Working together: effects of gender composition on job satisfaction and commitment

Abstract

Purpose

The purpose of this study is to estimate the effects of gender composition in the workplace on employee job satisfaction and commitment.

Design/methodology/approach

The data were collected on both the organizational and employee levels at three different points in time in organizations with more than 70 employees. Multi-level mixed-effects ordered logistics regressions were used to account for the multi-level nature of the data and the ordered nature of the dependent variables.

Findings

Employees in gender-balanced workplaces show higher levels of job satisfaction and commitment than those in female-dominated or male-dominated workplaces. The relationship is also based on the gender of the individual, as men show a significantly lower level of both job satisfaction and commitment when working in male-dominated workplaces than others, while for women, the effect is only significant for commitment.

Practical implications

Aiming for a balance in the gender composition of the workplace may improve employee attitudes, especially for men. The results also indicate that further research is warranted into why job satisfaction and commitment are significantly lower among men in male-dominated workplaces.

Originality/value

The relationship between gender and job satisfaction and commitment is well established, but less is known about the effects of gender composition on job satisfaction and commitment. Previous papers have focused on job satisfaction. This paper extends prior studies by estimating the effects of gender composition on both job satisfaction and commitment using multi-level regressions on a rich dataset.

Citation

Olafsdottir, K. and Einarsdottir, A. (2024), "Working together: effects of gender composition on job satisfaction and commitment", Employee Relations, Vol. 46 No. 9, pp. 60-75. https://doi.org/10.1108/ER-08-2023-0443

Introduction

Employee attitudes are important as studies have shown higher productivity where employee attitudes are more positive (Böckerman and Ilmakunnas, 2012; Oswald et al., 2015). Employee attitudes depended to some extent on gender, as women consistently reported higher job satisfaction than men in the last century (Blanchflower and Oswald, 1999; Clark, 1997; Sousa-Poza and Sousa-Poza, 2000a). This phenomenon has been framed as the gender gap paradox due to the concurrent evidence of gender wage inequality and worse working conditions for women than men. As this gap has been more evident in Anglo-Saxon countries than in other countries, it is referred to as the Anglo-Saxon gender gap paradox. Recent studies from this century suggest that this gap is diminishing (Green et al., 2018; Pita and Torregrosa, 2021).

Less attention has been given to how gender composition in the workplace may affect employee attitudes and whether there is a difference between female-dominated and male-dominated workplaces. In most countries, men and women tend to work in different occupations, leaving a sizable occupational gender gap. The dissimilarity index (DI) measured 49% in the European Union (EU) in 2019; hence, almost half of all employees would have to change occupations to have equal numbers of men and women (Eurofound and European Commission Joint Research Centre, 2021). It is therefore important to enhance the understanding of whether there is a difference in job satisfaction depending on the gender composition of workplaces.

Existing research has provided mixed results. On the one hand, Fields and Blum (1997) found higher levels of job satisfaction among men and women working in gender-balanced groups than those working in gender-balanced groups. Bender et al. (2005) found that job satisfaction measured highest in organizations with only women, while job satisfaction was significantly lowest among organizations with a majority of men. On the other hand, more recent studies show a positive relationship between gender diversity and overall employee job well-being (Haile, 2012; Clark et al., 2021). While these studies do not provide identical results, they indicate that gender composition matters for job satisfaction and call for further research. Existing research is mainly focused on two Anglo-Saxon countries, the USA and the UK; hence, using data from a Nordic country adds a new dimension and new context to the subject matter.

This paper contributes to the literature by using a rich dataset on employee attitudes to estimate the effects of the gender composition in the workplace on job satisfaction and commitment. While most studies focus on job satisfaction, we extend prior studies by estimating the effects of gender composition on employee commitment and job satisfaction. By using ordered logistic regressions, we account for the ordered nature of the dependent variables, and by using multi-level mixed-effects, we also account for the time factor and company-level data.

Our results indicate that employees in gender-balanced workplaces showed the highest level of job satisfaction as well as the highest level of commitment. In fact, the relationship seems to be asymmetric in that there is less job satisfaction and commitment in female-dominated workplaces and the least job satisfaction and commitment in male-dominated workplaces. Therefore, workplace gender integration and diversity management should be of interest to employees, managers and organizational researchers alike.

