Dixon et al. have written a provocative article, much of which I agree with. Nevertheless, I believe that they have neglected important data on the nature of contemporary prejudice and have oversimplified certain aspects of current theory and research on intergroup contact.
Dixon et al. base much of their argument on the claim that classical accounts of prejudice as a negative attitude do not do justice to the more ambivalent nature of prejudice in many societies. Dixon et al. are correct to point out that many modern manifestations of prejudice are, indeed, often more complex than mere antipathy; I acknowledged as much in the second edition of my book (Brown Reference Brown2010 – not cited in the target article). However, if we should be aware that the wolf of prejudice (towards women and some minority groups) can sometimes come disguised in benevolent sheep's clothing, it is also important that we remember that the prejudice experienced by many groups is far from “benevolent.” The two most obvious targets of overtly hostile prejudice in many Western societies are Muslims and immigrants.
In 2008, the percentages of nationally representative samples reporting unfavourable opinions of Muslims ranged from 23% (United Kingdom) to 52% (Spain) (Pew Research Center 2008). Perhaps not unconnected with this Islamophobia, the number of racially or religiously aggravated offences recorded in England and Wales totalled over 33,000 for 2008/2009 (Chaplin et al. Reference Chaplin, Flatley and Smith2011). Immigrants are another group that is frequently the target of unfavourable attitudes. Semyonov et al. (Reference Semyonov, Raijman and Gorodzeisky2006) reported how anti-foreigner sentiment in 12 European countries showed substantial increases in the 1990s: in 1988, between 10% and 30% of national samples agreed with such items as, “the presence of foreigners is one of the causes of delinquency and violence”; by 2000, those proportions had risen to between 40% and 70%. Similarly, McLaren (Reference McLaren2003) showed that attitudes about the expulsion of legally established immigrants in the 1997 Eurobarometer survey ranged from a low of 2.80 to a high of 3.73, where 3.0 was the midpoint and 5.0 the unfavourable pole of the scale.
These data do not paint a picture of Muslims and immigrants as the targets of benevolent prejudice. Rather, members of these groups are often viewed and treated in an overtly negative fashion. For them, at least, conceiving of prejudice as antipathy seems to be entirely appropriate. And, it seems to me, a reasonable goal of progressive policy making would be to get members of the majority group, not necessarily to like those stigmatised groups more, as Dixon et al. suggest, but at least to hate them a little less. Properly designed contact interventions might be one tool (among several) that can assist in the achievement of that goal (Brown & Hewstone Reference Brown, Hewstone and Zanna2005; Pettigrew & Tropp Reference Pettigrew and Tropp2011). Dixon et al. portray prejudice researchers as advocating contact as a major agent of social change. This is something of an oversimplification. Ever since Pettigrew and Tropp's (Reference Pettigrew and Tropp2006) meta-analysis of the contact-prejudice relationship, it has been clear that the correlation between contact and prejudice reduction is a modest one (typically between .2 and .3). It follows that most of the variance in prejudice must be accounted for by other factors. And some of us have been explicit about this (Brown Reference Brown2010, p. 279; Pettigrew & Tropp Reference Pettigrew and Tropp2011, p. 216).
Dixon et al. also oversimplify when they characterise all contact theories as having the goal of dissolving group boundaries. Although such ideas are to be found in some perspectives (e.g., Brewer & Miller Reference Brewer, Miller, Miller and Brewer1984; Gaertner & Dovidio Reference Gaertner and Dovidio2000), they have not gone uncontested. Theoretically, such assimilationist models cannot easily deal with the problem of generalisation. And, pragmatically, intervention policies based on them are unlikely to win much acceptance from minority groups and so will have little chance of success. It is for that reason that some of us have long advocated a “dual identity” approach to contact interventions, in which a deliberate attempt is made not to dissolve group boundaries completely, either physically or psychologically (Brown & Hewstone Reference Brown, Hewstone and Zanna2005).
