Summary
One-hundred eighty American student Ss formed an opinion about a factual or value issue and later received feedback about the agreeing or disagreeing opinion of another S, who was either attitudinally similar or dissimilar, and who scored either higher than or equal to the S in an “intelligence” test. The dependent variable was the S's opinion confidence after feedback. Similarity and ability were expected to interact with type of opinion, whether it was fact- or value-related, and whether it conformed to or was opposed to the S's opinion. The fourfold interaction did not emerge. However, disagreement significantly reduced confidence with factual, but not with value-related opinion. Also, agreement of a superior other increased and disagreement decreased confidence. The failure to find effects of attitude similarity is discussed with regard to task relevance and generality of the manipulation.