Abstract
Traditionally, scale reliability and item homogeneity are viewed as intrinsic properties of particular scales and items, and not as intrinsic to the content domain being measured. This paper shows that the reliability and item homogeneity of personality scales is in part dependent on the content domain being sampled, and that this "characteristic reliability" cannot be explained as a function of item ambiguity or scale length. Elsewhere we have suggested that responses to personality items are mediated by more abstract perceptions of the self (the "self-concept hypothesis"). We here suggest that the clarity or ambiguity of different self-perceptions may be a major determinant of item homogeneity and scale reliability. Arguing that self-perception is a special case of person-perception, we show that level of ambiguity in person perception (using different content dimensions of social judgment) is closely related to the level of scale reliability characterizing the same content domains. These results support the importance of self-concept processes in understanding what personality scales measure, and suggest the need for more commerce between self-theory and the technical concerns of personality psychometrics.
Get full access to this article
View all access options for this article.