The second and third parts of the paper describe the theoretical background and the importance of gender integration for the economy. In part four, we describe the data and methodology, while the fifth part presents the results of the regressions, followed by discussion and conclusion.

Key employee attitudes and the gender gap paradox

This study focuses on the effects of gender composition in the workplace on job satisfaction and commitment, two key employee attitude constructs commonly used in research where the relationship between different human resource management (HRM) practices and different employee related attitudes or outcomes are of interest (Chang, 2005; Edgar and Geare, 2005; White and Bryson, 2013; Hauret et al., 2022).

Job satisfaction reflects a positive or negative general attitude towards the job and the organization and is related to employee expectations as well as external comparisons (Locke, 1976; Cranny et al., 1992; Saari and Judge, 2004). The overall quality of jobs measured as job satisfaction, or an employee's evaluation of different job aspects, is considered an important dimension of employees' well-being (Clark, 2015). Job satisfaction is also considered one of the key outcome factors when studying the effect of human resource management (HRM) practices on employees (Nishii et al., 2008; Wu and Chaturvedi, 2009; White and Bryson, 2013; Hauret et al., 2022). It has been established that job satisfaction affects turnover intentions and is an important predictor of life satisfaction (Judge and Watanabe, 1993; Argyle, 2001). Furthermore, job satisfaction and other measures of employee attitudes have economic consequences, as productivity is higher in workplaces where employee attitudes are more positive (Böckerman and Ilmakunnas, 2012; Oswald et al., 2015). A longitudinal, organization-based study in Finland showed that increased job satisfaction led to increased value-added per hour worked (Böckerman and Ilmakunnas, 2012). A recent study among nurses, engineers and managers in Iceland (Einarsdottir et al., 2019) found no difference in job satisfaction between nurses, engineers and managers, which could be explained by the expert nature of these jobs.

Commitment can be viewed as an employee attitude towards the organization or the employer rather than the job itself. It was initially conceptualized and operationalized into three distinct measures: affective commitment (AC), normative commitment (NC) and continuance commitment (CC) (Meyer and Allen, 1991). AC describes attitudes towards the organization or the state when the employee experiences identification with the organization to which they belong, thus creating a willingness to contribute actively towards its success (Porter et al., 1974). It has been suggested that AC develops when employees have fulfilling work experiences and feel physically and psychologically comfortable in the organization and competent in their work (Meyer and Allen, 1991). Normative commitment (NC) is the sense of obligation to continue to work for the employer or the organization, while Continuance commitment (CC) describes an awareness of the social and economic cost associated with leaving the organization. This study focuses on AC and NC. Commitment is related to a number of individual and organization-level outcomes, such as employee turnover intentions, actual turnover and absenteeism, as well as productivity and efficiency (Loi et al., 2006; Podsakoff et al., 2009). Employee commitment can thus affect various organization and labor market outcomes through employee decisions on whether to quit, how much effort to put into a job and whether to participate in the organization and the labor market actively.

A persistent and consistent gender gap exists in relation to employee attitudes in the workplace, where women tend to be more content overall than men (Sousa-Poza and Sousa-Poza, 2000b; Clark, 1997), which has been explained by women's lower expectations (Clark, 1997) where they are less active in the labor market than men. This has been termed the Anglo-Saxon paradox, as even when controlling for job content, women in the UK and USA remain significantly more satisfied at work than men (Sousa-Poza and Sousa-Poza, 2000b, p. 137). However, this gender difference seems to be diminishing over time (Pita and Torregrosa, 2021) and one study in Britain indicated that it had disappeared there by 2014 (Green et al., 2018).

In countries with more equal opportunities for women and men and where women are relatively more active in the labor market, such as the Nordic countries (Hauret and Williams, 2017), the gender-job satisfaction paradox has been less evident, driven by various institutional labor market interventions (Kaiser, 2007). This is supported by other Nordic research (Eskildsen et al., 2004; Einarsdottir and Stefansdottir, 2008). In Pita and Torregrosa's (2021) study, with 2015 data from 35 countries, men in Denmark, Norway and Finland were more satisfied than women, while women in Sweden still reported slightly higher job satisfaction than men. This suggests that a different pattern may exist among neighboring countries despite economic and social similarities. On the one hand, women in Denmark, Norway and Finland may also have raised their expectations relating to work due to their exposure to high levels of gender equality early in life, in line with Clark's hypothesis (1997) regarding their lower expectations playing a significant role in the job satisfaction paradox (Perugini and Vladisavljevic, 2019). On the other hand, men's expectations may have been lowered, and their job satisfaction thus positively affected, or they may have been positively affected through a realized increase in gender diversity in the workplace.

