Journal of Personality and Social Psychology 2001, Vol. 80, No. 6, 876-893 Copyright 2001 by the American Psychological Association, Inc. 0022-3514/01/S5.00 DOI: 10.1037//0022-3514.80.6.876 Person Theories and Attention Allocation: Preferences for Stereotypic Versus Counterstereotypic Information Jason E. Plaks Steven J. Stroessner Columbia University Barnard College, Columbia University Carol S. Dweck Jeffrey W. Sherman Columbia University Northwestern University How do people respond to information that counters a stereotype? Do they approach it or avoid it? Four experiments showed that attention to stereotype-consistent vs. -inconsistent information depends on people's implicit theories about human traits. Those holding an entity theory (the belief that traits are fixed) consistently displayed greater attention to (Experiments 1 and 4) and recognition of (Experi- ments 2 and 3) consistent information, whereas those holding an incremental (dynamic) theory tended to display greater attention to (Experiment 1) and recognition of (Experiment 3) inconsistent information. This was true whether implicit theories were measured as chronic structures (Experiments 1, 2, and 4) or were experimentally manipulated (Experiment 3). Thus, different a priori assumptions about human traits and behavior lead to processing that supports versus limits stereotype maintenance. Each of us literally chooses, by his way of attending to things, what sort of universe he shall appear to himself to inhabit (James, 1890/ 1983, p. 416). We often encounter people acting in ways that are contrary to what we stereotypically expect of them (e.g., a woman who plays . baseball, a trucker who quotes Byron). On the surface, stereotype- inconsistent information appears to pose a direct challenge to the veracity of our stereotypes. However, we know both from every- day experience and from social cognition research (e.g., L. Johnston, Hewstone, Pendry, & Frankish, 1994; Maass, Salvi, Arcuri, & Semin, 1989; Weber & Crocker, 1983) that stereotypes are remarkably resistant to such challenges. This is troublesome for researchers and policy makers interested in reducing stereo- typing and prejudice, because it suggests that merely exposing people to stereotype-inconsistent information may not necessarily lead to a decrease in stereotyping. A number of researchers have examined social-cognitive pro- cesses that may contribute to the persistence of stereotypes. In typical studies, the researcher presents participants with both stereotype-consistent and stereotype-inconsistent information and then observes whether participants process the two types of infor- mation in different ways (e.g., Belmore, 1987; Hastie & Kumar, 1979; Hemsley & Marmurek, 1982; Hilton, Klein, & von Hippel, 1991; Sherman, Lee, Bessenoff, & Frost, 1998; Srull, 1981; see Stangor & McMillan, 1992). Perceivers routinely process consis- tent and inconsistent information differently, and this differential processing often appears to contribute to the preservation of per- ceivers' stereotypes (e.g., Crocker, Hannah, & Weber, 1983; Hilton et al, 1991). This line of research has led investigators to propose that per- ceivers might follow one of several processing strategies after identifying information as consistent or inconsistent with a stereo- type. These strategies can be grouped into three broad categories, two of which lead to stereotype preservation. First, perceivers may decrease their engagement with the stereotype-inconsistent infor- mation and instead focus on stereotype-consistent information, thereby accumulating in memory more confirming evidence than disconfirming evidence (e.g., Bodenhausen, 1988; Macrae, Hew- stone, & Griffiths, 1993; D. T. Miller & Turnbull, 1986; Snyder & Swann, 1978). Second, perceivers may increase their engagement with stereotype-inconsistent information to debunk it or otherwise reinterpret it in a manner that leaves the stereotype intact (e.g., Crocker et al., 1983; Devine & Baker, 1991; Lui & Brewer, 1983; Maass et al., 1989). Presumably, these two strategies are directed toward and contribute to stereotype persistence. Third, perceivers may increase their cognitive engagement with stereotype- inconsistent information because they consider such nonredundant information to possess greater informational value than consistent information (e.g., Bassok & Trope, 1984; Skov & Sherman, 1986). Presumably, such openness to stereotype-inconsistent information undermines the persistence of stereotypes. Because such processes have clear and important implications for stereotyping, researchers have attempted to uncover general rules about when people focus on inconsistent versus consistent target information (e.g., Sherman et al., 1998; Stangor & McMil- lan, 1992; Vonk, 1994). To this end, a good deal of research has been devoted to isolating the inherent properties of consistent and inconsistent information that may lead perceivers to focus on one at the expense of the other. For example, several researchers have proposed that inconsistent information is by its very nature more Jason E. Plaks and Carol S. Dweck, Department of Psychology, Co- lumbia University; Steven J. Stroessner, Department of Psychology, Bar- nard College, Columbia University; Jeffrey W. Sherman, Department of Psychology, Northwestern University. Correspondence concerning this article should be addressed to Jason E. Plaks, who is now at the Department of Psychology, Guthrie Hall 351525, University of Washington, Seattle, Washington 98195. Electronic mail may be sent to jep9@u.washington.edu.