Children’s belief revision during natural exploration in a children’s museum

• In Event: Belief revision in a social world: How other people shape children’s ability to revise beliefs

Abstract

Belief revision involves making inferences from prior knowledge and observed evidence. This process has mostly been studied in the laboratory (e.g., Macris & Sobel, 2017; Young, Alibali & Kalish, 2012) or in terms of its application to classroom environments (e.g., Chi, 2008; Ohlsson, 2008). Some investigations consider children’s belief revision through exploratory play (e.g., Cook, Goodman, & Schulz, 2011; Legare, 2012; Schulz & Bonawitz, 2007). Such measures demonstrate that children can change their beliefs but do not show that children are exploring with the purpose of learning new information.

Our goal is to study children’s belief revision as they explore a naturalistic environment. We recorded the behaviors of 6-9-year-olds (N=60, Mage = 94.25 months, 35 girls, 25 boys) as they interacted with a 1000sq foot museum exhibit based on spatial thinking with 17 different stations for exploration and play. Children wore an embedded GoPro camera, which allowed us to view their first-person perspective, and were recorded from a third-person perspective.

For each station, we defined a priori what goals children might construct, whether those goals were based on the museum’s intentions, and what actions might constitute revisions of beliefs they might form to meet those goals (see Table 1).

Children constructed their own goal at a station 64% of the time, significantly more than chance, Binomial test, p < .001. Age was unrelated to whether children constructed their own goal. Children engaged in more belief revision behaviors at a station when they were accomplishing goals that they had created for themselves (M = 1.73), as opposed to using exhibits in the way the museum intended (0.65), Wald X2(1) = 4.06, p = .04. This speaks to the autonomy of children’s belief revision in their own actions. However, children also engaged in more belief revision actions when they established their beliefs by initially watching others interact at the station (3.14) or being told by their grown-up prior to playing at the station (3.91), as opposed to freely exploring without guidance (0.72), Wald X2(2) = 23.06, p < .001. This is consistent with the benefits of guided exploration over free play.

Further, after children’s play, they watched the Go-Pro video of their own exploration, and were asked to reflect on what was happening in during their play. Children who engaged in more belief revision activities used more mental state terms (e.g., “think”) in their reflection, controlling for age, Wald X2(1) = 4.60, p = .03. This is consistent with hypothesis that children’s belief revision in situ potentially relates to their metacognitive understanding of why they are performing such behaviors during their play.

Overall, these data suggest that children naturally engage in belief revision, particularly when they construct their own goals, and have the metacognitive capacity to reflect on that belief revision when prompted to do so.

Authors

• David Marc Sobel, Brown University
  Presenting Author

• Laura Stricker, Brown University
  Non-Presenting Author

• Deena Skolnick Weisberg, Villanova University
  Non-Presenting Author