The authors describe a theoretical framework for understanding when people interacting with a member of a stereotyped group activate that group's stereotype and apply it to that person. It is proposed that both stereotype activation and stereotype application during interaction depend on the strength of comprehension and self-enhancement goals that can be satisfied by stereotyping one's interaction partner and on the strength of one's motivation to avoid prejudice. The authors explain how these goals can promote and inhibit stereotype activation and application, and describe diverse chronic and situational factors that can influence the intensity of these goals during interaction and, thereby, influence stereotype activation and application. This approach permits integration of a broad range of findings on stereotype activation and application.

Stereotypes activated upon initial exposure to a stereotyped individual may dissipate as the exposure continues. Participants observing a videotaped interview with a Black person showed activation of the stereotype of Black people following 15 s of observation but not following 12 min of observation. However, the discovery of a disagreement with the stereotyped individual may bring the dissipated stereotype back to mind. Participants who discovered, at the end of a 12-min videotaped interview with a Black person, that this person disagreed with them about the verdict in a court case showed activation of the stereotype of Black people, whereas participants who discovered instead that the Black person agreed with them did not. Participants who disagreed with a Black person also applied the Black stereotype to him, but this stereotype application was detected only on an implicit measure of application, not on an explicit measure.

Motivation may provoke stereotype use. In a field study of students’ evaluations of university instructors and in a controlled experiment, participants viewed women as less competent than men after receiving negative evaluations from them but not after receiving positive evaluations. As a result, the evaluation of women depended more on the favorability of the feedback they provided than was the case for men. Most likely, this occurred because the motivation of criticized participants to salvage their self-views by disparaging their evaluator led them to use a stereotype that they would otherwise not have used. The stereotype was not used by participants praised by a woman or by participants who observed someone else receive praise or criticism from a woman; all these participants rated the woman just as highly as participants rated a man delivering comparable feedback.

The accessibility of people's highest hopes and achievements can affect their reactions to upward comparisons. Three studies showed that, under normal circumstances, individuals were inspired by an outstanding role model; their motivation and self-evaluations were enhanced. However, when their most positive self-views were temporarily or chronically activated, inspiration was undermined, and individuals' motivation and self-evaluations tended to decrease. Another study found that role models inspired participants to generate more spectacular hopes and achievements than they would have generated otherwise. It appears that increasing the accessibility of one's best selves undercuts inspiration because it constrains the positivity of the future selves one may imagine and prevents one from generating the more spectacular future selves that the role model normally inspires.

The motivation to form a particular impression of an individual can prompt the inhibition of applicable stereotypes that contradict one's desired impression and the activation and application of stereotypes that support it. Participants, especially those high in prejudice, inhibited the Black stereotype when motivated to esteem a Black individual (because he had praised them). Participants motivated to esteem a Black doctor also activated the doctor stereotype. In contrast, participants motivated to disparage a Black doctor (because he had criticized them) inhibited the doctor stereotype. Participants motivated to disparage a Black individual also applied the Black stereotype to him, rating him as relatively incompetent. All these effects were driven by the self-protective motives of recipients of feedback from Black evaluators; detached observers showed no such effects.

The authors examined how the extent to which counterstereotypic individuals deviate from perceivers' stereotypes affects their impact on these stereotypes and found that extremely deviant group members provoke less stereotype assimilation than do moderately deviant ones. Extremely deviant examples can even provoke boomerang effects, that is, enhance the very stereotype that they violate. When participants whose prior stereotype were moderate or extreme were exposed to moderately or extremely deviant examples, the deviant examples' impact on stereotypes depended both on their extremity and on the extremity of perceivers' prior stereotypes. Boomerang effects were obtained only for extreme-stereotype participants exposed to extremely deviant examples and were mediated by perceptions of the typicality of the deviant examples. Open-ended explanations revealed that the atypicality of extremely deviant examples was used as grounds for dismissing them.

The same trait may imply different behaviors when applied to members of differently stereotyped groups. For example, these studies show that aggressive connotes physical violence when applied to a construction worker but verbal abuse when applied to a lawyer. Such stereotype-driven construals of traits can be more readily explained by a parallel-constraint-satisfaction model of impression formation (Z. Kunda & P. Thagard, 1996) than by more traditional models of representation commonly used by social psychologists. These studies show also that stereotypes affect predictions about a person's trait-related behavior even after the stereotype's impact on the trait itself has been undermined by individuating information about that person. The parallel-constraint-satisfaction model is better able to account for such a pattern than are earlier, serial models of stereotype use.

The authors propose that superstars are most likely to affect self-views when they are considered relevant. Relevant superstars provoke self-enhancement and inspiration when their success seems attainable but self-deflation when it seems unattainable. Participants' sell-views were affected only when the star's domain of excellence was self-relevant. Relevant stars provoked self-enhancement and inspiration when their success seemed attainable in that participants either still had enough time to achieve comparable success or believed their own abilities could improve over time. Open-ended responses provided rich evidence of inspiration in these circumstances. Relevant stars provoked, if anything, self-deflation when their success seemed unattainable in that participants either had already missed the chance to achieve comparable success or viewed their abilities as fixed and so unlikely to improve.

