Too young to correct: a developmental test of the three-stage model of social inference.

The 3-stage model of social inference posits that people categorize behaviors and characterize actors or situations effortlessly, but they correct these characterizations with additional information effortfully. The current article tests this model using developmental data, assuming that the less cognitively demanding processes in the model (i.e., categorization, characterization) should appear earlier in development, whereas the more demanding correction process should not appear until later in development. Using 2 different paradigms, Studies 1 and 3 found that younger children failed to take situational information into account while characterizing the actor. Study 2 found that younger children failed to take dispositional information into account while characterizing the situation. In contrast, in these 3 studies, older children used the available information to correct their characterizations of the actors and of the situations. Consistent with the 3-stage model, during elementary school years, children start to integrate additional information when drawing explicit social inferences. In Study 4, children of all age levels used a prior expectancy to draw a dispositional inference, ignoring situational information, suggesting that characterizations based on prior expectancies about an actor are a highly efficient process, not contemplated by the model. The 4 studies together illustrate how developmental data can be valuably used to test adult socio-cognitive models, to extend their validity, or to simply further inform those models.