A longitudinal study of the actor-observer attributional effect

Abstract

This study examined the operation of the Jones-Nisbett actor-observer effect—i.e. the tendency for persons to see their actions as due to situational causes while observers of the same actions ascribe them to the actor's personal dispositions—in the natural, non-laboratory environment of everyday interactions over an extended period of time. Subjects were asked to provide trait descriptions of themselves and a close acquiantance over a 3 1 2 week period by choosing five adjectives each, for the self and the other person, from a large list comprised of 200 words which had been rated high on meaningfulness from Anderson's ( Journal of Personality and Social Psychology, 9, 272–279, 1968) list of 555 words scaled for likeableness, on 4 days each week. We assumed that the use of a relatively homogeneous set of trait-adjectives would reflect a more dispositional,trait-like, attribution since it describes the person as relatively consistent over time and, presumably, situations. Conversely, the use of a more heterogeneous set of adjectives would represent attributions to the diversity of situations that the person would likely find himself or herself in over the duration of the study. We thus expected the actor-observer effect to manifest itself in greater trait variability in descriptions of the self than of an acquaintance. Neither a measure of evaluative consistency nor of descriptive consistency yielded greater variability in self-descriptions than in other-descriptions. Because individual differences in a tendency toward the actor-observer attributional effect was a salient feature of our findings, we explored a possible personality correlate of this tendency—self-esteem—and found that those who conformed more to the effect were lower in self-esteem than those conforming to it less. Also, on the measure of descriptive consistency, we found variability in descriptions of the self and of the other person to be positively correlated. The implications of our findings regarding the limiting conditions of the Jones-Nisbett effect were discussed.