The relationship between gender and commitment is less studied than the relationship between gender and job satisfaction. However, Einarsdottir and Stefansdottir (2008) found that women scored higher on commitment in Iceland than men. Hence, it is of interest to study whether the gender composition of the workplace affects these two key employee attitudes, job satisfaction and commitment, especially in a Nordic context that is considered relatively egalitarian.

Gender segregation and gender composition in the workplace

While most labor markets are gender-segregated both at the sectoral and occupational levels (Bettio et al., 2009), not much attention has been paid to the effects the gender composition of groups has on key employee attitudes, or more specifically, on job satisfaction and commitment at the organizational level. Gender composition in groups is particularly important in group decision-making. Teamwork, work councils and various forms of employee participation are used to an ever-wider degree in increasingly complex organizational settings requiring team efforts. For instance, empirical research indicates less willingness on behalf of women to compete and that they tend to shy away from competitive situations (Lackner, 2021) while they deal with conflicts and bullying in the workplace in a less confrontational manner than men (Johannsdottir and Olafsson, 2004; Davis et al., 2010; Rosander and Salin, 2023).

Men and women have been shown to behave differently depending on the gender composition of their group. In an experiment where men and women competed in an online business game, Apesteguia et al. (2012) found that when it came to business performance, teams with an equal gender mix performed best. Undergraduate and MBA students participated in an online business game in teams of three. The experiment showed that among the undergraduate students, three-women teams were outperformed by the other teams, while for MBA students, the best-performing teams were two men and one woman. In an experiment with undergraduate business students, Hoogendoorn et al. (2013) also found that teams with an equal gender mix performed better than male-dominated teams regarding sales and profits. Furthermore, gender differences relating to honesty when working in groups seem to exist. Although the effects are small, men seem more prone to lie than women. In an experiment by Muehlheusser et al. (2015), female groups lied the least and male groups the most, with mixed groups in between.

An experiment in the USA on gender differences in how people responded to conflict in the workplace suggested that women are more likely to engage in constructive behavior when conflicts arise, while men are more likely to engage in destructive behavior (Davis et al., 2010). A study in Sweden also indicated a faster escalation of confrontational conflicts into bullying for men than women in hostile work climates. This may be explained by men being more confrontational in their reactions to conflicts, creating reciprocal effects on work climates (Rosander and Salin, 2023). A study in banks in Iceland also indicated that men seek less help and use more assertive coping strategies than women in cases of workplace bullying, confirming gender stereotypes regarding appropriate behavior (Johannsdottir and Olafsson, 2004).

Studying the effects of gender composition on job satisfaction, Fields and Blum (1997) found that an employee's job satisfaction level was related to the gender composition of the employee's work group. They used ordinary least squares (OLS) with job satisfaction as a dependent variable on a random sample of over 1600 USA workers, controlling for various personal, job and contextual characteristics. According to their results, both men and women working in gender-balanced groups had higher levels of job satisfaction than those working in gender-homogenous groups. Bender et al. (2005) used the share of women as an explanatory variable in explaining job satisfaction in a sample of 1,800 people in the USA. Their results showed the highest level of job satisfaction in organizations with only women, while job satisfaction was significantly lowest among organizations with a majority of men (50–100%). Peccei and Lee (2005) studied the effects of gender composition on job satisfaction using the British Workplace Employee Relations Survey (WERS98) with over 23 thousand observations. Their results showed that the effects were asymmetrical for men and women, with more gender-balanced organizations having a greater positive impact on job satisfaction among men than women. However, the effects were quite weak while linear in nature. Haile (2012) also used the Workplace Employee Relations Survey (WERS) dataset (WERS2004) but found that gender diversity is associated with lower employee well-being among women. In contrast to these results, Nielsen and Madsen (2017) studied job satisfaction and gender diversity in the public sector in Denmark and found no evidence of any significant relationship between gender diversity and job satisfaction.

