Gender Differences In Pay Satisfaction And Pay Expectations.(Brief Article)

As noted above, the bivariate analysis finds no relationship between gender and pay satisfaction. When current salary is introduced into regression analysis of gender and pay satisfaction, consistent with earlier research, we find both salary and gender to be significantly related to pay satisfaction. This indicates that when controlling for salary women were more satisfied with their pay than were the men. The partial beta coefficients and level of significance for salary and gender are + .45, p [less than] .001 and - .11, p = .004, respectively. The variance in pay satisfaction explained ([R.sup.2]) by salary and gender is .18.

Table 3 reports the multiple regression analyses pertaining to the hypotheses being investigated. We find no support for the proposition that differences in career paths, job inputs, spouse's salary, career and pay importance and family financial situation account for gender differences in pay satisfaction. The partial beta coefficient for gender is -.11, which is identical to that observed when only current salary and gender are entered in the regression analysis of pay satisfaction (see regression model #1). Gender differences in pay satisfaction appear to be independent of work history and one's family situation. Inclusion of these variables does, however, result in a significant increase in variance explained in comparison to that observed when only salary and gender are entered ([R.sup.2] of .19 compared to .34). With respect to specific measures introduced in regression model #1 (presented in Table 3), the coefficient for the blue-collar/clerical job category was positive and significant (business jobs were the reference category for the dummy variable analysis of job title). In addition, family financial situation was positively related to pay satisfaction, while months in the labor force and spouse's salary were negatively related to pay satisfaction.

Hypothesis 1 predicted that gender differences in pay expectations will be accounted for, in part, by differences in career paths, job inputs, career and pay importance, family and financial need and spouse's earnings. The regression analysis lends some support to this hypothesis. While the simple correlation between gender and expected salary change is +.26 (see Table 1), the partial beta associated with gender is +.17 when measures pertaining to career paths, job inputs, career and pay importance and family financial situation are considered (see regression model #3). The coefficients for measures significantly related to expected salary were, with two exceptions, all positive. Only employment in the service/teaching and blue-collar/clerical job categories had negative coefficients. The variables with significant positive coefficients are: engineering job, industry (employment in manufacturing versus all other industries), months in the labor force, hours of work, amount of budget, career importance, pay i mportance, and family financial situation (see regression model #3).

When the variables specified in hypothesis 1 are entered into the analysis of expected percent change in salary, the partial beta for gender is +.11 (see regression model #6). This exceeds the simple correlation between gender and expected percent change in salary .08, see Table 1). Based on these findings, women appear to have lower expectations for pay than do men. This change is in the opposite direction implied by hypothesis 1. Other variables significantly related to expected percent change in salary are months in the labor force, length of service and family financial situation, all with negative coefficients.

Regression model #2 pertains to hypothesis 2. There is no support for the proposition that differences in satisfaction with selected facets of one's job account for gender differences in pay satisfaction. The partial beta associated with gender remains significant and increases somewhat (beta = -.12). Alternatively, introduction of the measures of satisfaction with selected job facets does significantly increase the proportion of variance in pay satisfaction that is explained. The variance in pay satisfaction explained increases from .34 to .45. The specific job facet satisfaction measures found to be significant are interesting job, security/benefits and advancement opportunities. Increases in 'satisfaction with any of the job facet measures is associated with increased pay satisfaction. The results pertaining to expected salary also provide no support for hypothesis 2. When measures of satisfaction with job facets are introduced to the analysis of expected salary, the partial beta associated with gender de clines from +.17 to +.16. Similarly, the results regarding expected percent change in salary do not support hypothesis 2. When measures of satisfaction with job facets are added to the analysis of expected percent change in salary, the partial beta associated with gender does not change (beta = .11, see regression #7).

