This article reviews 21 studies which manipulated specific variables to determine their impact on the construct validity of assessment centers. This review shows that the studies regarding the impact of different observation, evaluation, and integration procedures yielded mixed results. Conversely, dimension factors (number, conceptual distinctiveness, and transparency), assessor factors (type of assessor and type of assessor training), and exercise factors (exercise form and use of role-players) were found to moderate construct validity. On the basis of the review, practical recommendations are derived to maximize the probability that practitioners design and administer an assessment center with construct validity. Finally, new perspectives for future research are identified.