Actor and observer perceptions of the learning of a task

Abstract

The hypothesis that observers would attribute more generality to an observed sample of behavior than would actors was tested in four related experiments. The context involved learning the manual alphabet for the deaf. Actors were given a short learning period, a test, and feedback veridical with their actual performance. Observers watched the actor learn, via a one-way mirror. Procedural variations included the use of a confederate as the actor (learner), a design in which pairs of subjects were randomly assigned the observer and actor roles, and a replication in which the visual orientation of observers was shifted from the learner to the task. Dependent variables included judgments of the generality of the observed performance, predictions of the learner's score on a second alphabet test, and items referring to the learner and task. The results were strongly consistent with the hypothesis that observers attribute more generality to the acts of others than do actors themselves. Certain consistent divergences in the intercorrelations among judgments of actors and observers were noted, suggesting another facet of the actor-observer difference in perspective. There was only partial support for interaction effects related to the hypothesis that actors and observers would differ in the extent to which their inferences of ability and task difficulty would be in correspondence with performance feedback.