Children’s resolution of their own and others’ beliefs in light of counterevidence

• In Event: 1-029 - The emergence of scientific reasoning skills in early childhood

Integrative Statement

Children can revise their beliefs in the face of counterevidence (e.g., Macris & Sobel 2017; Rakoczy, Ehrling, Harris, & Schultze, 2015, Young, Alibali, & Kalish 2012). But are they sensitive to the contexts in which they encounter the counterevidence? The current study explores the extent to which social context – both in terms of how an initial belief is formed and how counterevidence is presented – affects 4- to 8-year-olds’ ability to revise their beliefs.

We investigated the degree to which two factors influence children’s tendency to revise an uncertain hypothesis in light of counterevidence: (1) the source of the initial hypothesis and (2) the nature of the counterevidence provided. Four- to 8-year-olds (N=104) were presented with ambiguous data about a potential cause (what type of flowers attract bees). Participants in the Other conditions witnessed an ignorant peer endorse one of two equally plausible hypotheses, while children in the Self conditions were asked to make an endorsement themselves. Following this, children received evidence indicating that the initial hypothesis was incorrect (using a 2x2 design). In the Verbal conditions, participants were told through testimony that the initial endorsement was incorrect and that there was counterevidence in support of the alternate hypothesis. Children in the Observed conditions witnessed three pieces of counterevidence that the initial endorsement was incorrect and that the other hypothesis was correct. Following the presentation of testimony or counterevidence, children in all four conditions were given the opportunity to state their current belief about the potential causes.

The proportion of children who revised their belief (i.e., responded that both the initial endorsement was incorrect and that the alternate hypothesis was correct) in each condition is shown in Figure 1. These data were analyzed via a Generalized Linear Model assuming a binary logistic response contrasting who initially made the hypothesis (Self vs. Other) and how counterevidence was received (Verbal vs. Observed) with age as a covariate. There was a main effect of counterevidence, with children revising their belief more in the verbal conditions, Wald χ2(1) = 5.52, p = .01. There was also a 3-way interaction among these factors, Wald χ2(1) = 4.64, p = .03. As children got older, they were less likely to revise their beliefs from verbal information and more likely to revise their beliefs from observed information when it was their initial belief (Self condition). When assessing another’s belief (Other conditions), all children were likely to revise when given verbal information, but older children were more skeptical of observed data.

These results show that belief revision is not a singular process. When revising their beliefs, children integrate not only how counterevidence is generated, but also whether they are evaluating others’ information or learning on their own behalf. Evaluating others’ information is more prone to verbal counterevidence. As children develop, they become more skeptical of others’ counterevidence for their own hypotheses, but also more persuaded by observed data.

Authors

• Zoe Finiasz, Brown University
  Presenting Author

• Deena Skolnick Weisberg, Villanova University
  Non-Presenting Author

• David Sobel, Brown University
  Non-Presenting Author