Exploring the nature of evaluative person descriptors through scale development

Research regarding the nature of evaluative person descriptors has been hampered because existing measures are limited. Moreover, recent work has identified potential facets of evaluative dimensions, but practical measures of these do not yet exist. Our goals were to (a) develop a measure—the Evaluative Person Descriptors Questionnaire (EPDQ)—that models both higher- and lower-order components of such descriptors, and (b) explore their nature through analyses of temporal stability, self-informant agreement, and convergent and discriminant validity. The EPDQ—which was developed following four rounds of data collection and analyses—taps three broad evaluative dimensions (Positive Valence [PV], Depravity, and Oddity) as well as four lower-order facets of PV. Results suggested that the EPDQ yielded temporal stability consistent with markers of the Big Five model. Self-informant agreement was more variable, with PV and its facets yielding lower agreement than Depravity and Oddity. We discuss implications for structural personality theory and research.