Dixon et al. castigate contact researchers as being reductionist in their hope to change individuals' hearts and minds. However, the charge of individualism overlooks one key component of the contact hypothesis and fails to recognise the contribution of one of its extensions, the extended contact hypothesis (Wright et al. Reference Wright, Aron, Mclaughlin-Volpe and Ropp1997). One of Allport's (Reference Allport1954) necessary conditions for contact was that there be “social and institutional support” for that contact. One reason why institutional support is important is that it creates a social climate in which egalitarian relationships between members of different groups are seen as the norm. Similarly, one of the mechanisms thought to underlie extended contact is that it helps to generate new in-group norms about the acceptability of such intergroup relationships (De Tezanos Pinto et al. 2010; Pettigrew et al. Reference Pettigrew, Christ, Wagner and Stellmacher2007). Such a normative explanation of behaviour change is far from the individualistic presentation that Dixon et al. give of contact research.
Finally, by their own admission, Dixon et al. pay little attention to intergroup emotions in their article. This was unfortunate because one of the significant contributions of this work has been to underline the variegated nature of people's feelings towards other groups and how these feelings are contingent on the specific nature of intergroup relationships (Smith Reference Smith, Mackie and Hamilton1993). An important lesson from research on group-based emotions is that we need to move away from a simple characterisation of intergroup orientations in terms of “like” or “dislike.” Emotions such as fear, disgust, hostility, guilt, and shame can be evoked in particular intergroup contexts, and these will have very different implications for people's subsequent behaviour (or behavioural intentions) towards out-groups (Brown et al. Reference Brown, Gonzalez, Zagefka, Manzi and Čehajić2008; Cottrell & Neuberg Reference Cottrell and Neuberg2005; Rees et al., in press). Understanding the conditions that elicit – or inhibit – these different emotions is a significant challenge for social psychology, and is one that takes us far beyond the questions of “negative evaluations” and “liking one another” that are the main foci of Dixon et al.'s critique.
You have
Access
Dixon et al. have written a provocative article, much of which I agree with. Nevertheless, I believe that they have neglected important data on the nature of contemporary prejudice and have oversimplified certain aspects of current theory and research on intergroup contact.
Dixon et al. base much of their argument on the claim that classical accounts of prejudice as a negative attitude do not do justice to the more ambivalent nature of prejudice in many societies. Dixon et al. are correct to point out that many modern manifestations of prejudice are, indeed, often more complex than mere antipathy; I acknowledged as much in the second edition of my book (Brown Reference Brown2010 – not cited in the target article). However, if we should be aware that the wolf of prejudice (towards women and some minority groups) can sometimes come disguised in benevolent sheep's clothing, it is also important that we remember that the prejudice experienced by many groups is far from “benevolent.” The two most obvious targets of overtly hostile prejudice in many Western societies are Muslims and immigrants.
In 2008, the percentages of nationally representative samples reporting unfavourable opinions of Muslims ranged from 23% (United Kingdom) to 52% (Spain) (Pew Research Center 2008). Perhaps not unconnected with this Islamophobia, the number of racially or religiously aggravated offences recorded in England and Wales totalled over 33,000 for 2008/2009 (Chaplin et al. Reference Chaplin, Flatley and Smith2011). Immigrants are another group that is frequently the target of unfavourable attitudes. Semyonov et al. (Reference Semyonov, Raijman and Gorodzeisky2006) reported how anti-foreigner sentiment in 12 European countries showed substantial increases in the 1990s: in 1988, between 10% and 30% of national samples agreed with such items as, “the presence of foreigners is one of the causes of delinquency and violence”; by 2000, those proportions had risen to between 40% and 70%. Similarly, McLaren (Reference McLaren2003) showed that attitudes about the expulsion of legally established immigrants in the 1997 Eurobarometer survey ranged from a low of 2.80 to a high of 3.73, where 3.0 was the midpoint and 5.0 the unfavourable pole of the scale.