Hypothesis 3 predicted that gender differences in pay will be accounted for, in part, by differences in current salary and intention to quit one's current job. Results shown in Table 3 provide modest support for hypothesis 3. When current salary and turnover intentions are added to the regression analysis, the partial beta coefficient associated with gender declines from .16 to .11 (see regression models #4 and #5). Clearly, the dominant variable in the regression analysis of expected salary is current salary (partial beta = .69). Note that with the inclusion of current salary, all career path and job input variables, except blue-collar/clerical job and months in the labor force, found significant in regression models 3 and 4 cease to be significant. Alternatively, spouse's salary increases in magnitude and continues to be significant. Both coefficients associated with turnover intentions are significant. Being undecided about leaving and definitely planning to leave are both positively related to expected s alary (+.05 and +.16, respectively).

The findings regarding expected percent change in salary pertaining to hypothesis 3 are counter to what was expected (model #8). The coefficient for gender increases from +.11 in model #7 to .18 in regression model #8. Men expect even greater percent increases in salary following their MBA studies when the effects of current salary and turnover intentions are considered. Current salary has the strongest relationship with expected percent change in salary (beta = -.41). Definitely, intending to leave one's current employer has a significant positive relationship with expected percent change in salary. With respect to the remaining variables entered into model #8, only satisfaction with advancement is significant (beta = .12).

Discussion and Conclusions

Our findings indicate that significant gender differences in pay satisfaction and pay expectations exist after controlling for variables identified in earlier studies of pay satisfaction and pay expectations. No support is observed for the hypothesis that differences in career path, job inputs, spouse's earnings, career and pay importance, and family financial situation account for gender differences in pay satisfaction. There is some support for the hypothesis that these factors account, in part, for gender differences in expected absolute earnings in a job following completion of the MBA. However, when pay expectations are measured as "percent change in earnings," the findings are counter to our hypothesis. When variables being investigated are entered, the partial coefficient for gender effects increases. Regarding our second hypothesis, there is no support for the proposition that differences in satisfaction with selected facets of one's job account, in part, for gender differences in pay satisfaction an d pay expectations.

Implications for Research

There is some support for our third hypothesis that differences in current salary and intention to quit account, in part, for gender differences in expected absolute salary. Alternatively, when the measure of pay expectations is percent change in salary, the findings are opposite those hypothesized; that is, the coefficient for gender increases. Men expect even larger percentage increases in salary.

The primary focus of this study was to test the validity of the Major and Konar model with a sample of experienced members of the labor force. The findings provide no support for the hypothesis that differences in career paths, job inputs, comparison standards and job facet importance account for gender differences in pay satisfaction. There is some support for the hypothesis that these factors account for part of the gender differences in absolute pay expectations. However, all variables, with the exception of career path and job inputs, cease to be significant when current salary and turnover intentions, are introduced to the analysis. Finally, in the analysis of expected percent change in salary, there is no support for the hypothesis that these variables account for gender differences. In summary, our results provide very little support for the Major and Konar model when the subjects have appreciable amounts of labor market experience.

In our judgement, future research addressing pay satisfaction and pay expectations of the labor force should include a measure of estimated earnings of others in general and a measure of estimated earnings of others of the same gender employed in one's current job as well as in the job to which one aspires. Evidence from other sources suggests these may be critical variables to include in studies of gender differences in pay satisfaction and pay expectations. First, McFarlin et. al (1989) found estimates of same gender earnings in the job to which one aspires to be more related to expected starting salaries among a sample of college students than estimates of earnings for the job in general (without specifying gender). Second, BLS data reporting 1993 annual earnings by occupation for those with a college degree show that the female-to-male earnings ratio in selected business occupations ranged from .72 to .86. The average earnings ratio of women to men for all business occupations in 1993 was .78 (Bureau of Labor Statistics, 1993). Evidence of the tendency to use a same-gender comparison in estimating average or going rates of pay for one's current job or for the job to which one aspires, coupled with labor force data revealing substantial gender differences in pay even when stratified by occupation, suggests these are important variables which will help account for gender differences in pay satisfaction and pay expectations.

Print Get the Mag Weekly Updates