Improving the Accuracy of Performance Evaluations: A Comparison of Three Methods of Performance Appraiser Training

Researchers in the area of rater training have relied almost exclusively on rater error measures to assess training effectiveness. A reduction in rater tendency to commit these errors subsequent to training is viewed as evidence that these raters have become more accurate in rating their employees. This assumed relationship between rater errors and rating accuracy has recently been questioned. This uncertain relationship between psychometric errors and accuracy was the focus of the current research effort. Supervisory personnel were trained under one of three training programs (psychometric error training, observation training, or decision-making training). Halo, leniency, range restriction and accuracy measures were collected before, and after training from the three training groups, and a no-training control group. The results suggested that while psychometric error training reduced rater errors, it also detrimentally affected rating accuracy. However, observation and decision-making training had no effect on, or increased error rate, but caused performance rating accuracy to increase after training. The need for a reconceptualization of rater training content and focus was discussed