Determinants of pay satisfaction: a discrepancy model evaluation

Attitudes toward pay like other aspects of job satisfaction have been researched extensively. Deficiencies in this research area (e.g., little systematic theory building and single variable predictions of pay satisfaction) have contributed to confusion and a lack of knowledge in this field. The research reported here utilized a cross-regional sample of managerial, professional, and technical employees from a large national oil company (N = 692). Major components of a model of pay satisfaction presented by Lawler (1971) were evaluated, using a multiple regression approach to statistically control the effects of the variables measured. In the first phase of testing the model, data gathered from the company's personnel data system were used as predictors of the JDI pay subscale and four other measures of pay satisfaction. The predictors included age, service, educational level, sex, monthly salary, performance appraisal, potential estimate, job grade, time since last salary increase, and amount of last salary increase. Then the contribution of a set of perceptual variables was evaluated. Self-ratings of performance, job demands, comparative pay equity, and a set of nonmonetary job factors were added to the predictive models of pay satisfaction. These measures were obtained from a pay satisfaction questionnaire. The comparative pay equity factor, an advancement factor, a benefits factor, employee gender (females tended to be more satisfied with their salary), and self-ratings of performance were consistently correlated with the measures of pay satisfaction, and the correlations were in the direction predicted by the Lawler (1971) model. Model predictions concerning self-ratings of job demands, actual monthly salary, and actual time since last salary increase, were only partially supported. While the weighted composite of company maintained measures was correlated with pay satisfaction, the results show the importance of the perceptual variables, particularly the pay equity measure. Finally, while the Lawler (1971) model was shown to have utility, the results of this and other studies were used to develop a modified discrepancy model of pay satisfaction.