Poster #81 - “Shape bias” goes social: Children categorize people by weight rather than race

• In Event: Poster Session 15

Abstract

Children tend to categorize novel objects according to their shape rather than their color, texture, or other salient properties – a tendency known as “shape bias” (Landau, Smith, & Jones, 1988; Diesendruck & Bloom, 2003). Research on shape bias has primarily been conducted in the object domain. In the present study, we investigated whether this bias also extends to the social domain, where it should lead children to categorize people according to their weight (their body shape) rather than their race (their skin color). We aim to advance basic research on the domain-generality or -specificity of shape bias, as well as to inform research on the emergence of weight-based prejudice, which is already robust by the preschool years (for reviews, see Puhl & Latner, 2007; Paxton & Damiano, 2017). Studies were preregistered at https://osf.io/c5gn7/?view_only=7ac9b9ed287341ad852602be6b20283f.
Study 1 included 50 five- and six-year-old participants (56% F, 42% M, 2% NB; 56% White, 16% Asian, 10% Multiracial, 4% Latinx, 4% Black, 1% did not report). Participants completed a lexical categorization task (adapted from Landau et al., 1988) in which they were asked to extend a novel label for a target individual to either a weight-matched, race-matched, or distractor individual. For example, participants were told, “He is a Zarpie” and asked, “Who else is a Zarpie?” See Figure 1 for an example. Participants completed five such trials in the social domain, and then five in the object domain (“This is a Wub, which one of these is also Wub?”). We compared children’s choices to chance-level (defined conservatively as 50%; see preregistration for chance level justification). A one-sample t-test showed that on the social trials, children tended to extend the label to the weight-matched individual significantly more than 50% of the time (t(49)=6.87, p<.001, M=3.88/5 trials), a large-sized effect (Cohen’s d = .99) shown in Figure 2. While children who showed shape bias in the object domain tended also to show shape bias in the social domain, the correlation did not reach significance (p=.052).
Study 2 includes 18 five- and six-year-old participants thus far (19% F, 56% M, 5% NB; 44% White, 22% Multiracial, 22% Asian, 6% Latinx, 6% Black; target n=20). Participants completed a task (adapted from Ahn et al., 2000) in which they were asked to generalize a novel kind-relevant property to either a weight-matched, race-matched, or distractor individual. For example, participants were told, “She has Zarpie deep inside her body” and asked, “Who else has Zarpie deep inside her body?” Stimuli were identical to Study 1. Trials pertained only to the social domain. Again, results showed that children tended to extend the label to the weight-matched individual significantly more than 50% of the time (t(17)=8.61, p<.001, M = 4.5/5 trials), a large-sized effect (Cohen’s d=2.03) shown in Figure 2.
Together, results suggest that children view weight as a highly informative and kind-relevant characteristic. This view may be supported by a domain-general tendency to conceive of shape as kind-relevant. We discuss implications for cognitive development and for weight bias.

Authors

• Rebecca Peretz-Lange, SUNY Purchase

• Melissa M. Kibbe, Boston University
  Presenting Author