Two information-processing mechanisms that could potentially contribute to judgmental discrimination against the members of stereotyped social groups were examined in two experiments, using a mock juror decision-making task.  Bost postulated mechanisms involve biased processing of judgment-relevant evidence.  The interpretation hypothesis asserts that the activation of stereotypic concepts influences the perceived probative implications of other evidence.  The selective processing hypothesis asserts that stereotype-consistent evidence is processed more extensively than is inconsistent evidence.  Judgment and memory data from the first experiment supported the general notion that stereotype-based discrimination emerged from biased evidence processing.  The specific pattern of results supported selective processing rather than interpretation biases as the critical process underlying observed judgmental discrimination.  The second experiment corroborated this conclusion by showing that a manipulation that prevents selective processing of the evidence effectively eliminated biases in judgments and recall pertaining to stereotyped targets.  Implications for a general understanding of stereotyping and discrimination are discussed.