A Quality Control Chart for Work Performance Appraisal

A substantial disagreement between total quality management/Deming's principles and traditional management falls in the area of work performance appraisal. In fact, Deming ranks the traditional “evaluation of performance, merit rating, or annual review” third in his list of the Seven Deadly Diseases of the western style of management. Deming advocates argue that many of the faulty management practices in performance appraisal originate from a failure to understand variation among workers and a failure to distinguish between the “common causes” and the “special causes” of variation. Deming emphasized quality control charts as a proper tool for monitoring the stability of a system, for distinguishing the special causes from the common causes, and for detecting who among the workers is performing within the system, out of the system on the good side, or out of the system on the poor side. However, the implementation of most control charts requires that the performance ratings be quantitative on the interval scale of measurement, which may not be the case in practice. Some merit systems use ratings that are only categorical on the ordinal scale of measurement (e.g., excellent, good, fair, and poor) or rank workers by arranging them in order of merit from 1, 2, last.

Deming, however, did not show how to construct a control chart for performance appraisal when the performance ratings are reported only on the ordinal scale of measurement. In this article, we propose a quality control chart that is particularly useful in the area of performance appraisal when the workers' ratings are categorical on the ordinal scale of measurement. The proposed chart can aid managers in implementing Deming's teachings on performance appraisal. The manager will then be able to understand variation among workers and to distinguish between the “common causes” and the “special causes” affecting a certain work system. The manager can then determine who among the workers is performing within the bounds of the system, out of the system on the good side, or out of the system on the poor side.