Two Caveats for the Use of Ideal Point Items: Discrepancies and Bivariate Constructs

摘要

Ideal point items in general and the spe-cific item response theory-based models outlined by Drasgow, Chernyshenko, and Stark (2010) hold considerable promise for the measurement of intermediate levels of constructs such as beliefs, traits, and atti-tudes. The aim of this commentary is to outline two caveats relating to the use of a certain type of ideal point item: one related to the assessment of discrepancies (e.g., person–environment fit) using ideal point items and one related specifically to possible computer adaptive testing appli-cations. Ideal point items for attitude or trait measures as outlined by Drasgow et al. appear to take one of the two forms. The first (assuming Likert items) is a statement of the general form "I have moderate levels of construct X," whereas the second takes the general form of "I have equal amounts of A and B" or "I am characterized by a balance between A and B," where A and B represent the two endpoints of a bipolar scale. This commentary focuses on this sec-ond possible form for an ideal point item, hereafter referred to as a double-barrel ideal point item.