Skip to main content
PMC home page
Annals of Work Exposures and Health logoLink to Annals of Work Exposures and Health
. 2018 Feb 19;62(4):426–437. doi: 10.1093/annweh/wxx114

Gendered Pathways to Burnout: Results from the SALVEO Study

Nancy Beauregard 1,2,✉, Alain Marchand 1,2, Jaunathan Bilodeau 1,3, Pierre Durand 1,2, Andrée Demers 3, Victor Y Haines III 1,2
PMCID: PMC5905644  PMID: 29471461

Abstract

Aim

Burnout is a pervasive mental health problem in the workforce, with mounting evidence suggesting ties with occupational and safety outcomes such as work injuries, critical events and musculoskeletal disorders. While environmental [work and non-work, work-to-family conflict (WFC)] and individual (personality) pathways to burnout are well documented, little is known about how gender comes to influence such associative patterns. The aim of the study consisted in examining gendered pathways to burnout.

Methods

Data were derived from the SALVEO study, a cross-sectional study of 2026 workers from 63 workplaces from the province of Québec (Canada). Data were analyzed using multilevel path analysis.

Results

Direct effects of gendered pathways were evidenced for work (e.g. decision latitude) and non-work (e.g. child-related strains) environmental pathways, as well as for individual pathways (i.e. internal locus of control). Indirect effects of gendered pathways were also evidenced, with women reporting higher levels of burnout compared to men due to lower levels of decision latitude and of self-esteem, as well as higher levels of WFC. Women also reported lower burnout levels through investing more time into domestic tasks, which could represent a recovery strategy to highly demanding work. WFC further mediated the associations between working hours and burnout, as well as the between irregular work schedules and burnout. These result suggest than men distinctively reported higher levels of burnout due to the specific nature of their work contract negatively impacting on WFC, and incidentally, on their mental health.

Conclusion

Study results supported our hypotheses positing that gender distinctively shapes environmental and individual pathways to burnout. OHS prevention efforts striving for better mental health outcomes in the workforce could relevantly be informed by a gendered approach to burnout.

Keywords: psychosocial work environment, non-work environment, gender differences, burnout, Canada, multilevel path analyses

Introduction

Empirical research has long emphasized the need to better understand pre-pathological mental health outcomes in order to sustainably foster a healthy workforce (Stansfeld and Candy, 2006; Seidler et al., 2014), with burnout attracting significant attention. Burnout is a psychological syndrome that encompasses symptoms of emotional exhaustion, cynicism, and self-inefficacy in work role performance and results from chronic exposure to excessive demands from the work environment (Maslach et al., 2001). Burnout prevalence in the general workforce is estimated between 4 and 13% (Norlund et al., 2010; Maske et al., 2016), with higher figures generally concentrated in at-risk occupations (e.g. human services, healthcare) (Bährer-Kohler, 2012). Burnout has been identified as a risk factor for occupational health and safety (OHS) outcomes such as work injuries, critical events at work (Nahrgang et al., 2011), and musculoskeletal disorders (Armon et al., 2010). Consequently, increasing knowledge on burnout and its explanatory dynamics may represent a key component to a comprehensive OHS prevention strategy.

Nevertheless, one current challenge faced by burnout and OHS research in general arises from the limited knowledge we have on gender inequalities in workers’ health outcomes (Campos-Serna et al., 2013; Malach-Pines and Ronen, 2016), potentially curtailing well-targeted prevention efforts in that respect. Tackling gender inequalities in occupational mental health invites us to conceive gender as a social determinant shaping differential exposure to constraints and opportunities for men and women in employment. The gendered processes underlying such differential exposure (e.g. social stratification, socialization) are posited to be associated with differences in mental health problems reported by men and women (Springer et al., 2012). Relatedly, the primary assumption that burnout is not evenly distributed among men and women seems supported. In a recent meta-analysis, Purvanova and Muros (2010) found that women tended to report somewhat higher levels of overall burnout than men. Further attempts to identify the underlying explanatory pathways leading to differences in burnout between men and women have so far been limited and inconsistent (Malach-Pines and Ronen, 2016). This study thus aims to shed additional light on gendered pathways to burnout relying on a quantitative assessment of a large sample of Canadian workers.