Hence, prior studies strongly indicate a gender difference in job satisfaction and commitment. There are also indications that the gender composition of teams and workplaces affects employee well-being and attitudes and that the effects are non-linear. This paper revisits the subject matter in a Nordic context and extends prior studies to include the effects of gender composition in the workplace on both job satisfaction and commitment. In both cases, we allow for a non-linear relationship in gender composition. Furthermore, the paper extends prior studies by expanding on the methodology used by Haile (2012) by using multi-level mixed-effects regressions to account for the data being multi-level with both a time factor and company-level data. We also use ordered logistic regressions to account for the data being ordered by nature.

Method

Participants and procedures

The data was collected among employees of Icelandic organizations at three different points in time: 2013, 2015 and 2019. The data collection was part of the International Cranet survey (Parry et al., 2021), in which human resource (HR) managers are surveyed on various issues of human resource management approximately every three years, using the same basic questionnaire in over 40 countries. The survey was sent electronically to HR managers in all organizations in Iceland with 70 or more employees. The HR managers provided information at firm level on the size of the organization, the sector in which the organization operates, the number of employees and the number of female employees versus male employees. The respondents were then asked whether another survey could be sent electronically to their employees by supplying either a random sample of 50 or, if they preferred, all employees.

A few months following the data collection among HR managers, employees in the organizations participating in the employee level survey received a questionnaire. They provided information on their job satisfaction and commitment, along with various background information on age, seniority, education, work hours and whether they held a management position. Collecting data on both the organization-level and employee-level allowed us to link firm-level information on the gender composition of employees and other organization-level background information with information on employee attitudes towards the job (job satisfaction) and the organization (commitment) as well as individual employee background information.

The dataset included a total of 3,047 answers from 100 public and private organizations. There were 1,463 responses from 2013, 814 from 2015, to 770 from 2019. Employees of 4 organizations answered in all three periods and 21 organizations answered in two of the three periods. Employees of 47 organizations answered only in the first period, employees of 19 organizations answered only in the second period and employees of 17 organizations answered only in the third period. Of the 3,047 responses, 1,418 (46.5%) were men and 1,629 (53.5%) were women. A total of 1,153 held managerial positions. Among the respondents, 56% worked in services, 24% in public services and 20% in manufacturing. The average age of respondents is between 40 and 49 years of age and the average workweek is 42 h. Around 60% of the respondents had a tertiary degree and only 10% had only primary education. Most of the respondents had seniority of over 5 years.

Measures

Four items with positive/negative statements were used to measure job satisfaction. Three questions were used from Judge et al. (2001), and the fourth one was added to include measurement of overall job satisfaction. The items included are (1) Most days I am enthusiastic about my job; (2) I find real enjoyment in my work; (3) I consider my job to be rather unpleasant (R = reverse scale is used); and (4) Overall, I feel fairly satisfied with my present job. Reverse scale responses were recoded such that a high rating referred to a more positive attitude in line with the other questions. Each question was answered on a Likert scale ranging from 1 to 5. Together, with equal weight, the four questions form the construct of employee's overall attitude towards their job or job satisfaction. The average response was 4.06, with a standard deviation of 0.79. The Cronbach Alpha reliability for job satisfaction is 0.85, which is acceptable.

The six highest loading items from Allen and Meyer's (1996) AC and NC scales were combined into one measurement of commitment in this study. The commitment indicator is made up of six questions: (1) I do not feel emotionally attached to this organization (R = reverse scale used); (2) I do not feel like a part of the family at my organization (R); (3) This organization has a great deal of personal meaning for me; (4) I do not feel obligated to continue working for my current employer (R); (5) The organization deserves my loyalty; and (6) I would not quit my job now because I have obligations to the employees. As with job satisfaction, each question was answered on a Likert scale ranging from 1 to 5. The responses for the reverse items were also recoded here and reversed in line with other items on the scale. The questions had equal weight in constructing the measure of commitment. The average response is 3.61, with a standard deviation of 0.77. Cronbach Alpha reliability for the construct is 0.79, which is acceptable.

To account for the gender composition in the workplace, the share of women and men was derived from the information provided in the survey by HR managers at the organizational level. From that information, the workplaces were divided into three categories: female-dominated, gender-balanced and male-dominated. In female-dominated workplaces, the share of women exceeded 60%, and in male-dominated workplaces, the share of men exceeded 60%. Others were labeled gender-balanced. Table 1 shows the gender distribution of participants in female-dominated, gender-balanced and male-dominated organizations.

