Changing Children’s Intergroup Biases and Social Stereotypes Through Statistical Counterevidence

• In Event: Novel Insights on the Development of Social Group Stereotypes and how to Change Them

Abstract

From a young age, children develop positive associations, attitudes, and stereotypes about their own social groups, and negative associations, attitudes, and stereotypes about other social groups (Baron & Banaji, 2006; Bigler & Liben, 1993; Dunham et al., 2016). Similar biases arise when children reason about minimally defined social groups (Dunham, 2018). Past interventions that exposed children to counterstereotypic exemplars effectively reduced bias and stereotyping among older children (8- to 10-year-olds), but were less effective for younger children (5- to 7-year-olds) (Gonzalez et al., 2017). In the present study, we aim to assess whether exposing 5- to 6-year-olds and 9 to 10-year-olds to counterevidence that is statistically representative of the entire social group, might change their biases and stereotypes. We hypothesized that the trait distribution in the sample would change children’s attitudes toward the group (e.g., observing a randomly drawn, mostly nice sample, would lead to more positive attitudes about the group).

US (50 5-6-year-olds and 25 9-10-year-olds) and Israeli children (48 5-6-year-olds) were randomly assigned to one of two minimally defined groups. We first measured children’s attitudes and essentialist beliefs about the ingroup (the Ingroup condition) or the outgroup (the Outgroup condition), as well as their expectations about the likelihood of drawing a nice child and a smart child from the groups. Then, participants saw a sample of 10 children, which varied in trait distribution (1 nice and 9 mean children or 9 nice and 1 mean children). Participants were told the sample was randomly drawn from one of the two groups, and were asked to guess which group the sample was from. Then, the experimenter revealed the group membership of the sample (in the Ingroup condition the sample was from the ingroup; in the Outgroup condition it was from the outgroup), and we re-measured their attitudes, essentialist beliefs, and expectations about the groups.

Children who observed the mostly nice sample were more likely to guess that the sample was from their ingroup than children who observed the mostly mean sample (US: p = .02; Israel: p = .08; Table 1). Children’s attitudes toward the groups became more negative after observing the mostly mean sample (US: p < .001; Israel: p = .04), but did not change after observing the mostly nice sample (Table 2). Children’s essentialist beliefs were not affected by the sample. US children’s expectations about the likelihood of drawing a nice child from the group decreased after observing the sample (p < .001), and their expectations decreased more after observing the mostly mean sample than the mostly nice sample (p = .07).

These initial findings suggest that statistical counterevidence can change children’s attitudes and beliefs about minimal groups. With the final sample size (N = 96 per age group per country), we will further explore whether the counterevidence has similar or different effects on children’s attitudes and beliefs about ingroup vs. outgroup.

Authors

• Rongzhi Liu, University of California - Berkeley
  Presenting Author

• Izabella Goldring, Bar-Ilan University

• Hilla Asraf, Bar-Ilan University

• Gil Diesendruck, Bar Ilan University

• Fei Xu, University of California - Berkeley