Children's Generalizations of Group Characteristics: Effects of Generic Versus Statistical Language

• In Event: Poster Session 08
  In Poster Session: PS 08 Topic - Cognitive Processes

Abstract

How does language impact children’s understanding of group characteristics? Generic and statistical statements are used daily to describe both social and non-social groups, with generic statements expressing a belief about a category as a whole, while statistical statements express a belief about a certain number or proportion of the members of a category. Generic statements are often used in everyday vernacular, even when they only carry a weak relation to the statistical criteria necessary to be classified as truth (Bian & Cimpian, 2017). These can often result in the propagation of harmful racial or gender stereotypes: for example, the phrases “Asians are smart” and “women are emotional” feature generic statements. The purpose of the current study is to explore how such generics focusing on social and non-social groups impact the inferences children make about characteristics of individual members of that group, particularly compared to statements featuring statistical language.
Currently, data collection is ~50% underway. Four-to-eight-year-old participants (N = 38 out of 80) were introduced to a novel group and 8 characteristics about that group in one of four between-subject conditions in a 2x2 design (Table 1). Half of the children were introduced to a fictional social group; the remaining half were introduced to a novel non-social group. Additionally, within each group type, the language used to describe the group either consistently used generic language (e.g., “Sumas like to run”), or statistical language (e.g., “half of Sumas like to run”). After being presented with the characteristics, participants were invited to choose whether or not to assign them to an individual group member (max = 8).
A power analysis (G*Power 3.1 software; Faul, Erdfelder, Lang, & Buchner, 2007) indicates that we are not yet able to test for all main effects and interaction effects of a medium size or smaller (e.g., age, group type). However, preliminary analyses on approximately half of the final sample suggests that there is a significant effect of language on children’s attributions of group characteristics. Specifically, independent samples t-tests indicate that children in the statistical language conditions (M = 4.77 out of 8, SD = 1.97) are significantly less likely than those in the generic language conditions (M=6.94 out of 8, SD = 1.39) to say that a specific group member possesses the characteristic t(36) = -3.76, p< 0.001 (Figure 1). Moreover, one-sample t-tests indicate that children in the generic language conditions attribute group characteristics to individual group members at rates significantly greater than chance t(15) = 19.98, p< 0.001. This is not the case for children in the statistical language condition.
Preliminary findings suggest that using statistical language to describe groups could reduce the likelihood that children will generalize group characteristics to individual group members. Moreover, findings could imply there are ways to describe group characteristics that lead children to create fewer generalizations about group members. By understanding how children interpret statistical language compared to generic language, we may better understand how stereotypes are formed, and how they can be avoided.

Authors

• Mercedes Munoz, Boston University
  Presenting Author

• Sarah Suárez, University of Minnesota
  Non-Presenting Author

• Sona Kumar, Boston University
  Non-Presenting Author

• Kathleen H. Corriveau, Boston University
  Non-Presenting Author