Finds that the subjective probability model continues to provide some degree of prediction for beliefs (of an individual for circumstances of a single event with multiple causes) prior to the exposure to a message, but that after exposure to a persuasive message, the model did not maintain the same level of accuracy of prediction. Offers several…

Uses meta-analysis to test three models of the "sleeper" effect (a theory proposing that in the long term low credible sources may actually attain greater attitude change in message recipients than high credible sources). Concludes that the sleeper effect does exist, and offers suggestions for practitioners and theorists. (SR)