The role of trait reasoning in young children's selective trust.

In recent years, ample research has shown that preschoolers choose selectively who to learn from, preferring, for example, to learn novel words from a previously accurate over a previously inaccurate model. But this research has not yet resolved what cognitive foundations such selectivity builds upon. The present article reports 2 studies that investigate whether preschoolers' selective trust is based on global impression formation (halo effects), on the close matching of past and future behavior or on trait-based inferences, and moreover whether the cognitive strategies used are the same for both positive and negative information (high and low competent models). Four- and 5-year-old children (N = 96) were presented with 2 high-competence models (strong vs. knowledgeable; Study 1) or 2 low-competence ones (weak vs. unknowledgeable; Study 2). In 5 subsequent task groups, which required strength and knowledge to different degrees, children were asked to choose between the 2 models. Children in both studies chose models selectively in accordance with their corresponding attributes, preferring the strong (or avoiding the weak) model for strength-related tasks and preferring the knowledgeable (or avoiding the unknowledgeable) model for knowledge-related tasks. This pattern of selective model choice held only for those children who correctly identified the attributes of both models (strong, smart), as indicated by their answers to trait questions at the end of the session. This suggests that trait reasoning plays a crucial role in young children's selective social learning.