Analysis

In the analysis, the constructs job satisfaction and commitment were used as dependent variables in the regressions. To account for the ordinal nature of the Likert scale, that is, a higher number being assumed to correspond to a higher outcome, with the distance between the numbers not necessarily being equal, ordered logit regressions were used to estimate the relationship between job satisfaction and commitment and the independent variables.

The data used in this study was collected on two different levels, organization level and employee level, while the data collection was conducted at three different points in time. Hence, each data collection is cross-sectional in nature, while the repeated observations give them some properties of time series. To make full use of the nature of the data, we use multi-level mixed-effects regression for the estimations. The first level is time, which takes account of the three different time periods when the data was collected. The second level is the organization level. Some organizations participated in all three time periods, some in two of the three and others only participated once. Since we can track organizations through time, we can account for organizational heterogeneity in the regressions. The third level is the employee level. It is possible that the same individual answers up the three times; however, we are unable to track individuals over time.

By using multi-level mixed-effects regressions, we are able to make full use of the nature of our data, controlling for each level, and to account for the ordinal nature of the Likert scale, we used ordered logit multi-level mixed-effects regressions. Controlling for time and organizations in a classic panel-data regression would give similar results; however, multi-level mixed-effects regressions are better suited.

The gender composition in the workplace is the independent variable that was the focus of this study. The gender composition was indicated using two dummy variables: One indicating a male-dominated workplace, where the share of men was larger than 0.6 and one indicating a female-dominated workplace, where the share of women was larger than 0.6. The left-out variable indicated a gender-balanced workplace, where the share of both men and women was between 0.4 and 0.6.

The other controls used in the regressions on each individual are, first of all, the gender of the individual, indicated by a dummy variable where 1 is female and 0 is male. In addition, we control for age, seniority, education, work hours and whether the individual holds a management position. Most of the information on the individual is in the form of dummy variables. The age was presented as a set of dummy variables by decade, and seniority was also presented as a set of dummy variables with years of seniority. Education was measured by the highest degree attained, also presented by a set of dummy variables. A dummy variable indicating whether the employee had a management position within the organization was also included, as prior studies suggest that job characteristics relating to independence in decision-making influences job satisfaction (Humphrey et al., 2007). Work hours were presented as the reported average weekly hours worked.

In addition to controls on the individual, we controlled for information on the organization, both industry and size. Organizations were classified by industry into three categories, the primary sector, services and public services, using a set of dummy variables. Size of the organization was measured as the natural log of the number of employees.

Results

The data show the highest job satisfaction and commitment level in gender-balanced workplaces, as shown in Figure 1. Both measures indicate a lower level of job satisfaction and commitment in female-dominated workplaces and the lowest level among male-dominated workplaces.

The results of the three-level mixed-effects ordered logit using job satisfaction and commitment, respectively, as dependent variables, are shown in Table 2.

Using job satisfaction as the dependent variable, we found that working in a male-dominated workplace negatively affects job satisfaction, and the effect is statistically significant. Working in a female-dominated workplace had a negative but insignificant negative effect on job satisfaction. Hence, job satisfaction measured highest in gender-balanced workplaces when controlling for individual- and organization-level variables. Furthermore, women showed significantly higher job satisfaction than men.

When using commitment as the dependent variable, we found significant differences for individuals working in female-dominated and male-dominated workplaces. Employees were significantly more committed in gender-balanced workplaces than in female-dominated or male-dominated ones. As with job satisfaction, women overall showed significantly more commitment than men.

The above results show a significant gender difference in measures of job satisfaction and commitment. This suggests there could be a difference between men's and women's attitudes depending on their work environment regarding gender composition. Figure 2 shows the gender difference in the data between job satisfaction and commitment and gender composition.

Figure 2 shows that the pattern of the highest level of job satisfaction in gender-balanced workplaces holds for men but becomes less pronounced for women. When it comes to commitment, the difference depending on the gender composition of the workplace seems to stem from divergent patterns among men and women, with women showing more commitment than men in gender-balanced and male-dominated workplaces.

To explore this relationship further, we ran the same regressions as before on men and women separately, shown in Tables 3 and 4.

