A substantial amount of research in the performance appraisal literature has focused on rater training as a means of improving performance ratings. Unfortunately the value of this research is somewhat equivocated by a lack of organization and integration. The present study provides an integration and a quantitative review of the rater training literature. A general framework for the evaluation of rater training is presented in terms of four rating training strategies (rater error training, performance dimension training, frame-of-reference training, and behavioural observation training) and four dependent measures (halo, leniency, rating accuracy and observational accuracy). Finally, a meta-analytic review is presented to assess the effectiveness of the rater training strategies across the four dependent measures.