Gendered environmental and individual pathways to burnout

Predominant models of work stress such as the Job Demand–Control–Support model (Karasek and Theorell, 1990), the Effort–Reward–Imbalance model (Siegrist and Peter, 2000), and the Job Demands–Resources model (Demerouti et al., 2001) have targeted multiple pathways linking the work environment to burnout. Work demands are features of the psychosocial work environment (e.g. psychological or physical demands, job insecurity, negative social interactions at work such as abusive supervision, interpersonal conflicts) and of the work contract (i.e. long hours, irregular shifts) conventionally seen as stressors leading to burnout development over time. Work resources are features of the psychosocial work environment (e.g. skill utilization, decision authority, organizational rewards, adequate social support at work) that contribute to reducing the negative impact of work stressors on burnout development. According to these theoretical models, adverse mental health outcomes are anticipated when chronic work demands deplete work resources that can effectively be accessed and mobilized by workers in dealing with such demands over time. Although there is ample evidence supporting associations between work demands, work resources, and burnout (Crawford et al., 2010), little is known about whether and how gender is an antecedent to an unequal distribution of work demands and resources a priori. Seidler et al. (2014), in a systematic review on associations between work demands, work resources, and emotional exhaustion, concluded to insufficient prospective evidence of differences between men and women in burnout and associated work factors.

Environmental pathways to burnout have been extended to the non-work environment. The non-work environment refers to pivotal life environments outside the workplace (e.g. family, networks, community, society) in which workers also evolve on a daily basis (Beauregard et al., 2011), and from which demands and resources influential to workers’ mental health can also emanate. Non-work demands (e.g. parental stress) and non-work resources (e.g. marital support) from the family domain have been the most often investigated in that matter (Leiter, 1990; Missler et al., 2013). Conflicting relationships between work and family domains are also deemed to place additional demands on workers. Work-to-family conflict (WFC) is a ‘form of interrole conflict in which the role pressures from the work and family domains are mutually incompatible in some respect’ (Greenhaus and Beutell, 1985). Both WFC and family-to-work conflict (FWC) are considered as stressors associated with burnout, WFC showing associations of a stronger magnitude (Reichl et al., 2014). Given that work and non-work stressors are antecedents of WFC/FWC (Michel et al., 2011), it can be argued that WFC/FWC plays a mediating role in the stressors–burnout association. One study corroborated that WFC- and FWC-mediated associations between work and non-work social support and burnout and that such associative patterns differed for men and women (Blanch and Aluja, 2012). Despite a narrowing gap in gender inequality in the division of housework observed in industrialized countries over the past decades (Scott and Plagnol, 2012; Altintas and Sullivan, 2016), a pervasive imbalance still places women at a disadvantage. Consequently, assessing gender differences in exposure to non-work demands and resources may provide additional insights as to whether this imbalance, as a gendered social process, extends to burnout. Individual pathways to burnout pertaining to individual resources from workers’ personality have also been documented. Locus of control and self-esteem are psychological traits linked with burnout (Rosse et al., 1991; Wang et al., 2010) and known to be sensitive to gender differences (Sherman et al., 1997; Kocalevent et al., 2014). To our knowledge, no study has assessed whether gendered environmental pathways (work, non-work, WFC/FWC) and individual pathways (personality) contribute cumulatively to explaining burnout differences in men and women.

Lastly, it has been suggested that a gendered differential exposure to stressors and resources from life environments (work and non-work) rather than a vulnerability to stress reactions most prominently accounted for differences between men and women in depressive symptoms (Marchand et al., 2016). In accordance with an exposure hypothesis, we seek to examine whether differences in burnout between men and women stem from environmental pathways that are gendered. Specifically, we posit that a gendered exposure to resources and demands from life environments (e.g. work and non-work) is associated with differences in burnout between men and women. Additionally, we seek to examine whether access to personal resources associated with personality (e.g. self-esteem, locus of control) is a gendered individual pathway explaining burnout differences in men and women. Figure 1 illustrates the analytical model guiding our empirical validation.

