Thinking Counterfactually Supports Children's Evidence Evaluation in Causal Learning

Often, the evidence we observe is consistent with more than one explanation. How do learners discriminate among candidate causes? The current studies examine whether counterfactuals help 5-year olds (N = 120) select between competing hypotheses and compares the effectiveness of these prompts to a related scaffold. In Experiment 1, counterfactuals support evidence evaluation, leading children to privilege and extend the cause that accounted for more data. In Experiment 2, the hypothesis that accounted for the most evidence was pitted against children's prior beliefs. Children who considered alternative outcomes privileged the hypothesis that accounted for more observations, whereas those who explained relied on prior beliefs. Findings demonstrate that counterfactuals recruit attention to disambiguating evidence and outperform explanation when data contrast with existing beliefs.