Reasons for skepticism about the efficacy of simulated social contact interventions.

Crisp and Turner (May–June 2009) argued that imagining intergroup interactions reduces intergroup prejudice. They argued that the procedure is remarkably effective, with “significant potential application for policymakers and educators seeking to promote tolerance for social diversity” (p. 238). We believe that such interventions, although appealing to many individuals, are problematic and that the authors’ conclusions are overly optimistic. We believe that simulated contact interventions are highly unlikely to produce meaningful attitude change. Simulated contact interventions are appealing, we suspect, in large part because they require so little effort or cost. Consider the case of anti-African-American prejudice in the United States. Simulated contact interventions do not require individuals to live in racially integrated neighborhoods or to send their children to integrated schools. Such interventions do not require financial or emotional investment in communities of color or require individuals to learn about and appreciate African American history or culture. Such interventions do not require the forging of actual bonds of friendships with African Americans, nor do they require individuals to imagine themselves seeking social contact with African Americans who differ from themselves in dress, manner, or values. Simulated contact interventions demand of prejudiced persons only that they pretend that they would engage in a one-time conversation with a single outgroup member who inhabits their same social world and actively seeks contact with them. Such efforts are shallow in our view. The appeal of this form of intervention may have colored the assessment of its effectiveness. We believe that there is little reason to be optimistic about the ability of simulated contact to reduce prejudice meaningfully. The first basis for skepticism concerns past work. The literature on social stereotyping and prejudice reduction is enormous; it spans multiple countries, disciplines, and decades. If a single point might be gleaned from this vast literature, it is that prejudice reduction is exceedingly difficult to achieve and maintain via shortterm intervention. The literature is littered with unsuccessful attempts and includes studies that have failed to find effects of extended contact (Feddes, Noack, & Rutland, 2009). A second basis for skepticism is that interventions often produce effects among only a subsample of participants (see Bigler, 1999). This is true of our own published interventions and is evidenced in studies of actual and simulated contact (e.g., Liebkind & McAlister, 1999; Stathi & Crisp, 2008). For example, in Study 1 of Turner, Crisp, and Lambert’s (2007) article, participants in the simulated contact group were significantly less biased than participants in the control group. The standard deviations reveal, however, that there was considerable variability in the simulated contact condition (SD 3.40) compared with the control condition (SD 1.66). This suggests that a significant portion of the simulated contact participants ended the procedure with intergroup bias levels equal to or greater than those of participants in the control condition. A third basis for skepticism is that the intervention literature relies heavily on self-report measures of attitudes. Crisp and Turner (2009) claimed that simulated contact has the strong potential to induce individuals to engage in cross-group contact. Nonetheless, these interventions typically fail to include even a single measure of intergroup behavior. Cameron, Rutland, Brown, and Douch (2006) did, however, examine the effect of intervention on “intended behavior.” They asked children how much they would like to interact with a refugee child and reported no significant effect of treatment condition. A fourth basis for skepticism is that when successful interventions include multiple measures of prejudice, they routinely produce significant effects for only a subset of the measures employed. Results reported by Turner et al. (2007) are illustrative. In Study 1, the authors examined the ingroup and outgroup preferences of participants who had been assigned to imagined contact versus control conditions. Participants were instructed to imagine interacting with an elderly person, yet there was no significant effect of condition on participants’ attitudes toward interacting with an elderly person; instead, there was an effect of treatment condition on interest in interacting with a young person. A fifth basis for skepticism is that, as acknowledged by Crisp and Turner (2009), the effects of simulated social contact have never been shown to be lasting. Outcome measures are typically given immediately (Stathi & Crisp, 2008) or within days of the intervention (Cameron & Rutland, 2006; Cameron et al., 2006; Turner et al., 2007). The broad intervention literature indicates that longterm assessments typically show decreases in effects. We expect the same to be true of imagined contact interventions. A sixth reason for skepticism is that the statistically significant effects reported in intervention studies are often practically unimportant. Participants’ attitudes rarely approximate nonbiased attitudes at posttest. Crisp and Turner (2009) cited Liebkind and McAlister (1999) as having reduced prejudice among Finnish children, but their intervention and control groups showed bias scores of 55.29 and 55.51, respectively, on a scale of 17 to 85, with a midpoint of 51. Stathi and Crisp (2008) reported that their use of an imagined contact intervention (Study 1) significantly increased the number of positive traits participants perceived themselves to share with the outgroup; the control and experimental groups projected only 2.60 and 3.68 traits, respectively, of 10 possible onto the outgroup. Furthermore, we are unconvinced that the small effects of imagined contact interventions are not the product of demand characteristics. Simulated social contact interventions are highly transparent in that they explicitly promote intergroup contact. If participants in these studies had been asked to estimate the experimenters’ outgroup attitudes (rather than the hypothesis of the study), we predict that they would have reported the experimenters to have positive views. Thus, subtle pressure to match the experimenters’ views appears present. In sum, we are deeply skeptical about the ability of imagined contact interventions to reduce prejudice. A single, brief, imaginary encounter with an outgroup member is unlikely to reverse or erase the psychological mechanisms that create prejudice.