that could have directly cued hostility" (p. 276). The absence of differences in the responses of high and low prejudiced participants in the Devine study may thus have occurred because of the direct activation of semantic associations involved in stereotype priming rather than because of a close association between the category alone and the stereotype.
Lepore and Brown (1997) further argued that "high-and low-prejudice people's representations of the social group may not differ in terms of content (at least for stereotype knowledge) but stronger links may have developed for different characteristics" (p. 277). Lepore and Brown reasoned that, as a consequence of this differential strength of associative links with the category, high and low prejudiced people would show divergent automatic stereotype activation as a function of category priming. Consistent with their hypothesis, using Devine's (1989, Study 2) priming and subsequent impression formation procedure, Lepore and Brown found that when only the category was primed, high prejudiced participants showed evidence of automatic negative stereotype activation, whereas low prejudiced participants did …show more content…
Because participants were given the time and opportunity to ascribe stereotypic traits deliberately to the particular categories, this process is considered to be controlled. Thus, the degree of the participants' endorsement of the cultural stereotypes was expected to vary as a function of prejudice (Devine, 1989; Esses et al., 1993; Lepore & Brown, 1997).
Participants' activation of cultural stereotypes, alternatively, was assessed with a word pronunciation task. Specifically, participants, who were classified as high or low in prejudice, were presented with a category prime (Black, White, or CCC
[a neutral baseline]) followed by a positive or negative Black stereotypic target word or nonstereotypic target word. Their task was simply to pronounce the target word. Response latency was the dependent measure. A number of studies have revealed that this procedure may produce a particularly sensitive measure of automatic processing because the paradigm does not foster task-specific strategies that can obscure the effects of automaticity (Balota & Chumbly, 1984; Balota &
Lorch, 1986; Bargh, Chaiken, Raymond, & Hymes, 1996; Joordens & Besner, 1992; Ratcliff & McKoon, 1988).