This chapter reviews the extensive literature on bias in favor of in-groups at the expense of out-groups. We focus on five issues and identify areas for future research: (a) measurement and conceptual issues (especially in-group favoritism vs. out-group derogation, and explicit vs. implicit measures of bias); (b) modern theories of bias highlighting motivational explanations (social identity, optimal distinctiveness, uncertainty reduction, social dominance, terror management); (c) key moderators of bias, especially those that exacerbate bias (identification, group size, status and power, threat, positive-negative asymmetry, personality and individual differences); (d) reduction of bias (individual vs. intergroup approaches, especially models of social categorization); and (e) the link between intergroup bias and more corrosive forms of social hostility.

Distinctions are made between global and specific, personal and social, and trait and state self-esteem, and these are used to structure a review of over 40 studies concerning social identity theory's hypothesis that (a) intergroup discrimination elevates self-esteem and (b) low self-esteem motivates discrimination. It is observed that researchers have tended to employ measures of global personal trait self-esteem in their investigations of this self-esteem hypothesis, and it is argued that measures of specific social state self-esteem are more consistent with social identity theory's assumptions. Although no convincing evidence is found for the self-esteem hypothesis in its full and unqualified form, it is argued that this is due to a lack of specificity in its formulation and it is suggested that a more qualified and specific version of the hypothesis may be more appropriate.

Drawing from the intergroup contact model and self-categorization theory, the authors advanced the novel hypothesis of a valence-salience effect, whereby negative contact causes higher category salience than positive contact. As predicted, in a laboratory experiment of interethnic contact, White Australians (N = 49) made more frequent and earlier reference to ethnicity when describing their ethnic contact partner if she had displayed negative (vs. positive, neutral) nonverbal behavior. In a two-wave experimental study of retrieved intergenerational contact, American young adults (N = 240) reported age to be more salient during negative (vs. positive) contact and negative contact predicted increased episodic and chronic category salience over time. Some evidence for the reverse salience-valence effect was also found. Because category salience facilitates contact generalization, these results suggest that intergroup contact is potentially biased toward worsening intergroup relations; further implications for theory and policy making are discussed.