The goal of the study reported in this article was to examine whether automatic processes in stereotype and prejudice activation are sensitive to task characteristics of the assessment procedure and whether these influences may account for existing inconsistencies that have recently been reported in the literature on automatic racial prejudice. Using a sequential priming paradigm with subliminal primes (“BLACK” and “WHITE”) to examine automatic prejudice, the study varied the judgment task in which the priming procedure was presented. Whereas half of the participants were asked to perform a lexical decision task (word/nonword), the remaining participants made evaluative judgments (good/bad). Results showed reliable influences of the judgment task on the observed pattern of priming effects. Moreover, the priming effects found in both conditions replicated the respective results reported in previous research that had used either evaluative or conceptual judgment tasks (Fazio, Jackson, Dunton, & Williams, 1995; Wittenbrink, Judd, & Park, 1997). In addition, the response time measure also showed different relationships with explicit measures of racial prejudice, depending on the judgment condition. In addition to their implications for the assessment of automatic stereotyping and prejudice these results suggest that automatic responses are not as invariant as it is sometimes posited.