These data do not paint a picture of Muslims and immigrants as the targets of benevolent prejudice. Rather, members of these groups are often viewed and treated in an overtly negative fashion. For them, at least, conceiving of prejudice as antipathy seems to be entirely appropriate. And, it seems to me, a reasonable goal of progressive policy making would be to get members of the majority group, not necessarily to like those stigmatised groups more, as Dixon et al. suggest, but at least to hate them a little less. Properly designed contact interventions might be one tool (among several) that can assist in the achievement of that goal (Brown & Hewstone Reference Brown, Hewstone and Zanna2005; Pettigrew & Tropp Reference Pettigrew and Tropp2011). Dixon et al. portray prejudice researchers as advocating contact as a major agent of social change. This is something of an oversimplification. Ever since Pettigrew and Tropp's (Reference Pettigrew and Tropp2006) meta-analysis of the contact-prejudice relationship, it has been clear that the correlation between contact and prejudice reduction is a modest one (typically between .2 and .3). It follows that most of the variance in prejudice must be accounted for by other factors. And some of us have been explicit about this (Brown Reference Brown2010, p. 279; Pettigrew & Tropp Reference Pettigrew and Tropp2011, p. 216).
Dixon et al. also oversimplify when they characterise all contact theories as having the goal of dissolving group boundaries. Although such ideas are to be found in some perspectives (e.g., Brewer & Miller Reference Brewer, Miller, Miller and Brewer1984; Gaertner & Dovidio Reference Gaertner and Dovidio2000), they have not gone uncontested. Theoretically, such assimilationist models cannot easily deal with the problem of generalisation. And, pragmatically, intervention policies based on them are unlikely to win much acceptance from minority groups and so will have little chance of success. It is for that reason that some of us have long advocated a “dual identity” approach to contact interventions, in which a deliberate attempt is made not to dissolve group boundaries completely, either physically or psychologically (Brown & Hewstone Reference Brown, Hewstone and Zanna2005).
Dixon et al. castigate contact researchers as being reductionist in their hope to change individuals' hearts and minds. However, the charge of individualism overlooks one key component of the contact hypothesis and fails to recognise the contribution of one of its extensions, the extended contact hypothesis (Wright et al. Reference Wright, Aron, Mclaughlin-Volpe and Ropp1997). One of Allport's (Reference Allport1954) necessary conditions for contact was that there be “social and institutional support” for that contact. One reason why institutional support is important is that it creates a social climate in which egalitarian relationships between members of different groups are seen as the norm. Similarly, one of the mechanisms thought to underlie extended contact is that it helps to generate new in-group norms about the acceptability of such intergroup relationships (De Tezanos Pinto et al. 2010; Pettigrew et al. Reference Pettigrew, Christ, Wagner and Stellmacher2007). Such a normative explanation of behaviour change is far from the individualistic presentation that Dixon et al. give of contact research.
Finally, by their own admission, Dixon et al. pay little attention to intergroup emotions in their article. This was unfortunate because one of the significant contributions of this work has been to underline the variegated nature of people's feelings towards other groups and how these feelings are contingent on the specific nature of intergroup relationships (Smith Reference Smith, Mackie and Hamilton1993). An important lesson from research on group-based emotions is that we need to move away from a simple characterisation of intergroup orientations in terms of “like” or “dislike.” Emotions such as fear, disgust, hostility, guilt, and shame can be evoked in particular intergroup contexts, and these will have very different implications for people's subsequent behaviour (or behavioural intentions) towards out-groups (Brown et al. Reference Brown, Gonzalez, Zagefka, Manzi and Čehajić2008; Cottrell & Neuberg Reference Cottrell and Neuberg2005; Rees et al., in press). Understanding the conditions that elicit – or inhibit – these different emotions is a significant challenge for social psychology, and is one that takes us far beyond the questions of “negative evaluations” and “liking one another” that are the main foci of Dixon et al.'s critique.