Assessors and use of assessment centre dimensions: a fresh look at a troubling issue

Previous studies on the construct validity of assessment centres have generally produced puzzling results. The premise of this study is that these prior studies were relatively one-sided. Actually, most previous studies were field studies, which typically used the multitrait-multimethod approach to distinguish between two sources of variance (i.e., exercises and dimensions). Therefore, this study aims to shed light on the issue of assessment centre construct validity by addressing substantive and methodological concerns inherent in previous research. In this study, 85 industrial and organizational psychology students and 39 managers rated videotaped assessment centre candidates in three exercises on six dimensions. Results from generalizability analyses showed that assessors' ratings were relatively veridical. In addition, when assessors rated candidates whose performances varied across dimensions and whose performances were relatively consistent across exercises, they were reasonably able to differentiate among the Various dimensions. They also rated such candidate profiles similarly on the various dimensions across exercises. When assessors rated a candidate profile without clear performance fluctuations across dimensions, distinctions about dimensions were more blurred. Results from student and managerial assessors were similar, although managers distinguished somewhat less between the various dimensions. The research and practical implications of these findings are discussed. Copyright (C) 2001 John Wiley & Sons, Ltd.