Gender differences in pay expectations among
different occupational groups
Niklas Fransson and Anders Biel
Department of Psychology Göteborg University
Fransson, N., & Biel, A. Gender differences in pay expectation among different occupational groups. Göteborg Psychological Reports, 2000, 30, No. 7. Previous research, almost exclusively using student samples, has suggested that women tend to expect lesser pay for own work than men. Those results have been interpreted as being at least partly responsible for the gender wage gap evident in most developed countries. The main purpose of the present study was to test whether the same gender difference exist among people in the labor market and to explore which variables that mediate pay expectations. Questionnaire data from 329 males and 553 females from 4 different occupational groups were analyzed. The obtained results partly corroborated previous findings in that women and men from 2 of the 4 occupational groups differed in their pay expectations and that 3 out of 4 occupational groups displayed a gender wage gap. The two strongest predictors of pay expectation were current salary and years in occupation. Neither predictor has been tested in previous research using student samples. Those latter findings led to the conclusion that factors predicting pay expectation among students differ from factors predicting pay expectations among people in the labor market.
Key words: Gender wage gap, pay expectation, labor market
Author note: This research was supported by the Swedish Council for Social Research Grant 98-0242:1A to Anders Biel. We wish to thank Rez Shirazi for helpful comments and ideas in the initial stages of this study. Correspondence concerning this article should be addressed to Niklas Fransson, Department of Psychology, Göteborg University, Box 500, SE 405 30 Göteborg, Sweden. E-mail: [email protected]
No. 7:30, 2
Introduction
In recent years wage differences between males and females have been widely discussed and debated in western communities. Swedish national statistics (e.g., Statistics Sweden, 1998; Landsorganisationen, 1999) as well as international (ILO, 1996; as cited in Anker, 1997) have consistently shown that a wage gap between men and women exists both between and within occupations. The magnitude of the wage gap between men and women within different occupational groups vary, but the direction of the inequality is stabile, men have higher wages than women (Statistics Sweden, 1998). Despite the fact that men earn more than women, there is no difference in pay satisfaction between the two groups (Crosby, 1982; as cited in Major, Vanderslice, & McFarlin, 1984). An intriguing and thoroughly investigated problem has been to explain why members of a disadvantaged group tolerate such an injustice. Psychologists have mostly been interested in individual factors (e.g., perception of fair pay, social comparison standard) that might account for wage differences between the sexes (Jackson, Gardner, & Sullivan, 1992; Major & Konar, 1984; Rutte & Messick, 1996). One important factor in explaining this phenomenon is social comparison standard, meaning whose outcome (e.g., wage) one compare own outcome with (Bylsma & Major, 1994; Major, 1989; Major & Forcey, 1985). The comparison process can partly explain the gender wage gap and its maintenance. It has for example been shown that judgments of contentment and satisfaction (e.g., regarding wage) are based on comparisons (preferably) with similar others rather than on assessments of objective status (for a critique and overview see Kruglanski & Mayseless, 1990). Another potentially important factor, and of main interest in the present study, is pay expectations. According to Jackson et al. there has been little research investigating gender differences in self-pay expectations. The few studies that have examined this topic have almost exclusively used student samples (Major & Konar, 1984; Martin, 1989; McFarlin, Frone, Major, & Konar, 1989). However, McFarlin et al. argued that factors predicting people’s pay expectations at career-entry differ from factors that predict pay expectations of people already in the labour market. Their argument is that the availability of different kinds of data (e.g., salary of co-workers, actual pay received by same gender individuals in other occupations) that individuals could use when they estimate their pay expectations differ between these two groups. To understand why this could be the case, one can envision students as making a prediction about future entry pay based on expectations. People already in the labour-market can use back-casting to estimate an entry pay. Furthermore, members of the latter group have personal experience of factors considered as relevant when salaries are adopted. Finally, they can use their own present salary as an anchor. Hence, factors that predict expected entry pay among students may differ from factors that predict estimated entry pay among employees. These differences, between students and people in the labour-market, probably determine in which way they are motivated to engage in actions to increase their actual salary. Hence, students are motivated by the gap between expected pay and actual salary statistics and people in the labour-market are motivated by actual salary differences between males and females and between different occupations. Previous studies based on student samples (mainly in business) have consistently found that women, as compared to men, have lower self-pay expectations (Jackson et al., 1992; Major & Konar, 1984; McFarlin et al., 1989).