Figure 1.

Figure 1.

Open in a new tab

Tested model of the associations between study variables.

Method

Participants and procedures

The analyses were based on data from the SALVEO study, a cross-sectional study conducted in 2009–2012 in the province of Québec (Canada). The SALVEO study followed a two-stage design. First, n = 63 workplaces identified from a list of 500 client companies of a large insurance company were randomly sampled, for a response rate of 41.0% significantly higher than what is reported elsewhere in organizational research (Baruch and Holtom, 2008). Participating workplaces did not significantly differ from non-participating workplaces based on insurance claims for mental health issues. Sampled workplaces were from the private sector and reflected a wide range of industrial activities (n = 19 from the manufacturing sector, n = 44 from the services sector), with a union rate of 34.9% and an average workplace size of 247.1 workers. Second, a random sample of workers was selected within workplaces, yielding n = 2162 participants for a response rate of 73.1%. Respondents answered during working hours a self-administered questionnaire aimed at evaluating work and non-work determinants of workers’ mental health. The Ethics Committees of the University of Montreal, McGill University, Laval University and Concordia University approved the study.

Measures

Burnout

Burnout was evaluated using the 16-item Maslach Burnout Inventory—General Survey (Schaufeli et al., 1996) based on a four-point Likert scale (1 = strongly disagree, 4 = strongly agree). Drawing from a unidimensional approach of the burnout concept (Brenninkmeijer and VanYperen, 2003), we relied on a composite score from subscales of emotional exhaustion (five items), cynicism (five items) and professional inefficacy (six items) to examine burnout as a global syndrome. The composite score had good reliability (Cronbach’s α = 0.89).

Work–family conflict

WFC (four items; α = 0.79) and FWC (four items; α = 0.74) were assessed using a scale developed by Gutek et al. (1991). Items were measured on a four-point Likert scale (1 = strongly disagree, 4 = strongly agree).

Work factors

Psychosocial work environment was evaluated using recommended instrumentation for the Demand–Control–Support model (Karasek and Theorell, 1990) and the Effort–Reward–Imbalance model (Siegrist, 2002). The Job Content Questionnaire (Karasek et al., 1998) measured skill utilization (six items; α = 0.80), decision authority (three items; α = 0.79), psychological demands (nine items; α = 0.73), social support from colleagues (four items; α = 0.83), and social support from the supervisor (four items; α = 0.89). The Effort–Reward–Imbalance questionnaire (Siegrist and Peter, 1996) measured physical demands (one item; ‘my job is physically demanding’), recognition (six items; α = 0.82), career perspectives (four items; α = 0.69), and job insecurity (two items, α = 0.65). Both instruments were assessed using a four-point Likert scale (1 = strongly disagree, 4 = strongly agree). Abusive supervision was measured using Tepper’s 15-item scale with a five-point Likert scale (1 = I cannot remember him ever using this behavior with me, 5 = He uses this behavior very often with me; α = 0.91) (Tepper, 2000). Interpersonal conflicts were measured with a five-item scale developed by Harvey et al. (2006) based on a four-point Likert scale (1 = never, 4 = very often; α = 0.91). The nature of the work contract was evaluated based on the number of working hours per week and on the frequency of exposure to irregular or unpredictable work schedule (1 = never, 4 = all the time).

Non-work factors

Single-parent family status was derived from respondents’ marital status and parental status of minor children in the household (1 = single parent, 0 = other possible combinations in marital and parental status). Marital strains (four items; α = 0.70) and parental strains (four items; α = 0.60) were two scales derived from Wheaton’s work (Wheaton, 1994) based on a binary response format (false/no; true/yes). Domestic tasks measured the number of hours spent on housework chores per week, whereas caregiving accounted for the number of hours spent on taking care of an elderly parent and/or a handicapped person per week.

Individual factors

Personal resources included psychological traits associated with locus of control and self-esteem. Locus of control (internal) was measured using a seven-item scale (α = 0.84) developed by Pearlin and Schooler (1978). Self-esteem was measured using Rosenberg’s 6-item scale (α = 0.87) (Rosenberg, 1979). Both instruments were assessed using a five-point Likert scale (1 = strongly disagree, 5 = strongly agree). In terms of sociodemographic factors, sex categories were coded 0 = men and 1 = women. Age, measured in years, was included as a covariate given cumulative evidence of its association with burnout (Brewer and Shapard, 2004).

