Cognitive consistency has become, in social psychology, a basic assumption in many theories and experiments. It is, for example, a fundamental postulate of congruity theory (Osgood and Tannenbaum, 1955), dissonance theory (Festinger, 1957(Festinger, , 1964Aronson, 1968), and balance theory (Abelson & Rosenberg, 1958). Recently, however, Freedman (1968) has questioned the importance and generalizability of cognitive consistency, suggesting that with the exception of some very contrived situations,.~.it accounts for but a small portion of response variance. Orloff and Fish (1973) provided support for'such limitations. Their data indicated that when intelligent Sis agreed with a persuasive speech, they also evaluated the speech as b eing logical and the speaker as credible. That is, persuasion scores were highly correlated with perceived message logic and with perceived speaker credibility. However, these correlations were significantly smaller for low I. Q. yes. These findings led to the conclusion that high I. Q. Sus may change their attitudes for different reasons, or at least, make a greater effort to hold cognitions consistent with their attitudinal positions, suggesting a &dquo;moderator&dquo; function for intelligence. Should these findings replicate, an attempt to modify the assumptions of consistency theory in line with different levels of intellectual ability would seem in order. That is, if there is a greater need for cognitive consistency among more intelligent people, any tendency to change one's attitude for reasons of cognitive consistency would appear to be greater in this group. Such theoretical modifications appear essential if the previously cited theories are to remain viable.This study is a test of the generalizability of these results across stimulus conditions. In the present study, demand characteristics were controlled. In addition, the sex of the speaker delivering the persuasive message, and the topic of the message were altered. Method 81 Male and 119 female undergraduates were divided into four experimental groups and one control group with N = ~+0 per group.Experimental Ss were presented with either a logical or an illogical speech, preceded by either a high or low speaker credibility introduction. Both speeches argued for a change from a becameral to a unicameral system of government. Following the speech, _Ss completed a post-experimental questionnaire designed to assess the degree of persuasion, perceived logicality of the message and perceived credibility of the speaker. Following completion of this questionnaire, all Ss completed form I of the Wonderlic Personnel Test (Wonderlic, 1959), a short speed test of I. Q.In order to minimize demand characteristics, it was stressed to all Ss during the introduction, that the E's were not trying to persuade them of anything, but rather, were primarily interested in obtaining their &dquo;evaluations&dquo; of the speech.
ResultsConsistency theory predicts that a speech perceived as logical and a speaker perceived as credible will be more pers...