No. 7:30, 3
Furthermore, Major et al. (1984) in two hiring simulation experiments showed that among identically qualified job applicants those communicating low self-pay expectations received less pay than job applicants with higher pay expectations. Taken together, these results indicate that, due to lower pay expectations, women will end up with lower salaries than men. One purpose of this study was therefore to test if the difference between genders in pay expectation found in previous studies with students as subjects could be replicated in a sample of people already in the work force. A second aim was to get a clearer view of which variables that mediate (influence) gender differences in self-pay expectation. Finally, we tested if these variables could explain differences in pay expectations within four different occupational groups.
In relation to these aims two hypothesis were formulated:
Hypothesis 1. In general (a) and within each occupational group (b) women will have lower pay expectations than men. This hypothesis stems from previous research showing that women generally have lower self-pay expectations than men (Jackson et al., 1992; Major & Konar, 1984; McFarlin et al., 1989).
Hypothesis 2. Based on previous research job facet importance (Jackson et al., 1992), and social comparison standard (Major & Konar, 1984) are expected to be strong mediators of pay expectation. Furthermore, Desmarais and Curtis (1997) showed that when previous income was salient it strongly influenced estimates of entitlement for work. Hence, current salary is also anticipated to mediate pay expectations.
Method
Procedure and participants. A mail-back questionnaire was sent to 1710 individuals in the labour market. Participants were selected through a stratified random sampling procedure from mailing lists maintained by two labor unions and one larger organization. These sub-samples corresponded to employees in either male-dominated or female-dominated jobs. The national union of metalworkers fulfilled the criterion of male-dominated occupations 500 of their members (250 males) were approached. Based on national statistics (Statistics Sweden, 1998) one trade worker union (Commercial Employee’s Union) was selected to represent female-dominated jobs. Participant from this union will henceforth be called trade workers. Members of this union mainly work as salespersons within a wide range of retail commerce. A total of 400 members (200 males) were sent the questionnaire. Respondents in female-dominated jobs were also chosen from a large local hospital with approximately 15000 employees. From the hospital staff two sub-samples were selected. One consisted of 400 (109 males) untrained nurses and hospital orderlies and the other contained 410 trained nurses (95 males). It should be noted that respondents from the metalworker group, and especially those from the trade worker group, represent a wider variety of occupations than respondents from the groups of untrained and trained nurses.
All three organizations were promised a descriptive report for their cooperation. Single participants did not receive any compensation. After 3 remainders usable questionnaires were obtained from 882 respondents, corresponding to a final response rate of 51.6%.
Data analyses are based on the final sample consisting of 329 males (Age: M = 40.7, SD = 10.61) and 553 females (M = 43.5, SD = 9.70). Male participants reported an average of 14.6 years of full-time job experience in their current occupation (SD = 10.75). The equivalent figure was 16.6 years for females (SD = 9.45).
Questionnaire Job facet importance
Job facet importance was assessed by a set of 15 job characteristics (e.g., having a high salary, having friendly co-workers) identified in previous research (Jackson et al., 1992; Major & Konar, 1984). Respondents rated the importance of each job characteristic on a seven-point rating scale ranging from ”not at all important” to ”very important”.
Social comparison standard
Respondents were asked whom they compared themselves with when they decided what a reasonable pay was for their own work. They were provided with 5 comparison standards. Two items measured high social comparison standard: their superiors or the best paid segment in their profession. One item measured low social comparison standard: the lowest paid segment in their profession. Finally, two items, ”men in my occupation” and ”women in my occupation”, measured gender specific social comparisons. For each item, three response options were provided: “yes”, “no” or “uncertain/don’t know”. In forthcoming analysis “uncertain/don’t know” replies were treated as no comparison.
