In two Luchins-type impression formation experiments, both using an afteronly design, nonvolunteer college students were exposed to short communications describing the personality of a well known sports figure positively (i.e., friendly, outgoing, generous) and/or negatively (i.e., unfriendly and unsocial). The following results were obtained: 1. Primarily based on an Einstellung hypothesis, it was posited that cognitive "organization" would decrease and indecision increase, the less familiar was the subject of the communications and the less salient were impressions of him. It was predicted that one result would be an increased incidence of neutral ratings on items asking raters for their impressions of the subject. The findings strongly support this hypothesis. The tendency to rate neutral was more prevalent when impressions of the subjects were nonsalient, or minimally salient, than when they were salient, and most prevalent when nonsalience interacted with unfamiliarity. 2. It also was predicted from Einstellung that under conditions of nonsalience a kind of primacy effect would result. That is, when cognitive organization is reduced, the first communication because it supplies the recipient with new organization should have greater potency than the last. Therefore, greater incongruent impression change (i.e., weakening original impressions) should result when recipients' initial exposure is to a communication dissonant with their original impressions than to a communication consonant with their original impressions. In this study, since recipients' original impressions of the subject were positive, the hypothesis would be confirmed if more negative responses resulted from exposure to a negative-positive sequence than to a positive-negative sequence of communications. The hypothesis was in fact supported with raters highly familiar with the subject, but was not supported with raters unfamiliar with the subject. If a positive correlation exists between interest and familiarity, then it is possible to explain this finding by suggesting that the low familiarity group may have lacked sufficient interest to attend to the stimuli. 3. The results also indicated a tendency toward recency when salience was high. Since salience may "naturally" interact with familiarity, the role of familiarity in primacy-recency would appear, therefore, to be more complex than has been intimated in the literature. In fact, a recent study by Sears and Freedman (28) finds that even the expectation of unfamiliarity may serve as a potent catalyst for increasing the effectiveness of a persuasive appeal. Although the experiments reported here originally were intended to clarify the familiarity relationship in impression formation, their results raise many more questions about that variable than they answer. A U curve hypothesis, designed to reconcile the apparent inconsistency between an unfamiliarity-rccency finding in opinion change and unfamiliarity-primacy in impression formation, received only weak support, and a final decision accepting or rejecting that hypothesis was deferred until more definitive evidence can be gathered. [ABSTRACT FROM AUTHOR]