At a very basic level, a situational judgment test (SJT) is a series of situations and associated behaviors relevant to each situation. Although a sizable body of research relevant to SJTs has accrued, little is known about how properties of situations and behaviors, as fundamental units in SJT design, are related to properties of SJTs in terms of the information provided by scores at the item-level. In this study, theory and empirical research relevant to situations, interactionism, and trait activation provide a foundation for the argument that situational and behavioral characteristics would explain item-level variability in relationships with external variables, namely other individual difference characteristics and criterion-related validities. Ninety items from three SJTs were coded with regard to item stem situational cues and factor-five model personality trait expression associated with the response options. Mixed support was found for the study's assertions in analyses pertaining to two types of SJT scores (stem-level scores and response option-level scores). Response option personality trait expression was significantly related to response option correlations with like personality characteristics. Further, models predicting item-level correlations for both stem scores and response option-level scores explained a respectable proportion of between-item variability in correlations with external variables. Finally, there was strong evidence that the effect of response option trait expression clustered around or varied significantly across the item stems within which the response options were nested. However, results were inconsistent with regard to the effect of situational characteristics on stem score-level correlations with external variables. Additionally, results pertaining to interactions between situational characteristics and behavioral characteristics in predicting response option-level correlations with external variables were mixed. The dissertation concludes with a discussion of the study's implications for trait activation and the design of SJTs and other similar measurement procedures that rely on the sampling of situational content (e.g., work samples, assessment centers).
Read