Analytical strategy

Multilevel path analyses were performed using Mplus software package (Mùthen and Mùthen, 2006). This analytical strategy was retained to account for the hierarchical structure of the data where workers (Level 1) were nested in workplaces (Level 2). Intra-class coefficients for burnout, WFC, and the majority of work and non-work variables confirmed the necessity of adjusting path analyses for workplace variations (data not shown). All study variables were measured at the workers’ level. Missing values were low, ranging from 0.2% for burnout and 3.0% for caregiving hours, with non-respondents showing on average no differences in burnout, WFC, or gender. Accordingly, analyses were conducted on respondents with complete cases only (n = 2026).

Weighted least-squares parameter estimates method of estimation was retained because of dependent categorical variables included in the models (Mùthen and Mùthen, 2006). The fit of the model was based on a multi-criteria assessment including a chi-square test, root mean square error of approximation (RMSEA), comparative fit index (CFI), and standardized root mean square residual (SRMSR). A non-significant chi-square test, RMSEA < 0.06, CFI > 0.95, and SRMSR < 0.08 are indicative of a good model fit (Hu and Bentler, 1999; Tabachnick and Fidell, 2013). We followed Mùthen and Mùthen’s model constraints’ approach for multilevel models to test for significance of indirect effects (Mùthen and Mùthen, 2006). Further, as the bidirectionality of conflicts steaming from work and non-work domains is well acknowledged, we also investigated whether family–work conflict (FWC) could play a similar mediating role as WFC (Reichl et al., 2014). Because previous multilevel path analyses did not support a significant association between FWC and burnout, we concentrated our efforts on the analytical model displayed in Fig. 1. The analytical model posits differences between men and women as inferred from respondents’ self-classification into either sex category (Risman and Davis, 2013). By positing sex categories as an antecedent of work and non-work demands and resources, as well as their subsequent consequences (i.e. WFC, burnout), we therefore sought to capture gendered exposure processes at the basis of differences between men and women in the explanatory dynamics of burnout.

Results

Table 1 presents means, standard deviations, and percentages for study variables for men and women, and the total sample. Differences between men and women from t-tests were found for burnout, WFC, work (7 out 13 indicators), and non-work factors (2 out of 5 indicators), as well as individual factors (2 out of 4 indicators). Table 2 reports on Pearson’s correlations offering preliminary support for associative patterns between study variables.

Table 1.

Sample descriptive statistics.

Men, n = 1038 Women, n = 988 Total, n = 2026
Mean (%) SD Mean (%) SD Mean (%) SD Minimum–Maximum
Mental health
 Burnout 1.37 0.96 1.49 0.99 1.43**a 0.98 0–6
Work domain
 Skill utilization 18.06 3.49 17.39 3.22 17.73*** 3.37 6–24
 Decision authority 8.83 2.00 8.41 1.98 8.63*** 2.00 3–12
 Psychological demands 23.24 3.85 23.67 3.89 23.45* 3.87 10–36
 Physical demands 2.27 0.98 1.69 0.85 1.99*** 0.96 1–4
 Support from colleagues 12.50 1.92 12.56 1.96 12.53 1.94 4–16
 Support from supervisor 11.87 2.55 12.00 2.65 11.93 2.60 4–16
 Recognition 15.58 2.62 15.78 2.63 15.68 2.63 5–20
 Career perspectives 10.38 2.39 10.32 2.37 10.35 2.38 4–16
 Job insecurity 3.77 1.29 3.78 1.31 3.78 1.30 2–8
 Conflict at work 7.36 2.23 7.40 2.25 7.38 2.24 4–20
 Abusive supervision 18.87 6.69 18.26 6.01 18.58* 6.37 15–69
 Working hours (week) 41.40 5.33 38.72 5.62 40.10*** 5.64 14–67
 Work schedule (irregular) 1.57 0.82 1.45 0.74 1.51*** 0.79 1–4
Family domain
 Single-parent family 4.0 13.0 8.5 0–1
 Marital strains 0.45 0.91 0.44 0.91 0.44 0.91 0–4
 Parental strains 0.18 0.54 0.24 0.60 0.21* 0.57 0–3
 Domestic tasks (hours/week) 4.22 4.70 5.00 4.53 4.60*** 4.64 0–26
 Caregiving (hours/week) 0.86 2.69 0.87 2.84 0.87 2.76 0–19
WFC
 Work–family conflict 9.65 3.45 10.19 3.52 9.91*** 3.49 4–20
Individual
 Sex (women) 48.8 0–1
 Locus of control 19.67 4.76 19.39 4.42 19.54 4.59 0–28
 Self-esteem 19.59 3.43 19.21 3.45 19.40* 3.44 2–24
 Age 39.79 10.94 41.49 10.78 40.62*** 10.89 17–70
Open in a new tab