The authors propose a parallel-constraint-satisfaction theory of impression formation that assumes that social stereotypes and individuating information such as traits or behaviors constrain each other's meaning and jointly influence impressions of individuals. Building on models of text comprehension (W. Kintsch, 1988), the authors describe a connectionist model that can account for the major findings on how stereotypes affect impressions of individuals in the presence of different kinds of individuating information; how stereotypes, behaviors, and traits affect each other's meaning; and how multiple stereotypes jointly affect impressions. Most of these findings can be modeled by constraint networks, which suggests that they may be due to relatively automatic processes that require little conscious inference. The authors also point to a small number of phenomena that involve more controlled processes. The advantages of the authors' parallel model over serial models are discussed.

People encountering deviants who violate a stereotype try to maintain the stereotype by subtyping the deviants. They use the deviants' additional attributes to justify subtyping them. Participants read about counterstereotypic targets. Participants who were given no additional information about targets, and so had no grounds for subtyping them, did generalize from them and changed their stereotypes. However, participants who were told that targets had an additional, neutral attribute appeared to use it as grounds for subtyping them; their stereotypes remained unchanged. Participants came to view the neutral attributes as atypical of the stereotype and as associated with deviance, that is, as good reasons for subtyping the deviant. Neutral attributes blocked generalization from truly counterstereotypic targets but not from overly stereotypic ones, suggesting that their effect was due to participants' attempts to explain away individuals who strongly challenge their stereotypes.

People prefer controllable risks over less dangerous uncontrollable ones. Two kinds of evidence suggest that people′s belief that they are more capable than others at controlling dangerous outcomes contributes to this preference: In Study 1, the preference for controllable risks was reduced when subjects were choosing risks for another person rather than for themselves. In Study 2, this preference was stronger when the ability required to control outcomes was conducive to an "I am above average" bias than when it was not. The preference for controllable risks over less dangerous uncontrollable ones was obtained for hypothetical scenarios as well for actual choices and was shown to reduce subjects′ expected outcomes.

People believe that they perform fewer health-threatening behaviors than their average peer Men such beliefs were challenged by information about the actual average behavior frequencies reported by their peers, subjects shifted their own self-views and reported engaging in these behaviors less frequently than controls did. Evidently, this biased reconstruction of their own past behavioral patterns was designed to permit subjects to maintain the belief that they were superior to their peers. This interpretation was strengthened by the finding that no shifts in self-reported behavioral frequencies occurred for subjects given inflated peer averages to which most subjects were able to view themselves as superior without biasing their self-ports.

Several studies indicate that directional questions influence self-conceptions. For example, subjects asked "are you happy with your social life?" rated themselves as happier than did subjects asked "are you unhappy with your social life?". This occurs because subjects use a positive-test strategy to test directional hypotheses about themselves: Subjects were more likely to search their memories for examples of the hypothesized characteristic than for examples of its opposite. The resulting biased set of memories biased subjects′ working self-conceptions. This effect depends on the variability of the relevant self-knowledge. Directional questions biased self-conceptions only when the self-knowledge was inconsistent enough to permit support for opposing hypotheses. When we examined relatively consistent domains of the self, or relatively consistent individuals, neither the content of recruited memories nor self-conceptions were biased by directional questions.

When stereotypes affect judgment about individuals in the presence of individuating information, they may do so by affecting the construal of that information. Therefore stereotypes may affect judgment even if perceivers subsequently neglect the stereotypes or base rates and base their impressions only on these stereotype-driven construals of the individuating information. Two experiments showed that stereotypes affected judgments about targets in the presence of the same ambiguous individuating information that was open to multiple construals, but not in the presence of the same specific disambiguated construals of that information. A third experiment showed that the effects of stereotypes on target ratings in the presence of ambiguous individuating information were mediated by the construals of this information. Thus all subjects relied predominantly on the individuating information, but when it was ambiguous, it was construed differently, depending on the stereotype.

We propose that people motivated to hold particular beliefs about another person attempt to construct justifications for these desired beliefs. In so doing, they construct general beliefs from which the desired ones may be inferred. In Study 1 subjects motivated to see a target as high or low in ability (because he was to be their partner or opponent in an upcoming game) differed as expected in their beliefs about the target and also differed in their general beliefs about the ability of the target's peers and about the causal determinants of performance at a task in which the target had been observed. In Study 2 subjects motivated to see a member of a stigmatized group, schizophrenics, positively (because they expected to interact with him) expressed more positive stereotypes of that group than less motivated subjects (who expected only to observe such an interaction). In both studies the differing general beliefs may have been constructed so as to justify different desired beliefs about the targets.

It is proposed that motivation may affect reasoning through reliance on a biased set of cognitive processes--that is, strategies for accessing, constructing, and evaluating beliefs. The motivation to be accurate enhances use of those beliefs and strategies that are considered most appropriate, whereas the motivation to arrive at particular conclusions enhances use of those that are considered most likely to yield the desired conclusion. There is considerable evidence that people are more likely to arrive at conclusions that they want to arrive at, but their ability to do so is constrained by their ability to construct seemingly reasonable justifications for these conclusions. These ideas can account for a wide variety of research concerned with motivated reasoning.

Four studies examined how people combine social concepts that have conflicting implications (e.g., Harvard-educated and carpenter). Several kinds of evidence indicated that such combinations are guided by causal reasoning that draws upon both causal relations contained within the constituent concepts and on broader world knowledge. Open-ended descriptions of members of combinations contained explicit causal descriptors, as. well as emergent attributes not used to describe members of constituents. Ratings of the likelihood that combination members possessed various attributes were not fully predicted by comparable ratings of constituents. Causal reasoning appeared to be most pervasive for combinations viewed as more surprising, suggesting that surprise may have triggered the generation of causal accounts.