Two Statistical Aids for Determining the Optimal Number and Width of Pay Grades

Abstract

This paper develops and discusses two statistical aids that can be used by compensation specialists for determining the optimal number and width of pay grades. The aids assume that the point method of job evaluation has been used. The first aid is the adjusted ratio of the between-grade to within-grade sum of squared deviation of job evaluation points, similar to the Analysis of Variance F-ratio. This L-ratio attempts to estimate the optimal grouping of jobs such that the between paygrade variation is maximized, while the sum of withinpay grade variations is minimized. The second aid is the standard error of measurement of the job evaluation method. This tool helps explore the boundaries of the pay grades and provides information regarding the likelihood of misclassifying a job due to measurement error. It is suggested that the two aids could be used to help compensation specialists to choose objectively from among a number of alternative pay grade structures on top of other factors affecting pay grade determination.