5. Ibid. 6. T. C. Brook, "Communicator-Recipient Similarity and Decision Change," Journal of Personality and Social Psychology, I (1965) pp. 650-654. 7. Hovland, et al., Communication. 8. Elaine Walster and L. Festinger "The Effectiveness of 'Overhead' Persuasive Communications," Journal of Abnormal and Social Psychology, 65 (1962) pp. 395-402. 9. L. Festinger and N. Maccoby, “On Resistance to Persuasive Communication," Journal of Abnormal and Social Psychology, 68 (1964) pp. 359–366. 10. J. Mills, "Opinion Change as a Function of the Communicator's Desire to Influence and Liking for the Audience," Journal of Experimental Social Psychology, II, (1966) pp. 152-159. THE LIKABILITY AND SELF-INTEREST OF THE BY VERNON A. STONE AND HARROGADDE S. ESWARA Evidence provided reveals that it is not always safe to assume that a likable source will be more effective than an unlikable source, and that self-interest (as reflected by occupation) interacts with other source variables rather than manifesting itself as a main effect. Recent studies have shown that the most favorable source is not always the most effective agent of persuasion. Bauer has found that enduring attitude change in the face of counterpropaganda may be better wrought in some cases by a source of questionable trustworthiness than by a high-trust source.1 Also in contradiction to common sense notions, Walster, Aronson and Abrahams have found that in some cases a low prestige source may be more persuasive than a high prestige communicator.2 The latter finding came in a test of the hypothesis that "regardless of his prestige, a communicator will gain in effectiveness when he advocates a position opposed to his own best interests and will lose effectiveness when he advocates a selfish positon." 3 To test this hypothesis, the Walster group used a prosecutor as the high prestige source and a criminal as the low prestige source. The topic was court power-whether there should be more of it, which would presumably be in the best interest of the prosecutor, or less court power, in the best interest of the criminal. There were four conditions, each source taking each side of the argument. As predicted, the criminal was more effective in changing attitudes when he argued against his best interest. But contrary to expectations, the prosectuor was almost equally effective whichever way he argued. This indicated an interaction between prestige and self-interest whereby the *From "The Likability and Self-Interest of the Source in Attitude Change," Journalism Quarterly, XLVI, No. 1 (Spring 1969), pp. 61-68. Reprinted with the permission of Journalism Quarterly, copyright holder. low prestige source gained in effectiveness by arguing against his own best interests, but made no difference for the high prestige source. Walster and her associates did not accept the interaction interpretation, however. They held the results might have been in part an artifact of the high school student subjects' being more familiar with arguments for strong courts than with arguments against them. They conducted a second experiment in which the sources were a prosecutor and a criminal from the supposedly less familiar Portuguese court system. In this experiment, as predicted, each source was more effective when he argued against his self-interest. Support was claimed for the hypothesis that such will be the case regardless of prestige. A test of that hypothesis would demand that there indeed be differences in the perceived prestige of the sources. In the first experiment by the Walster group, the prosecutor was clearly rated higher than the criminal on prestige. But in the second experiment, with the setting shifted to Portugal, the criminal was rated almost as high as the prosecutor. It appeared possible that prestige differences might have interacted with self-interest in the first experiment but were too small to do so in the second. The present study was addressed to the conflicting questions posed by the two earlier experiments: Does arguing against his best interest enhance the persuasiveness of a communicator, regardless of his prestige? Or does such enhancement come only for low prestige sources? The approach taken here differed from that of Walster and her associates in three basic ways. First, to rule out the possibility of differential effectiveness of messages arguing for and against an issue, the same message was used in all conditions. Second, occupation was used for the manipulation of self-interest rather than of prestige; the manipulation of positive and negative source characteristics was undertaken independently of occupation. Third, in ascribing the source characteristics constituting global prestige, credibility and personal likability were manipulated separately. The latter strategy was intended to provide a test of the relative effects of credibility and likability on attitude change. The source characteristics of expertness, honesty and influence which constituted the prestige index used by the Walster group may be seen as essentially components of credibility in that they deal with the credentials of the source in relation to the message. But Aronson and Golden have demonstrated that opinion change may be influenced by characteristics of the source which are more or less irrelevant to his credentials. Moreover, Rarick has separated "cognitive" and "affective" components of source prestige and found both to be positively related to attitude change." Credibility, in turn, may be considered in terms of such components as expertness and trustworthiness. Although probably closely related in most cases, they do not necessarily have to be. One can conceive of a source which is considered highly expert but untrustworthy and another which is trustworthy