a P values for differences between men and women.

*P < 0.05.

**P < 0.01.

***P < 0.001.

Table 2.

Correlation matrix.

1 2 3 4 5 6 7 8 9 10 11
1. Sex (women) 1
2. Burnout 0.061 1
3. Skill utilization −0.099 −0.408 1
4. Decision authority −0.105 −0.383 0.628 1
5. Psychological demands 0.055 0.266 0.192 0.055 1
6. Physical demands −0.303 0.079 −0.095 −0.109 −0.024 1
7. Support from colleagues 0.016 −0.338 0.265 0.229 −0.128 −0.110 1
8. Support from supervisor 0.024 −0.415 0.318 0.352 −0.174 −0.152 0.373 1
9. Recognition 0.038 −0.520 0.325 0.350 −0.235 −0.154 0.502 0.623 1
10. Career perspectives −0.014 −0.432 0.440 0.382 −0.098 −0.117 0.304 0.397 0.512 1
11. Job insecurity 0.003 0.436 −0.173 −0.210 0.211 0.093 −0.269 −0.341 −0.463 −0.318 1
12. Conflict at work 0.009 0.354 −0.113 −0.125 0.300 0.141 −0.266 −0.281 −0.401 −0.217 0.255
13. Abusive supervision −0.048 0.403 −0.165 −0.246 0.225 0.163 −0.213 −0.518 −0.496 −0.303 0.284
14. Working hours (week) −0.238 0.002 0.174 0.123 0.238 0.091 −0.012 −0.003 −0.012 0.077 0.011
15. Irregular work schedule −0.079 0.092 0.133 0.079 0.223 0.114 −0.048 −0.041 −0.077 0.008 0.094
16. Single-parent family 0.149 −0.018 0.002 −0.013 0.012 −0.03 0.008 −0.007 −0.011 −0.008 0.005
17. Marital strains −0.004 0.142 −0.035 −0.012 0.077 −0.024 −0.084 −0.047 −0.092 −0.090 0.092
18. Parental strains 0.052 0.064 0.000 −0.004 0.080 0.011 −0.028 −0.03 −0.064 −0.066 0.094
19. Domestic tasks (week) 0.084 −0.022 −0.008 −0.027 0.061 0.034 −0.059 −0.023 −0.037 −0.01 0.073
20. Caregiving (hours) 0.001 −0.025 −0.015 −0.004 0.03 0.047 −0.047 −0.011 −0.006 −0.018 0.042
21. WFC 0.078 0.463 −0.006 −0.062 0.450 0.035 −0.198 −0.202 −0.302 −0.164 0.328
22. Locus of control −0.031 −0.470 0.324 0.311 −0.107 −0.112 0.292 0.267 0.394 0.280 −0.295
23. Self-esteem −0.054 −0.390 0.298 0.260 −0.026 −0.037 0.245 0.167 0.274 0.214 −0.195
24. Age 0.078 −0.117 0.156 0.100 0.04 −0.083 −0.044 −0.052 −0.008 0.080 0.019
12 13 14 15 16 17 18 19 20 21 22 23 24
1. Sex (women)
2. Burnout
3. Skill utilization
4. Decision authority
5. Psychological demands
6. Physical demands
7. Support from colleagues
8. Support from supervisor
9. Recognition
10. Career perspectives
11. Job insecurity
12. Conflict at work 1
13. Abusive supervision 0.443 1
14. Working hours (week) 0.121 0.056 1
15. Irregular work schedule 0.142 0.056 0.205 1
16. Single-parent family 0.022 0.018 −0.01 −0.012 1
17. Marital strains 0.072 0.069 −0.005 0.074 −0.149 1
18. Parental strains 0.058 0.021 0.017 −0.004 0.167 0.155 1
19. Domestic tasks (week) 0.016 0.02 −0.005 0.012 −0.006 0.067 0.143 1
20. Caregiving (hours) −0.002 0.051 0.019 0.003 0.036 −0.014 −0.006 0.166 1
21. WFC 0.247 0.207 0.213 0.237 0.01 0.156 0.129 0.086 0.031 1
22. Locus of control −0.234 −0.247 0.029 −0.013 −0.018 −0.204 −0.144 −0.061 −0.070 −0.360 1
23. Self-esteem −0.114 −0.087 0.003 0.008 −0.038 −0.136 −0.123 −0.019 −0.022 −0.230 0.547 1
24. Age −0.100 −0.015 0.031 0.017 0.043 −0.021 0.094 0.168 0.155 0.016 0.022 0.054 1
Open in a new tab

