Biases in memory for and use of inconsistent beliefs in stereotyping.

Some research has suggested that level of inconsistency of information that invalidates expectancy-based stereotypes or trait labels is one of several moderating variables that determine the memorability and use of inconsistent information (e.g. Rojahn & Pettigrew, 1992). Two experiments were designed to identify variables that moderate memory for and use of counter-stereotypic social-political beliefs that were attributed to a stereotypically identified group. In Expt 1, recall was found to be inversely related to level of inconsistency. In Expt 2, location of counter-stereotypic beliefs within a list and the order of counter-stereotypic beliefs (most-to-least vs. least-to-most inconsistent) determined the form of the relationship between recall and level of inconsistency. When counter-stereotypic beliefs were presented relatively early in a list and were ordered from least-to-most inconsistent, a positive relationship obtained; otherwise, the relationship was negative. Stereotyping was less extreme when counter-stereotypic beliefs were presented early in the least-to-most inconsistent ordering. Cognitive processes that may have produced these sets of findings were discussed.