Files
Abstract
Accumulating evidence suggests that assessment centers (ACs), a popular leader selection tool, fail to predict relational forms of performance, such as organizational citizenship behaviors and transformational leadership (LoPilato et al., 2016). It has been speculated that the failure to predict relational performance reflects bias in ACs toward rewarding candidates with a proclivity to dominate rather than cooperate, and more inclined to get ahead than get along (Hoffman et al., 2015; Meriac et al., 2014). The present study contributes to the literature by investigating this apparent deficiency in three primary ways. First, criterion-related validity analysis showed that AC ratings predicted subsequent ratings of task performance on the job, but not relational or leadership performance. Second, we investigated whether construct validity issues might lead ACs to be less predictive of relational outcomes. Specifically, we confirmed previous findings that ACs tend to reflect personality traits indicative of the propensity to get ahead to a greater extent than to get along. Further, while we found that the dark traits exert a curvilinear effect on job performance, they show a positive linear relationship with AC ratings, indicating that ACs fail to capture the negative effects associated with high levels of dark traits. Finally, we identified a boundary condition of the criterion-related validity of ACs; for candidates high in dark traits, higher performance in ACs corresponded to lower relational effectiveness at work. Implications for research and practice are discussed.