TRAITEDNESS AND THE ASSESSMENT OF RESPONSE PATTERN SCALABILITY

Scalability statistics are used to measure the degree to which an examinee's item responses to a self-report measure conform to a specified pattern. This article proposes that a scalability index based on item response theory, called Z(L), offers promise for detecting individual differences in traitedness. The Z(L) index is described and applied to data from 1,000 Ss who were administered the Multidimensional Personality Questionnaire (A. Tellegen, 1982). Low split-half reliabilities were found for Z(L) index scores computed within 11 personality scales. However, further data scrutiny revealed that valuable information about an individual's trait structure can still be gleaned from a scalability analysis. It is concluded that scalability indices can be used, albeit judiciously, to explore the fit between an individual's behavior and a personality trait construct.