Title

AN ANALYSIS OF THE RELATIVE CONTRIBUTIONS OF VARIOUS FACTORS IN SALARY ADMINISTRATION: A COMPARISON OF TWO JOB EVALUATION METHODOLOGIES (COMPARABLE WORTH, WAGE DETERMINATION)

Date of Completion

January 1987

Keywords

Psychology, Industrial

Degree

Ph.D.

Abstract

Both the comparable worth controversy as well as the current federal posture, which emphasizes the importance of basing employment tests on information that is directly related to job content, have led employers to inspect their wage allocation systems. In the present study, 66 professional jobs were evaluated with both a priori point factor and policy capturing structured methods of job evaluation. The intent was to determine which method yielded the more equitable salary hierarchy, and to test for ways to measure and adjust for sex-based wage discrimination. Regression model terms in addition to factors related to job content included a measure of seniority, a supply and demand proxy, a measure of occupational prestige, and an index of the percent female of each occupation. Additionally, several salary measures were evaluated for their appropriateness as criterion measures.^ Contrary to the commonly accepted notion of a "universal" set of factors, different factors were found across the two approaches to job evaluation. The policy capturing structured approach allowed for the identification of a new factor, measuring technical processing activities, which appears to reflect the blue collar component of white collar jobs. The policy capturing approach was thus determined to be the more flexible approach for job analysis purposes (the identification of job requirements). The a priori approach was seen as superior for purposes of job evaluation (salary setting). This determination was founded upon evidence suggesting that female incumbents rated themselves lower than their male counterparts rated themselves in evaluating their job requirements, and that a priori job ratings were found to be more veridical estimates of job duties since they reflected the mean of four raters' evaluations.^ In constructing the salary schedule, it was recommended: that incumbent salaries rather than market salaries be used as the criterion measure; that the function relating supply and demand to market salary be empirically determined and used to adjust salaries predicted from job content information; and that, where benchmark jobs are chosen to construct the salary hierarchy, these jobs be neither male nor female dominated in the marketplace. ^

COinS