Pay expectation
Our pay expectation measure differed from measures used in previous research assessing self-pay expectation (e.g., Jackson et al., 1992; Major & Konar, 1984). Since our respondents already received salary, we could not use a straight-forward measure of self-pay expectation such as: ”How much money do you expect to earn your first year working full time” (Major & Konar, 1984). Rather, pay expectations was assessed by 2 items: ”how much do you think persons within your occupation should receive in entry-pay” and ”what is a reasonable salary after 5 years in your occupation”. The measure used is thus henceforth referred to as pay expectation. Answers were provided in Swedish crowns/month before tax.
Results Mediators of gender differences in pay expectations
A number of factors have been proposed in the literature to explain the gap in male and female pay expectation (e.g., Desmarais & Curtis, 1997; Jackson et al., 1992; Major & Konar, 1984). Among the most prominent, and included here, were job facet importance, social comparison standard, current salary, and years in occupation.
Measures of job facet importance were based on a Factor analysis (principal component, varimax rotation with Kaiser normalization) of the importance ratings of 15 job facets. This analysis resulted in 5 factors that explained 71.4 % of the variance. Due to low reliability scores with the other items in the same factor two of the 15 items used to asses job facet importance were dropped from the analysis. Composite scores (means) were calculated for 4 of these factors while the fifth factor only included one single item
(high salary). The four identified factors matched 4 of the 5 dimensions found by Jackson et al (1992) and were labeled accordingly: Personal development opportunities (5 items, α = .79; e.g., opportunities to use special skills and abilities, opportunities for personal development), Job advancement opportunities (2 items, r = .73; good promotional opportunities, opportunities to advance within their profession), Pleasant work environment (3 items, α = .63; e.g., friendly co-workers, highly cooperative co-workers), and Job accommodation to family life (2 items, r = .64; possibility to be home with children, opportunity to adjust work time to child care). In the analysis the entire sample was used since we had no specified hypothesis that the mediators of gender differences would differ across occupations. We also wanted to explore which factors that display gender differences in the total sample. In order to make subjects current salary comparable across subjects full-time salary was computed for those working part-time. Results for these hypothesized mediators and the two pay expectation measures are presented in Table 1. Pay expectation. For the whole sample across occupations there was no significant sex differences in either entry-pay expectations (t880 = .12, ns) or pay expectations after 5 years (t880 = .67, ns). Thus, there was no support for H1a, that women in general have lower pay expectations.
Current salary and years in occupation. The two sociodemographic variables assumed to mediate the gap in female and male pay expectations were both highly significant. As can be seen men had higher current salary than women. This is in line with national statistics showing that men generally earn more than women in Sweden. However, years in occupation showed the opposite pattern. Women had been working significantly longer than men in their current occupation. If years in occupation would constitute a good explanation to the gender gap in pay expectation it would thus indicate that women tend to work in the same occupation longer than men which somehow lead them to have lower pay expectations.
No. 7:30, 6
Table 1. Pay expectation and hypothesized mediators of gender differences in pay expectation (N = 882) Pay measure Mediator Gender Females Males Statistics Entry-pay expectation 14676 15229 .12
Pay expectation after 5 years 18207 18864 .67
Current salary 15434 16570 5.96**
Years in occupation 16.58 14.55 -2.84**
Job facet importancea Personal development
opportunities 6.07 5.74 -5.06**
Job advancement opportunities 5.30 5.13 -1.65
Pleasant work environment 6.60 6.35 -5.02**
Job accommodation to family life 5.51 4.78 -6.67**
High salary 5.84 5.83 -.22 Comparison standardb Their superiors 30.6 29.2 .18 Highest paid 58.0 61.5 1.01 Lowest paid 38.8 40.3 .18 Men in my occupation 54.3 56.1 .25 Women in my occupation 62.2 39.8 39.0**
Note.a Mean ratings on seven point scale ranging from 1 ”not at all important” to 7 ”very important”. Statistic t.
b
Percentage of respondents making comparison. Statistic χ2 **p < .01
Job facet importance. As can be seen in Table 2 women scored significantly higher than men on 3 of the 5 factors that measured job facet importance. The largest mean difference was obtained for Job accommodation to family life, followed by Personal development opportunities and Pleasant work environment. These results are practically identical to the results obtained by Jackson et al. (1992) using a student sample.