Table 3 shows that the significant negative relationship between job satisfaction and working in a male-dominated workplace only holds for men, while it is not significant for women.

Similarly, the negative relationship between commitment and working in a male-dominated workplace still holds for men, while commitment is significantly lower for women in female-dominated workplaces (see Table 4).

All the regressions show a significant and fairly large effect of holding a management position. There is no significant difference between the three sectors, manufacturing, services and public services, except that men in the public services show significantly lower levels of commitment. Job satisfaction increases with age (the comparison group is Age 18–29), while it seems to fall with seniority (the comparison group is Seniority less than 1 year). The correlation between these two variables is 0.45. Job satisfaction was highest among those with tertiary education (the comparison group was primary school education). While the data was collected at three different points in time, the results show no difference between the time periods.

Discussion

Finding ways to improve the well-being of both men and women in the workplace is a win-win situation for both employers and employees, as it has been shown to increase productivity (Böckerman and Ilmakunnas, 2012; Oswald et al., 2015). This paper finds that employees, both men and women, are more satisfied with their jobs and show more commitment to their organization in a gender-balanced workplace than if either gender dominates. Employees working in female-dominated workplaces showed lower job satisfaction and commitment than those in gender-balanced workplaces, while the lowest level of job satisfaction and commitment was in male-dominated workplaces.

Fields and Blum (1997) used OLS in their estimates to find a positive link between job satisfaction and gender composition. Haile (2012) expanded on the methodology by accounting for unobserved heterogeneity in the individual and organization. We account for unobserved heterogeneity at the organization level, while we are unable to do that on the individual level. However, we take the analysis further and expand prior studies by using multi-level mixed-effects ordered logistic regression. Furthermore, while Haile (2012) assumes a linear relationship, we show that the relationship is non-linear.

The results support those of Fields and Blum (1997), who found that both men and women working in gender-balanced groups had higher levels of job satisfaction than those working in gender-homogenous groups. Their research used a sample of 1,600 individuals, while our sample included over 3,000 observations. Our results are also in line with Bender et al.'s (2005) result that job satisfaction was significantly lowest among organizations where the majority of employees were men. However, he found that job satisfaction measured highest in organizations with only women. Peccei and Lee (2005) found a weak link between job satisfaction and gender similarity. They define gender similarity as gender composition in the respondent's local occupational group, which might not reflect the gender composition in the workplace as a whole. On the other hand, Haile (2012) found that gender diversity was associated with lower employee well-being among women.

The main contribution of this paper lies in the fact that the richness of the data set allowed us to separate the sample by gender, as the data had enough men working in female-dominated workplaces and enough women working in male-dominated workplaces to make such a differentiation possible. We found that the relationship between gender composition in the workplace and job satisfaction and commitment is highly dependent on gender. Job satisfaction of women was highest in gender-balanced workplaces, but not to a significant extent. However, for commitment, we found it was significantly lower for women in female-dominated workplaces. Furthermore, men showed a significantly lower level of both job satisfaction and commitment in male-dominated workplaces than in other workplaces.

Our results suggest that men would benefit most from increased gender integration in the workplace, possibly through spillover effects from women's positive attitude towards both the job and the employer. It could also occur through a culture of stronger social relations, less competitive culture, more honesty, less confrontational and less assertive behavior, along with more constructive and help-seeking behavior, when coping with conflicts and bullying situations. Women are more likely to bring these to the workplace, as indicated by prior studies (Johannsdottir and Olafsson, 2004; Davis et al., 2010; Muehlheusser et al., 2015; Lackner, 2021; Rosander and Salin, 2023). Hence, employers of male-dominated workplaces could improve productivity in the workplace by improving the gender balance or perhaps by adopting a less assertive and confrontational and more constructive work culture. In other words, by adopting a more female-oriented work culture.

