Deriving Global Dimension Ratings in Assessment Centers: to Clinically Combine, Unit-Weight, or Optimally-Weight?

An Assessment Center is a procedure in which the performance of a job candidate is evaluated in job simulations. In evaluating each candidate, performance ratings on individual behavioral dimensions within individual exercises must be combined across exercises to form Global Dimension Ratings {GDRs), and these GDRs must further be combined to form an Overall Assessment Rating (OAR). Despite the evidence in the personality assessment and the decision-making literature supporting a mechanical combination of ratings into a global assessment rating, clinical methods of combining ratings continue to be used in assessment centers. The predictive validities of three methods of combining dimension ratings across exercises to derive GDRs for assessment center dimensions (e.g., Decisiveness) were compared; these methods were traditional Clinical combining , Unit-weighting , and Optimal-weighting (i.e., regression-weighting). The criteria for the Clinically combined, Unit—weighted, or Optimally weighted GDR predictors were Behaviorally Anchored Rating Scale (BARS) measures of on-the-job performance on the same behavioral dimensions as those rated in the assessment center. No statistically significant differences were found between the predictive validities of the Clinically combined, Unit-weighted, or Optimally-weighted GDRs. However, if practitioners are considering replacing the traditional Clinical combination method with a mechanical combination method, several cautions are noted