Social comparison standard The five items that measured social comparison standard revealed one gender difference. Within the same occupation, men are less prone to compare their salary with the salary of women.
Gender differences in pay expectations
Respondents were divided into 4 groups according to which type of occupation they had. An independent sample t-test was conducted for each group with gender as independent variable and entry-pay expectation, pay expectation after 5 years in that occupation, and current salary, expressed as full time salary, as dependent variables. Results are presented in Table 2.
No. 7:30, 7
Table 2. Mean salary per month, in Swedish crowns, for two measures of pay expectations and for current salary for males and females in 4 occupational groups.
Occupational group Pay measure Metalworkers M t Trade workers M t Trained nurses M t Assistant nurses M t Entry-pay Male Female 14840 14744 .31 14236 12953 3.42** 16548 16665 -.43 15293 14343 3.33** After 5 years Male Female 18583 18029 1.59 17784 16868 2.32* 20426 20550 -.32 18661 17379 3.40** Current salary Male Female 17073 15212 5.28** 16076 13314 5.52** 17347 17221 .45 15544 14331 3.75** **p < .01 *p < .05
Apart from trained nurses males and females differed significantly in current salary. Males had higher wages than women, in particular among trade workers. Among trade workers and assistant nurses, men had higher pay expectancies than women. This was true for both entry pay and pay after 5 years in occupation. In the other two groups, pay expectancies did not differ between women and men. It looks as if trained and assistant nurses, and trade workers adapt to existing conditions. For assistant nurses and trade workers, existing salary differences among men and women were reflected in differences in pay expectations. Trained nurses were also consistent in that women and men did not differ in any of the pay measures. Metalworkers on the other hand seem to be less anchored on their present salary. Our hypothesis H1b that women within each occupation was thus supported by the results for trade workers and assistant nurses. Notice also that all groups of respondents found it reasonable that employees in their own occupation have a higher salary after 5 years in occupation than what they presently earned themselves, despite that their average full-time experience in their present occupation was close to 16 years.
Factors influencing pay expectations
In order to test the importance of the hypothesized mediators (Years in occupation, Salary, Job facet importance, and Social comparison standard) and gender on pay expectation for each occupation, 2 separate regression analyses were performed with each pay measure as dependent variable. The significant predictors (mediators) for each occupation and the overall result for the whole model are depicted in Table 3.
No. 7:30, 8
Table 3. Significant predictors of pay expectations within each occupational group.
Group and Entry-pay After 5 years
predictor β t β t Metalworkers Salary .32 4.14** .35 4.52** R2 = .08, F (13, 178) = 2.27** R2 = .08, F (13, 178) = 2.25** Trade workers Salary Years in occupation Job facet importance Personal develop opp.
.41 -.22 .19 5.05** -2.94** 1.98* .47 -.18 5.90** -2.39* R2 = .23, F (13, 141) = 4.60** R2 = .24, F (13, 141) = 4.78** Assistant nurse Salary
Job facet importance Job acc. to family life Comparison standard Superiors (1 = yes, 2 = no) .30 -.16 4.22** -2.12* .29 -.16 4.14** -2.37* R2 = .12, F (13, 181) = 3.10** R2 = .15, F (13, 181) = 3.54** Trained nurses Salary Years in occupation .14 2.14* -.16 -2.28* R2 = .03, F (13, 234) = 1.54 R2 = .06, F (13, 234) = 2.12* **p < .01, *p < .05
The results for the group of metalworkers (N = 192) showed that current salary was the only independent variable that was consistently related to both types of pay expectation. In a similar fashion the results for the group of trade workers (N = 155) showed salary to be consistently related to pay expectation. Dissimilar to the results for metalworkers years in occupation was significantly related to both types of pay expectation, indicating that individuals that have been working as trade workers for a shorter time period have both higher entry-pay expectation and pay expectation after 5 years than individuals that have been working as trade workers for a longer period. A third significant result for the group of trade workers was that individuals that perceived personal development opportunity as an important aspect of a job had higher entry pay expectations than individuals that perceived personal development opportunity as a less important aspect of a job.
