Social cognition and social perception.

Abstract

constructs that represent them, no matter what the person's process­ ing goals during priming. Bargh & Pietromonaco ( 1982, Bargh et al 1986) showed that trait adjectives prime relevant trait constructs even when those adjectives are presented subliminally so that subjects are unaware of even the presence of the priming stimuli. Once a social construct is activated by environmental data, however, it constitutes a theory-driven influence on the interpretation of subsequent environmental events. But the accessible con­ struct will only have this influence if it is applicable (Le. related) to the data. Increases in the accessibility of available constructs in memory can also be the result of internal sources, such as one's current goals or needs (Bruner 1957, Higgins & King 1981 , Wyer & Srull 1981 , Srull & Wyer 1986). SOCIAL COGNmON AND PERCEPTION 375 Potentially the most influential of these top-down determinants of accessibility is the frequency with which one has used a certain construct to process social input (Higgins et al1982, Higgins et al1985, Bargh et al1986, Wyer & Sru11 1986). Frequent and consistent use of a construct, perhaps in response to constancies in one's particular social environment, appears to result in a chronic likelihood of using that construct across all situations. Higgins et al (1982) and Bargh and Thein (1985) have demonstrated that when people form an impression of a target person, they are more sensitive to information that is relevant to their chronically accessible constructs than that which is relevant to other available constructs. Moreover, people's greater sensitivity to stimuli related to chronically accessible constructs does not depend on that construct being primed or preactivated (Higgins et al 1982, Bargh et al 1986). And such stimuli are more likely than others to be de­ tected in the environment in the flrst place ( Bargh 1982, Bargh & Pratto 1986). Salience effects are a related example of the interplay between data-driven and theory-driven influences. Features of the current environmental context may activate certain temporary expectancies for the situation, so that people or events that do not flt these expectancies are more distinctive or salient and receive greater attention. Moreover, the resultant focus of attention on those particular features that make the person or event salient appears to result in the activation of other abstract representations in memory of which those features are a component, such as stereotypes. Such easily discriminated features as a person's sex, race, or age are particularly likely to be salient features of a social situation, especially when those features are in the contextual minority (Taylor et al 197 8 , Hamilton 1979, Ashmore & DelBoca 1981 , Deaux & Lewis 1984, McCann et al 1985). McArthur & Friedman (1980), for ex­ ample, varied the age, sex, and race composition of groups of stimulus persons and found that subjects' stereotypic beliefs influenced the overall positivity of their trait ratings for a group member only when that target person was in the contextual minority (see also Hamilton & Rose 1980). It would seem, therefore, that theory-driven influences on the encoding and interpretation of social information can themselves be determined by data­ driven means ( see Higgins & Lurie 1983, Bargh 1984). The interpretation and meaning given to a person's behavior and to social events are in large part a function of the relative accessibilities of relevant social constructs and stereotypes, which in turn are partly a function of features of the recent and current social environment. Moreover, there is evidence suggesting that stereotype activation by the contextual salience of stereotype-consistent in­ formation occurs under memory set conditions (Mills & Tyrrell 1983). The effects of salient information on the likelihood of stereotypic influences on encoding, therefore, is apparently not limited to occasions in which people have the goal of understanding the behavior of others. As Fiske et al (1982)