Abstract
A method is presented which unconfounds the intensity of trait ascriptions from the perceived variability of trait-related behaviors. In accordance with the actor-observer hypothesis, 56 individuals judged their friends' behaviors as more consistent than their own and attributed higher trait intensities to themselves than to their friends. This result was obtained from trait-ascription data gathered from two independent groups of actor and observer subjects. It was only partly supported by dependent descriptions of self and friend, suggesting that in this case, additional judgmental processes of social comparison may be involved.