The reported research considered the applicability of a dual process framework to stereotype change. Three experiments manipulated task involvement, source credibility, and message quality and measured stereotype change. Under conditions of high involvement, stereotype change was effected only by message cues. In low-involvement conditions, there was an additive effect of source and message cues. This credibility effect was due to an enhancement of the impact of stereotype-disconfirming information when presented by a highly credible source. These findings show differential processing of stereotype-disconfirming information under hig4 and low involvement, consistent with the dual process approach. Processing time, recall, and thought listing also showed differential processing of the counter stereotypical message under high-and low-involvement conditions. Results are discussed in terms of the models of persuasion, and implications for traditional stereotype change research are considered.