| Abstract: | “Actor” subjects engaged in a brief getting-acquainted conversation while “observer” subjects watched. Ratings of and causal attributions for behavior during the conversation were obtained immediately upon its conclusion and three weeks later from the same subjects and from an independent group of subjects not asked for immediate attributions. Comparison between immediate and the independent, “postponed” assessments showed that attributions by actors and observers tended to emphasize situational factors more over time, and personal factors less. Attributions by the “repeated assessment” subjects did not show a significant change over time, but their delayed attributions were significantly less likely to emphasize situational factors than were the delayed attributions of the “postponed assessment” group. The relative accuracy of these attributions could not, of course, be directly assessed. However, valid attributions presumably must be based upon valid perceptions or memories of the behaviors they seek to explain, and reliable criterion ratings of behaviors in the present study could be obtained from the experimenters. Actors' and observers' delayed behavior ratings agreed less with these criteria than did their immediate ratings, an indirect indication that attributions do not increase, and may decrease, in accuracy over time. |
