Abstract
Denis J. Hilton, Richard H. Smith, and Sung Hee Kim proposed in 1995 that making dispositional inferences requires the use of John Stuart Mill's method of agreement. However, their data were based on "incomplete"information that may have obscured the use of Mill's method of difference. Consistent with this suspicion, a study with "complete" information revealed the joint operation of both Mill's methods. Dispositions to the actor required not only low distinctiveness but also low consensus, whereas dispositions to the stimulus required not only high consensus but also high distinctiveness. In contrast, causal explanations to the person, stimulus, or occasion were determined by the method of difference only.