Developmental differences in the use of distinctiveness, consensus, and consistency information for making causal attributions.

The present study examined developmental differences in the use of distinctive-ness, consensus, and consistency information for making causal attributions. First, third and sixth graders and college students were presented with brief story pairs consisting of an act manifested by an agent toward a target person. Each story in a pair was accompanied by a different level of a particular type of information (e.g., high consensus for one and low consensus for the other). Subjects were asked to make causal inferences about both the agents and the targets. The results revealed significant age-related differences in the ability to use each type of information in the manner predicted from Kelley's causal attribution model. Young children's use of distinctiveness information yielded the predicted agent attributions significantly more often than it yielded the predicted target attributions, while the reverse was true for consensus information. These findings were interpreted in terms of the causal principles involved: Information was used in the predicted manner at a younger age when a covariation principle was required than when a discounting principle was required.