Transsituational variability in human performance within assessment centers

Abstract

Current models in the psychology of personality describe behavior as largely a function of the situation (i.e., situation specific) rather than reflective of broad underlying stable traits (i.e., consistent). Because of sample size limitations and inappropriate levels of abstraction in performance measures, this “interactionist” explanation of behavior has rarely been formally evaluated in industrial/organizational psychology. In this study 2056 assessment center candidates were rated on eight skills or abilities (traits) measured over five situational exercises (situations). Person, situation, and trait variance components were identified via the multitrait—multimethod matrix analysis. The results revealed major sources of variance from person and person × situation components, high convergent validity in the ratings, but a lack of discriminant validity across specialized skills and abilities. Assessees performed differently in the various situational exercises (i.e., person × situation interaction), although differential ratings were probably due largely to halo error. The findings are discussed in the context of substantive and methodological issues relating to transsituational variability in behavior.