However, evaluations of current assessment systems have shown that subjects were not rated multidimensionally, that ratings made over a significant period of time were not interrelated, and that the objective measure most closely associated with performance ratings was age. Using data obtained during a longitudinal study of performance assessment in an Australian police department, this author argues that plausible alternative interpretations should be allowed to compete with traditional I-O technical assessments. These alternatives would require that the performance ratings analyzed here be treated as reflections of rater characteristics, rather than ratee characteristics. The end result would be different sets of police performance criteria, and different indications of validity, applicable to the various stages of an officer's career.