The implicit association test (IAT) is typically used to assess nonconscious categorization judgments that are "under control of automatically activated evaluation" (Greenwald, McGhee, & Schwartz, 1998, p. 1464) and that are usually considered independent of explicit judgments. The present study builds on recent work suggesting evidence of short-term modifiability of the IAT effect. Specifically, we show that reading a short text that describes a novel, fictional scenario, within which the to-be-evaluated categories are embedded, can produce substantial and immediate modulations of the IAT effect. This modulation effect does not occur when subjects are simply instructed to think about counterstereotypical associations (Experiment 1A and 1B). In Experiment 2, we use a variant of the IAT to show that scenario modulation cannot be explained in terms of strategic criterion shifts. These results suggest that a newly acquired knowledge structure targeting the abstract, category level can produce behavioral effects typically associated with automatic categorization.
A dominant theme in research on social cognition is that categorizations and evaluative judgments often occur implicitly (e.g., Bargh, 1999; Bodenhausen & Macrae, 1998; Brewer, 1988; Devine, 1989; Fiske & Neuberg, 1990). According to this view, people we encounter trigger automatic activation of category-specific, schematic knowledge that was acquired throughout our learning history. Such automatic categorization is not only held responsible for negative stereotypes and their behavioral consequences, but it is also deemed off-limits for explicit, verbal report. Therefore, much recent work has focused on developing behavior-based assessment techniques that allow tapping the automatic categorization system.
The most prominent method in this regard is the so-called implicit association test (IAT; Greenwald, McGhee, & Schwartz, 1998; see also Devine, 2001), which involves intermixing two types of speeded, two-choice discriminations. The first concerns the categorization of interest. For example, participants may be asked to discriminate between names of flowers and insects. The second discrimination concerns an evaluative dimension such as that between pleasant and unpleasant words. The critical manipulation is the match between stimulus-response mappings for the two types of discriminations. For a person who favors flowers over insects, a compatible mapping implies using one key for flowers or pleasant words and the other for insects or unpleasant words. In the case of an incompatible mapping, the same key is used for flowers or unpleasant words and the other key for insects or pleasant words. The IAT effect is the response time (RT) difference between these two mapping conditions. IAT effects have been found for numerous social and nonsocial category evaluations, are usually very robust (on the order of 100 msec or more), and are obtained even for participants who show no preferences in explicit self-report. Therefore, the IAT effect is assumed to be resistant to explicit influences and thus a relatively pure reflection of implicit, category-relevant associations (e.g., Ashburn-Nardo, Voils, & Monteith, 2001; Greenwald et al., 1998; McConnell & Leibold, 2001).
The fact that much human categorization may be automatic in nature is important because automatic processing can affect thinking and action without our knowing (Blair, Ma, & Lenton, 2001) and is usually thought to develop over time with practice. As a result, automatic categorization ought to be resistant to change (Bargh, 1999). Interestingly, however, recent work with the IAT suggests some short-term modifiability of automatic category judgments. For example, Dasgupta and Greenwald (2001) showed that after exposing participants to examples of positively viewed blacks, the typical black versus white IAT effect expressed by white subjects is reduced. Similarly, Blair et al. …