Women were significantly more content in the workplace than men when controlling for various organizational and individual variables regarding both job satisfaction and commitment. These results are in line with prior research relating to the job satisfaction gender gap (Sousa-Poza and Sousa-Poza, 2000b) but contrary to prior research indicating the gender effect to be less evident in the Nordic countries (including Iceland, where this study takes place) than elsewhere (Eskildsen et al., 2004; Einarsdottir and Stefansdóttir, 2008; Kaiser, 2007). It is also contrary to conclusions from prior studies that the gender satisfaction paradox and female advantage is a disappearing phenomenon (Green et al., 2018; Pita and Torregrosa, 2021). In line with Clark's (1997) hypothesis, women's expectations may still be lower than men's in the Icelandic context, making Iceland more like Sweden than the other Nordic countries. Our results also confirm that women are more committed to their work than men, in line with prior studies in the same cultural and institutional context (Einarsdottir and Stefansdottir, 2008) suggesting that women may put more effort into their jobs and be less likely to quit. The high positive effect of holding a managerial position on both job satisfaction and commitment can be explained by the characteristics of managerial jobs, especially the independence in decision-making, which has been shown to explain a large part of job satisfaction (Humphrey et al., 2007).

The research presented here is based on an extensive data set gathered at three separate points in time and at two levels. In the regressions, we take full advantage of the richness of the dataset in terms of using multi-level regression with time as the first level and organization as the second. Still, there are limitations as the assignment into the three groups, male-dominated workplaces, female-dominated workplaces and gender-balanced workplaces, is not random as people tend to choose their workplace or employers choose their employees. Hence, it is not possible to rule out that selection bias might affect the results. Furthermore, wages have been shown to affect job satisfaction (Clark, 1997; Shields and Price, 2002), while wage information was not available in our dataset.

Our results give reason to continue to explore the relationship between gender composition of workplaces and employee attitudes, both by using datasets that include information on wages and by applying methods to correct for selection bias. Furthermore, for policy purposes, research is warranted to explore why men show lower job satisfaction and commitment in male-dominated workplaces than in other types of workplaces and how to remedy that. The research can also be extended in the direction of other background variables than gender, such as migration status and ethnic background, age of employees and gender of supervisor, to name a few possibilities.

While the dataset is rich, it has some limitations. The data was collected among organizations with 70 or more employees and, therefore, does not necessarily reflect the situation in smaller workplaces. While we can follow organizations through time, we are unable to track individuals over time, which would improve our estimates.

Conclusion

This paper sought to shed light on the relationship between gender composition in the workplace and employee attitudes, thus extending prior studies. Using a rich dataset of over 3,000 observations, the results showed that working in a gender-balanced workplace provides the highest job satisfaction and commitment for both men and women. Estimating the effects separately for men and women, women still showed the highest job satisfaction in gender-balanced workplaces, while the difference was not significant. However, for men, job satisfaction and commitment were significantly lower in male-dominated workplaces than in gender-balanced workplaces. As for commitment, men showed a significantly lower level of commitment in male-dominated workplaces, while women showed significantly lower levels of commitment in female-dominated workplaces. These results indicate that there may be value for organizations to improve the gender balance among their workforce through increased diversity to improve the well-being of their employees as well as productivity. In other words, there is an indication that women's positive attitude towards their jobs and employers may positively affect the attitude of their male co-workers when workplaces become increasingly gender-balanced, while some of women's commitment might get lost in female-dominated workplaces. In practice, this suggests that gender balance should be considered in the last stages of recruitment decision-making.

Figures

Figure 1

The relationship between job satisfaction and commitment and gender composition

Figure 2

The relationship between job satisfaction and commitment and gender composition by gender

Table 1

Gender distribution of respondents across types of organization

	Female-dominated	Gender-balanced	Male-dominated	Sum	N
Women, %	47.1	29.2	23.7	100	1,631
Men, %	22.9	24.6	52.5	100	1,416
Total					3,047

