In two studies, subjects ®lled out a questionnaire requiring them to choose between internal and external explanations of desirable or undesirable events. They were also asked to ®ll out the same questionnaire from another person's point of view, either a member of the ingroup or a member of the outgroup. The ®rst study used students as subjects and the events were about student life. The second study used employees for subjects and the events were about the working world. As predicted on the basis of internality norm theory, internal explanations were generally found to be chosen more often when the respondent or imagined respondent was said to be a worthy person. The internality scores were the highest for oneself and for the ingroup member, regardless of the desirability of the events; they were the lowest for the outgroup member. These internality attribution eects do not seem to stem from the well-known ingroup-favouring attributions (ultimate attribution error). Hence, the internality norm construct, irrespective of event desirability, provides a new pathway for exploring the evaluative eects of intergroup relations.