The results for the group of assistant nurses (N = 195) are similar to the results for metalworkers and trade workers regarding the whole model and salary. Thus, both the model and salary were highly significant across both dependent variables. In a similar way as for metalworkers, but dissimilar to trade workers and trained nurses, years in occupation was not significantly related to any type of pay expectation. One of two unique results for this group was the relationship between superiors as social comparison and pay expectation after 5-years namely. The negative sign of the beta weight indicate that people that compare themselves with their superiors have higher pay expectations after 5 years than people who do not make such a comparison.
Secondly, assistant nurses that perceived job accommodation to family life (e.g., opportunity to adjust working time to child care) as a less important aspect of a job had higher entry pay expectations than assistant nurses that perceived that aspect of a job as highly important.
The results for trained nurses (N = 248) differ from the three previously analyzed groups (metalworkers, trade workers and assistant nurses). The most striking difference was that the only the model for pay expectations after 5 years was significant (p < .05) while both models were highly significant (p < .01) for the other three occupational groups. The low degree of explained variance for the model in this group makes interpretation of β-weights for individual independent variables vulnerable to type 1 errors. With this in mind, we can note that a differentiating result refer to salary. Salary was a significant predictor for both types of pay expectations for all previous groups while for trained nurses it was only significant for entry-pay expectation. In a similar way as for trade workers, but dissimilar to metalworkers and assistant nurses, years in occupation was negatively related to pay expectation after 5-years. This result indicates that individuals that have been working as trained nurses for a shorter time period have higher pay expectation after 5 years than individuals that have been working as trained nurses for a longer period.
Across all four groups a non-significant result deserves attention, namely that gender did not give any significant contribution to the explained variance for any group on either dependent measure. The most consistent predictor was present salary that proved significant for both pay measures for metalworkers, trade workers and assistant nurse and for entry-pay for trained nurses. The other human capitol variable, apart from salary, was years worked in occupation. This variable was a significant predictor of both pay measures for the group of trade workers and for pay expectation after 5 years for trained nurses. In all cases more years worked in the current occupation meant lower pay expectation.
Discussion
It has been argued that differences in self-pay expectation between men and women could explain the gender wage gap (Jackson et al., 1992). One argument is that self-pay expectation determines pay satisfaction and entitlement. Since women expect lower wages, they will also be satisfied with lower wages. A second argument is that employees settle wages according to expectations. Hence, existing wage differences are conserved. Both arguments have been supported by empirical studies using university students at different levels of education (Jost, 1997; Major et al., 1984). In the present study we tested if a gender difference in pay expectation exists among people working within 4 different occupations. Obtained results only partly corroborate the previous findings. Hence, trade-workers and assistant nurses displayed gender differences on both types of pay expectation, while no gender differences on either type of pay expectation was obtained for metalworkers or for the total sample. Since trained nurses did not display a gender wage gap no gender difference on pay expectation was expected. A possible reason for the obtained non-significant result for the total sample is that the relatively high and equal level of pay expectation for females and males in the large group of trained nurses obliterate an overall gender effect. If the group of trained nurses, which was not predicted to display any gender difference in pay expectation, is excluded
from the test an overall gender effect is obtained for both entry-pay expectation (t608 = 3.23, p < .001) and pay expectations after 5 years (t608 = 4.10, p < .001). A tentative conclusion is that if a gender wage gap exists within an occupation, pay expectations for males and females will also differ while if there is no gender wage gap, there will be no difference in pay expectation.
The analysis of gender differences on the hypothesised mediators of pay expectations showed that both current salary and years in occupation were significant. Men had received a higher salary than women while women had worked for a significantly longer time in their occupation than men. The dimensions of Job Facet Importance that were rated to be of different importance by men and women were Personal development opportunities, Pleasant work environment and Job accommodation to family life. In all three cases women conceived the dimension as more important than men. The obtained results are very similar to the results obtained by Jackson et al. (1992) using students as respondents. This suggests that there exists a gender difference in what is regarded as important facets of a job. At the same time such differences may have small effects on pay expectations.
