A study was conducted to test the hypothesis that for an "uninvolving topic" the persuasiveness of a highly credible communicator would be mediated by the favorability of the communicator-related thoughts generated, while for an "involving topic" the effect of the communicator's credibility would be mediated by the favorability of the message. Sixty undergraduate students in four experimental groups and one control group listened to radio interview tapes concerning a policy change that would or would not affect them and, involving speakers related to or unconnected with the issue. Subjects then completed a booklet containing the dependent measures. All subjects accurately recalled the topic of the persuasive message and whether the proposed policy change was to go into effect immediately (high involvement) or in 10 years (low involvement). Subjects in the high credibility condition rated the speaker as more knowledgeable, qualified, expert, and skilled, and also expressed stronger agreement with the message than did subjects in the low credibility conditions. The results suggest that while listening to the persuasive message, uninvolved subjects thought primarily about the speaker, while the involved subjects thought primarily about the persuasive message. This pattern is consistent with the argument that uninvolved subjects use a heuristic suggested by the communicator's credibility to evaluate the message, while involved subjects evaluate the message according to the strength of the argument. (HTH)