but inept. The credibility component chosen for manipulation in the present investigation was professional expertness. Likability was treated as including only those characteristics which generally are considered to make a person personally likable but are not necessarily related to his expertness in a profession or an area of knowledge. Predictions were made for expertness rather than likability because expertness was considered closer to the global prestige variable used in the study suggesting the present one. The main objective was a test of the opposing predictions suggested by data from the study by Walster and associates. The prediction derived from their hypothesis was for a main effect of self-interest whereby a source arguing for such selfinterest would be more persuasive than a source arguing for such selfinterest, regardless of expertness. The opposing prediction, suggested by the first experiment by the Walster group, was for an interaction whereby an inexpert source would be more effective when arguing against than when arguing for his best interests, but an expert source would be equally effective whichever way he argued. METHOD Subjects were 95 male and 73 female students enrolled in an undergraduate radio-television course at the University of Wisconsin. The experiment was conducted by the senior author during a regular class period. Subjects were randomly assigned to the eight experimental conditions in a 2x2x2 factorial design. The manipulated variables were (1) the self-interest (occupation), (2) the expertness and (3) the likability of the source. The main dependent variable in the before-after design was attitude change on the issue of courtroom television. All materials were contained in a single mimeographed booklet titled "Free Press-Fair Trial Study." Booklets for the eight conditions were identical except for the page describing the source. The attitude change message was represented to the subjects as one of the responses in a nationwide survey requesting opinions on the issue of television in the courtroom from a random sample of lawyers and newsmen. The message, a 36-line argument against courtroom television, was constructed largely of excerpts from the United States Supreme Court decision of 1965 in the Billie Sol Estes case. To point up the self-interest differences of the occupational variable, the booklet's introduction explicitly noted that television journalists generally consider it against their best interest to have their cameras banned from courtrooms and that criminal lawyers generally favor such a ban on grounds that the presence of television cameras may interfere with a fair trial for their clients. It was also noted that the news media and the bar have collided on the issue many times in recent years. The source was introduced as JRB, a respondent in the nationwide survey. At the start of a two-paragraph description, he was further introduced as either (1) a television journalist or (2) a lawyer specializing in criminal cases. Expertness was manipulated in the first paragraph by describing the source as either (1) a successful practitioner in his profession who had received high recognition from his colleagues or (2) an unsuccessful practitioner who had been described by colleagues as "lacking in sound. judgement and professional competence." Thus professional prestige was used as an indicator of expertness. Likability was manipulated in the second paragraph of the source description by noting that "people who know JRB on a day-to-day basis" describe him as either (1) pleasant, unpretentious, sensitive to the feelings of others, etc., or (2) unpleasant, pretentious, insensitive, etc., characteristics which would tend to make a person unlikable though he might be a highly respected expert in his profession. No explicit information was provided regarding the source's trustworthiness. Any perceptions of trustworthiness were left to be inferred from the descriptions of the professional expertness and personal likability of the communicator. All attitudes were assessed on bipolar 7-point scales before and after the attitude change message was read. Attitude toward the issue was assessed by four agree-disagree items expressing the issue of television. and free press-fair trial in slightly different ways. The source was evaluated on six dimensions: (1) his reliability as a source of information on the issue at hand, (2) expertness on the issue, (3) bias, (4) trustworthiness, (5) likability as a person and (6) general attractiveness. The message was evaluated by (1) how well made and (2) how sincere were the arguments.7 RESULTS Source Evaluation Factors. A factor analysis of correlations between source evaluation items yielded two factors (Table 1). Reliability and expertness loaded high on the first factor; scores on those two items were summed for the expertness used in the study. Likability and attractiveness evaluations loaded high on the second factor and were summed for likability scores. Because the loadings for trustworthiness and bias were distributed equally between the two factors at the start of the experiment, those characteristics were treated separately. Manipulation Checks. The pre-message ratings in Table 2 show that the source manipulations yielded the intended high-low distinctions both in the expertness and likability conditions. The midpoint of each evaluation continuum was 8. The mean initial expertness rating given sources in high expert conditions was 9.48 and in low expert conditions 5.79. The difference between likability ratings in the likability conditions was even greater a mean of 10.21 for subjects who read the high likability source description as against 4.74 for those reading the low likability description. |