Social categorization within the contact setting
One important limitation of the initial theorizing of the contact hypothesis is that it did not specify how individuals should categorize the ingroup and the outgroup during contact in order to maximize its effects (Allport, 1954). Social categorization changes individuals’ appraisal of the situation (Tajfel & Turner, 1979), and as a consequence it can determine how contact will shape outgroup attitudes. Therefore, understanding the contact effects by taking into account social categorization is important both theoretically and practically. At the theoretical level, this integration can provide insights on how to maximize contact effectiveness and identify the boundary conditions of the positive effects of contact. At the practical level, it can (a) shed light on how the contact setting should be structured for interventions to be effective and (b) relatedly inform the implementation of activities that individuals should carry out as part of interventions.
During the 80s, scholars developed models aimed at extending Allport’s (1954) contact hypothesis, starting from a consideration of the basic premises of social identity theory (SIT; Tajfel & Turner, 1979). SIT distinguishes interpersonal from intergroup behavior, with the two being positioned at the opposite ends of a continuum. At the interpersonal end of the continuum, relations between individuals are solely determined by their personal characteristics and preferences; at the intergroup end, relations between individuals are uniquely determined by group differences. The major models that considered the role of such categorization processes will be outlined later.
The decategorization model
Brewer and Miller (1984, 1988) argued that, since prejudice stems from the categorization into ingroups and outgroups, interactions between individuals belonging to different groups should take place at the interpersonal end of the continuum. In other words, individuals should interact on the basis of their personal characteristics, such as personalities and idiosyncrasies, instead of their group memberships. Repeated interpersonal interactions would lead individuals to adopt a more individuated mode of thinking, would reduce the importance attached to outgroup stereotypes and prejudice, and would use individual information as the basis of evaluation. Evidence for the model has been mainly provided by laboratory studies (e.g., Bettencourt, Brewer, Croak, & Miller, 1992; N. Miller, Brewer, & Edwards, 1985). Indirect evidence for the effectiveness of the decategorization model is also provided by studies testing the effects of cross-group friendships on prejudice- reduction, which imply intimate personal relations and key components of personalized contact, such as self-disclosure (Davies et al., 2011).
The model has two main limitations. First, relinquishing important social identities may be unrealistic or undesirable (Hewstone, 1996). Considering, for example, a visible identity such as (in some cases) ethnicity, an individual may not be able to forsake it or may not be willing to do so. Second, if decategorization is successful and individuals relate with outgroup members on the basis of their personal characteristics, they may be unable to associate the (individual) outgroup member they have contact with the outgroup category as a whole (Brown & Hewstone, 2005). In other words, decategorization would sever the link between outgroup members encountered and the larger outgroup category, thus impeding the generalization of attitudes from the individual to the outgroup as a whole (but see McIntyre, Pao- lini, & Hewstone, 2017).
The intergroup contact model
Hewstone and Brown (1986; Brown & Hewstone, 2005) argued that, being a fundamental human process that serves to provide order to the social world and to improve self-esteem (Bruner, 1957; Tajfel & Turner, 1979), social categorization cannot be eliminated. In fact, as mentioned earlier, individuals may be unwilling or unable to relinquish their group identities. Therefore, Hewstone and Brown argued that it is important that group identities are salient during contact or, in other words, that individuals from different groups interact with each other at the intergroup end of the interpersonal—intergroup continuum postulated by SIT (Tajfel & Turner, 1979). Maintaining group salience during contact allows one to recognize the other individual as a member of the outgroup and to associate him/ her with the outgroup as a whole, allowing the positive attitudes developed during contact to generalize outside the contact setting.
This model received support from experimental (e.g., Brown, Vivian, & Hewstone, 1999, Study 1), correlational (e.g., Voci & Hewstone, 2003), and longitudinal studies (e.g., Binder et al., 2009), showing that group salience is a key factor for generalization to occur, even when considering implicit attitudes (Vezzali, 2008). A main limitation of the model is that group salience per se can increase intergroup anxiety, which can be detrimental for intergroup relations (Vezzali, Capozza, Mari, & Hichy, 2007). In addition, when negative contact occurs, group salience can lead to an increase in prejudice (Graf & Paolini, 2017).
The common ingroup identity model and the dual identity model
According to Gaertner and Dovidio (2000), contact should lead individuals to perceive each other as members of a superordinate group that includes both the ingroup and the outgroup. This new common ingroup identity satisfies social categorization and allows to extend the advantages reserved to the ingroup (i.e., more positive attitudes) to - now former - outgroup members. The common ingroup identity model has received ample empirical support from experimental (e.g., Gonzalez & Brown, 2003), correlational (e.g., Eller & Abrams, 2004), and longitudinal research (e.g., Mahonen & Jasinskaja-Lathi, 2015). The positive effects of a common ingroup identity generalize to indirect forms of prejudice, such as humanity attribution and implicit prejudice (Capozza, Trifiletti, Vezzali, & Favara, 2013; Vezzali, 2008; cf. Chapter 5).
This model suffers, however, from the same limitations as the decategorization model. Specifically, individuals may not want or be able to shift their group identity, abandoning the ingroup for the sake of a superordinate group. Moreover, to the extent that outgroup individuals one has contact with are categorized based on a superordinate identity, the association between them and the larger outgroup category is severed. Therefore, categorization of known outgroup members in the superordinate category may not allow generalization of positive attitudes to the outgroup as a whole.
To address these issues, Gaertner and Dovidio (2000) conceptualized dual identity, where ingroup and outgroup categories do not fade but are maintained within a larger superordinate identity. In fact, by maintaining both - ingroup and outgroup - original identities as well as establishing a superordinate identity, the dual identity approach does not require individuals to forsake their social identities and satisfies the need for group distinctiveness (Jetten, Spears, & Postmes, 2004). Importantly, salience of original identities allows the generalization of attitudes to the whole outgroup category. There are several correlational (e.g., Gaertner, Rust, Dovidio, Bachman, & Anastasio, 1994; Stathi, Pavetich, et al., 2020) and experimental (e.g., Gonzalez & Brown, 2006) studies supporting the benefits of adopting a dual identity approach for the generalization of outgroup attitudes.