Evaluating interview criterion-related validity for distinct constructs: A meta-analysis

Abstract

The employment interview is used to assess myriad constructs to inform personnel selection decisions. This article describes the first meta-analytic review of the criterion-related validity of interview-based assessments of specific constructs (i.e., related to task and contextual performance). As such, this study explores the suitability of the interview for predicting specific dimensions of performance, and furthermore, if and how interviews should be designed to inform the assessment of distinct constructs. A comprehensive search process identified k = 37 studies comprising N = 30,646 participants (N = 4449 with the removal of one study). Results suggest that constructs related to task (ρ =.30) and contextual (ρ =.28) performance are assessed with similar levels of criterion-related validity. Although interview evaluations of task and contextual performance constructs did not show discriminant validity within the interview itself, interview evaluations were more predictive of the targeted criterion construct than of alternative constructs. We further found evidence that evaluations of contextual performance constructs might particularly benefit from the adoption of more structured interview scoring procedures. However, we expect that new research on interview design factors may find additional moderating effects and we point to critical gaps in our current body of literature on employment interviews. These results illustrate how a construct-specific approach to interview validity can spur new developments in the modeling, assessment, and selection of specific work performance constructs.