Investigated a merit pay system within the context of a large transit organization on the West Coast. More then 800 employees (aged 26–67 yrs) completed a survey questionnaire; they had an average length of service of 13.5 yrs, ranging from 1 to 44 yrs. One finding of the study was that an employee's performance score was significantly related to the job worth score, as measured by the job grade for each employee's position. This finding suggests a form of systematic bias in performance evaluation scores. The present authors suggest a general procedure (consisting of 4 delineated steps) for evaluating any merit pay program.