Causal inference in young children: evidence preference and evidence accumulation

Two experiments examined evidence preference and evidence accumulation in causal inference among late-preschool-aged children (68–77 months), focusing on three types of evidence: informant testimony, observation of others’ experimental outcomes, and self-generated exploratory outcomes. Experiment 1 (75 recruited; 59 valid) presented the three forms of evidence in conflict to investigate children’s selection preferences and the influence of evidence order. The results showed that children were more likely to trust evidence obtained through their own exploration, and evidence order produced a significant recency effect under some conditions. Experiment 2 (98 recruited; 85 valid) presented two sources of evidence supporting one another while the third contradicted them, examining whether evidence accumulation altered endorsement. The results showed that most children selected the conclusion jointly supported by two converging sources rather than that supported by a single conflicting source. These findings indicate that late-preschool-aged children can process evidence from multiple sources, are sensitive to differences among evidence sources, and can identify and use the accumulation of converging evidence. Thus, causal inference at this developmental stage is characterized by both evidence preference and evidence integration. Educational practice should provide opportunities for self-generated exploration and comparison across multiple sources of evidence to promote scientific thinking.