Source(s): Created by authors

Table 2

Regression of gender composition on job satisfaction and commitment

Variable	Job satisfaction		Commitment
	Odds ratio	Std. err	Odds ratio	Std. err
Gender (1 = woman, 0 = man)	1.2698	0.0359***	1.3246	0.1735**
Female-dominated workplace	0.8808	0.0819	0.9066	0.0378**
Male-dominated workplace	0.8050	0.0452***	0.7600	0.0629***
Management position	1.6688	0.1235***	1.6061	0.1663***
Working hours	1.0080	0.0061	1.0083	0.0057
ln(Number of employees)	1.0000	0.0001	0.9999	0.0001
Services sector	1.0248	0.0568	0.9298	0.0712*
Public services	0.9491	0.1851	0.7884	0.0988
Age 30–39	1.0631	0.0905	0.9155	0.1617
Age 40–49	1.1912	0.1454	1.0179	0.0263
Age 50–59	1.3803	0.1229***	0.9179	0.1521
Age 60–69	1.7425	0.2725***	1.0166	0.0973***
Seniority 1–2 years	0.6089	0.0047***	0.7697	0.0070***
Seniority 3–5 years	0.5575	0.0487***	0.6697	0.0669
Seniority 5–10 years	0.5860	0.0319***	0.8191	0.1015
Seniority over 10 years	0.5661	0.0170***	1.0148	0.0367
Technical education	1.0076	0.0463	0.9362	0.0812***
Secondary school degree	0.8313	0.1046	0.8238	0.0234
Bachelor's degree	1.1570	0.1142	0.9419	0.0746***
Master's or Ph.D.	1.1502	0.0731**	0.8694	0.0311
N	3,046		3,046

Note(s): *** significant at 1% level, ** significant at 5% level and * significant at 10% level

Source(s): Created by authors

Table 3

Regression of gender composition on job satisfaction by gender

Variable	Men		Women
	Odds ratio	Std. err	Odds ratio	Std. err
Female-dominated workplace	0.7892	0.1547	0.9407	0.0920
Male-dominated workplace	0.7254	0.0373***	0.8653	0.0829
Managerial position	1.8094	0.2125***	1.5821	0.2198***
Working hours	1.0059	0.0131	1.0082	0.0050
ln(Number of employees)	1.0001	0.0002	1.0000	0.0001
Services sector	1.0104	0.1005	1.0769	0.0619
Public services	0.8793	0.1781	1.0138	0.1479
Age 30–39	0.9952	0.0781	1.1356	0.2447
Age 40–49	1.1157	0.1654	1.2899	0.2693
Age 50–59	1.1688	0.2545	1.6433	0.3261**
Age 60–69	1.7830	0.6423	1.5860	0.4789
Seniority 1–2 years	0.5983	0.0638***	0.6203	0.0182***
Seniority 3–5 years	0.5202	0.0404***	0.5916	0.0664***
Seniority 5–10 years	0.6344	0.1077***	0.5493	0.0209***
Seniority more than 10 years	0.6109	0.0549***	0.5206	0.0310***
Technical education	1.1723	0.2597	0.8752	0.3237
Secondary school	0.9454	0.1050	0.8168	0.1611
Bachelor's degree	1.2874	0.3800	1.0979	0.1162
Master's or Ph.D.	1.3343	0.1504**	1.0437	0.1448
N	1,416		1,630

Note(s):*** significant at 1% level, ** significant at 5% level and * significant at 10% level

Source(s): Created by authors

Table 4

Regression of gender composition on commitment by gender

Variable	Men		Women
	Odds ratio	Std. err	Odds ratio	Std. err
Female-dominated workplace	1.0044	0.1749	0.8717	0.0366***
Male-dominated workplace	0.6969	0.0555***	0.8688	0.0940
Management position	1.8580	0.1367***	1.3797	0.2584*
Working hours	1.0033	0.0103	1.0119	0.0049**
ln(Number of employees)	1.0000	0.0002	0.9999	0.0001
Services sector	0.8978	0.1371	0.9974	0.0556
Public services	0.6319	0.0619***	0.8982	0.1094
Age 30–39	0.8689	0.1391	0.9721	0.1707
Age 40–49	0.8850	0.0717	1.1820	0.1207
Age 50–59	0.8437	0.0325***	1.0018	0.3109
Age 60–69	0.8418	0.2072	1.1953	0.1793
Seniority 1–2 years	0.8091	0.0094***	0.7339	0.0341***
Seniority 3–5 years	0.6197	0.0796***	0.7230	0.1048**
Seniority 5–10 years	0.8343	0.1144	0.8310	0.0915*
Seniority over 10 years	1.1210	0.0582**	0.9493	0.0265*
Technical education	1.0724	0.2659	0.8334	0.1737
Secondary school degree	0.8622	0.0998	0.8276	0.0218***
Bachelor's degree	1.1408	0.3143	0.8705	0.1313
Master's or Ph.D.	1.0010	0.1989	0.7936	0.0982*
N	1,416		1,630

Note(s): *** significant at 1% level, ** significant at 5% level and * significant at 10% level

Source(s): Created by authors