects to log-term ones. Agr__ee n n n n __n__n____ Disagree 1. I like tasks that require little thoght oce I've leaned them. Agree _:__n_n__ n_n____n__ Disagree J.Theideaofrelyingonthoghttomakemywaytothetopappealsto me. Agree __ :n___:____ __ n_n___n___ Disagree K. I really enjoy a tek that involve coming upwithnewsolutionto problen. Agree ___:__n___n__n__:__n__ Disagree L.Learningnnewwaystothinkdoen't ecitemeverymuch. Agree __ n__ n____n____ n_n___n__ Disagree M. I prefer my life to be filled with puzzle that I must solve. Agree __ r__: n_n___n__n__ Disagree N. O. P. Q. R. Elaboration, arguments, evidece and style The notion of thinking dntractly is annealing Agree 3 n n 127 to me. Disag ree I wonld prefer a task that is intellectual, difficult, and important toonethatissomewhat importantbutdoenotrequiremuchthonght. Agreen::____ _______n:___ Disagree I feel relief rather than satisfaction after ‘conpleting a task ttat required a lot of mental effort. Agree __ n_z : n n 1 It's enough for me that something gets the job or why it works. Agree _: n n n :_ Disagree donenI Idon't care how Disagree Iusuallyedupdeliberatingabontissueevenwhentheydonot affect me personally. Agree _n n n n n : Disagree Elaboration, arguments, evidece and style 128 Appendix D Supplementary tables Elaboration, arguments, evidece and style 129 Table 2 Conparison of scores for the Source's belief in the propostion W W PAl. It is because heroin and PA1.* It is because heroin addicts hypodermic syringe are illegal use improper means of injection tlat that addictsuseimpropermeannof thedrugandhypodermic syringe injecting the drug. cannnnot be obtained legally. Means: .87 .54 g - 67.94, df - 1/239, p - .000 P81. Manny of the heroin related PBl. Many of the heroin related deatlneachyeararethereultof deattneachyeararethereultof improper mean of injecting the improper meann of injecting the drug. drug. Means: .86 .84 E I 017' df - V239, p - 068 PC1. The laws that ban the sale PC1.* Reducing the signnificant and use of heroin actually cause a nunber of heroin related deatrn wonld significant nunber of heroin actually eliminate the need for laws relateddeathseachyear. thatbanthesaleanduseofheroin. Means: .81 .66 E - 15.10, df ‘ V239, P ' .000 PA2. Because street heroin is PA2. Because street heroin is impure, many addicts overdose on impure, many addicts overdose on the drug. the drug. Means: .84 .85 g- .02, df-l/239, p- .90 P82. It is because heroin is PB2.* It is because street heroin illegal that addicts use impure is impure that heroin is illegal. heroin. Means: .71 39 z - 53.22, an. - 1/239, p - .ooo' PC2.Itisbecauseheroinis PC2.*Itisbecansemannyaddicts illegalttatmanyaddicts overdose overdosethatheroinisillegal. on the drug. Means: .53 .58 _F- .90,df-l/239,p-.35 Elaboration, argnments, evidece and style Table 2 (cont.) PA3. If the use of heroin was legalized, the price of the drug wonld be greatly reduced. Mean: .87 E . 48.77, df ' P83. If the price of heroin was greatly reduced, addicts conld afford other esential health related products. Means: .82 E I 9.59, df ' PC3. If heroin was legalized, addicts conld afford other essential health related products. Meann: .79 _F- - 91.67, df - PM. Usersofadrugthatis not physically damaging, can lead normal live. Means: .77 E 3 1.55, df 3 P84. The use of heroin cause little physical damage. Means: .75 z . 018' at - PC4. Heroin users can lead a normal life. Means: .74 E " 041' at . 130 PA3.* If addicts received proper health care, they wonld not be forced to use all of their reonrce to pay for the drug. .61 U239, p 8 .000 P83. If the price of heroin was greatly reduced, addicts conld afford other essential health related products. .70 U239, p I- .002 PC3.* If the price of heroin was greatly reduced, the sale of heroin conld be legalized. .40 U239, p 8 .000 PA4* If yon lead a normal life, thantheuseofadrugneednot be physically damaging. .72 1,239] P ' .215 P84. The use of heroin cause little physical damage. .73 U239, p B .68 PC4.* If heroin users lead normal live, thenthereisnoneedto keepthesupplyanduseofthedrug illegal. .71 U239, P 3 .52 Elaboration, arguments, evidece and style 131 Tdnle 3 Conparisons of reasoning errors for each argnmnent and the argnmnents conbined laid me .P; also Argument 1: Predicted on the basis of mean scores for the individual proposistion: .21 .06 Ontained: .33 .18 14.98 .000 Argument 2: Predicted on the basis of mean scores for the individual proposistions: .25 .07 Ontainned: .28 .21 3.47 .064 Argument 3: I Predicted on the basis of mean scores for the . individual proposistion: .23 .08 Ontained: .28 .16 12.70 .000 Argument 4: Predicted on the basis of mean scores for the individual proposistion: . 18 . 16 (btained: .19 .17 .64 .18 Total for the argunents combined predicted on the basis of mean score for the individual proposistions: 1.08 .72 Ontainned: 1.07 .72 23.78 .000 Elaboration, arguments, evidece and style 132 Table 4 Factor analyse of belief and attitude scale Orthogonal rotation Oaliqne rotation Factor 1 Factor 2 Scale Attitude Belief Correct-incorrect .79 .39 .46 True—false .81 .31 .58 Probdnle—improbable .74 -. 05 .93 Likely-unlikely .69 -.03 .85 Good-bad .83 .91 -.06 Valuable-worthless . 86 . 92 -. 01 Wise-foolish .90 .91 .02 Pleasant-unpleasant .73 .71 .04 Desireable-undesireable .81 .77 .07 Eigenvalue 6.06 6.03 0.95 t of variance .67 .67 .11 Reliability .94 .93 .89 Correlation between factors .73 (.80) Elaboration, arguments, evidece and style 133 Table 5 Correlation matrix NcogValidEVidStyleFprttElabDRE as as s Ncog 1.00 Valid -.03 1.00 Evid .00 .01 1.00 Style -.02 -.03 -.02 1.00 Exp .13 .06 .04 -.04 1.00 Att -.06 .04 .02 -.09 .15 1.00 El» .31 .07 .09 -.04 .03 .00 1.00 DRE -.15 -.30 .01 .02 .02 .10 -.14 1.00 m -.02 .20 .16 .03 .27 .36 .11 .03 1.00 ME -.08 .17 .04 -.02 .21 .54 .00 .09 .70 1.00 B -.05 .01 .00 -.03 .15 .73 .00 .13 .34 .51 1.00 Where: Att IAttitude Exp - Expectancy B - Belief ME - Message evaluation DRE - Detection of reasoning error PE - EVidece evaluation Elab - Elaboration Evid - EVidece stregth Moog - Need for cognition Style 8 Message style (aliteration) Valid - Argument validity