Regarding social comparison standard only on of the five measures used proved to be more frequently used by either gender, viz. ”women in my own occupation”. Women used this category as comparison more frequently than men did when judging what a reasonable pay for own work is.
With the obtained gender differences across the mediators and previous research (e.g., Jackson et al., 1992; Major & Konar, 1984) in mind, it was somewhat surprising that gender was not a good predictor of either kind of pay expectation for any occupational group. Rather the outstanding predictor was current salary. Salary was a significant predictor for both pay expectation measures for three out of four occupational groups. The second best predictor for pay expectation across the four occupational groups was years in occupation. The number of years the respondent had worked in their current occupation significantly predicted both types of pay expectations for trade-workers, and pay expectation after 5 years for trained nurses. In all three cases more years on the job resulted in lower pay expectations. A plausible reason is that people use their own entry-pay and salary after 5 years as a reference point when estimating a reasonable entry-pay and pay after 5 years in their occupation. At the same time, people with longer work experience do not fully pay regard to that there has been a general salary increase over the years. Another possibility is that people with extended work experience provide a lower estimate to warrant the fairness of their present salary. A practical implication would be that newcomers to the workforce would expect to receive a considerably higher entry salary than what people with longer experience in the workforce think is a reasonable salary for newcomers. For obvious reasons neither current salary nor years in occupation have been used as predictor variables in previous research with students (not receiving salary) as respondents. The positive results for current salary and years in occupation and the lack of effect of gender support the argument by McFarlin et al. (1989) that factors predicting pay expectation differ between people educating themselves for the labor market and people in the labor market.
No. 7:30, 11
Anker, R. (1997). Theories of occupational segregation by sex: An overview. International Labour Review, 136, 315-339.
Bylsma, W. H., & Major, B. (1994). Social comparisons and contentment: Exploring the psychological costs of the gender wage gap. Psychology of Women Quarterly, 18, 241-249. Desmarais, S., & Curtis, J. (1997). Gender and perceived pay entitlement: Testing for effects of experience with income. Jornal of Personality and Social Psychology, 72, 141-150.
Jackson, L. A., Gardner, P. D., & Sullivan, L. A. (1992). Explaining gender differences in self-pay expectations: Social comparison standards and perceptions of fair pay. Journal of Applied Psychology, 77, 651-663.
Jost, J. T. (1997). An experimental replication of the depressed-entitlement effect among women. Psychology of Women Quaterly, 21, 387-393.
Kruglanski, A. W., & Mayseless, O. (1990). Classic and current social comparison research: Expanding the perspective. Psychological Bulletin, 108, 195-208.
Landsorganisationen (1999). Röster om facket och jobbet: Synen på lönesättning och löneskillnader. Linköping, AB Dangårds Grafiska.
Major, B. (1989). Gender differences in comparisons and entitlement: Implications for comparable worth. Journal of Social Issues, 45, 99-115.
Major, B., & Forcey, B. (1985). Social comparisons and pay evaluations: Preferences for same-sex and same-job wage comparisons. Journal of Experimental Social Psychology, 21, 393-405.
Major, B., & Konar, E. (1984). An investigation of sex differences in pay expectations and their possible causes. Academy of Management Journal, 27, 777-792.
Major, B., Vanderslice, V., & McFarlin, D. B. (1984). Effects of pay expected on pay received: The confirmatory nature of expectations. Journal of Applied Social psychology, 14, 399-412.
Martin, B. A. (1989). Gender differences in salary expectations when current salary information is provided. Psychology of Women Quarterly, 13, 87-96.
No. 7:30, 12
McFarlin, D. B., Frone, M., Major, B., & Konar, E. (1989). Predicting career-entry pay expectations: The role of gender-based comparisons. Journal of Business and Psychology, 3, 331-340.
Rutte, C. G., & Messick, D. M. (1996). Detecting salary discrimination against male and female managers. European Journal of Social Psychology, 26, 727-740.
Statistics Sweden. (1998). Statistical yearbook of Sweden 1999 (vol. 85). Stockholm: Statistics Sweden.