Do within-dimension ratings in assessment centers really lead to improved construct validity?

Abstract

Abstract. In their meta-analysis on the construct validity of assessment centers (ACs), Woehr and Arthur (2003) found that within-dimension ratings of candidates’ performance had better construct validity than within-exercise ratings. However, this finding conflicts with previous experimental studies. In these studies, comparable results were only obtained when the rating approach was confounded with other factors, which are also likely to affect construct validity but which were not taken into account by Woehr and Arthur. In the present investigation, we identified factors that are often associated with a within-dimension rating approach. Most of these factors were also confounded with the rating approach in the studies from Woehr and Arthur’s meta-analysis. Meta-analytic moderator analyses showed that the time at which assessors rated candidates, the exchange of information between assessors, and assessor rotation moderate AC construct validity. The effects of the rating approach can probably be attributed to these confounding factors.