n = 2026. All correlations above 0.045 are significant at *P < 0.05.

Table 3 presents the results of the multilevel path analysis model. Fit indices show an acceptable fit to the data. Examination of direct effects indicates that WFC and burnout shared common explanatory factors. As such, work factors relative to psychological demands and job insecurity were associated with higher WFC and burnout levels and conversely for recognition at work. Working hours showed a positive association with WFC but a negative association with burnout. Other common direct effects pertained to psychological traits of locus of control and self-esteem, both associated with lower WFC and burnout levels. Contributive factors specific to WFC encompassed working irregular shifts, experiencing marital strains, and being a woman. Similarly, low skill utilization, low decision authority, low career perspectives, interpersonal conflicts, abusive supervision, less caregiving hours and domestic tasks, and WFC were distinctively associated with elevated burnout levels.

Table 3.

Results of the direct effects from multilevel path analysis for burnout.

Direct effects
Sex WFC Burnout
Work domain
 Skill utilization −0.817*** 0.028 −0.065***
 Decision authority −0.491*** 0.040 −0.021*
 Psychological demands −0.013 0.244*** 0.023***
 Physical demands −0.548*** 0.042 −0.025
 Support from colleagues 0.067 −0.012 −0.018
 Support from supervisor 0.032 0.001 −0.009
 Recognition 0.240 −0.062* −0.021*
 Career perspectives −0.121 −0.025 −0.030***
 Job insecurity −0.031 0.353*** 0.089***
 Conflict at work 0.041 −0.015 0.026***
 Abusive supervision −0.358 0.006 0.020***
 Working hours (week) −2.714*** 0.090*** −0.006*
 Work schedule (irregular) −0.286*** 0.431*** 0.029
Family domain
 Single-parent family 0.557*** −0.064 −0.050
 Marital strains −0.004 0.173* 0.015
 Parental strains 0.051* 0.232 −0.032
 Domestic tasks (week) 0.651** 0.021 −0.015***
 Caregiving (hours) −0.067 −0.008 −0.013**
WFC
 Work–family conflict 0.723*** 0.067***
Individual
 Sex (women) 0.035
 Locus of control −0.301 −0.182*** −0.018***
 Self-esteem −0.405* −0.045* −0.036***
Random part
 σ2 (workplace) 0.728** 0.056**
 σ2 (workers) 7.507*** 0.490***
Goodness of fit
 χ2 (df) 299.188 (106)***
 CFI 0.975
 RMSEA 0.030
 SRMSR 0.037
Open in a new tab

n = 2026. Unstandardized coefficients reported. Analyses were adjusted for age.

*P < 0.05.

**P < 0.01.

***P < 0.001.

Differences between men and women in burnout levels were found in a preliminary multilevel regression, after single adjustment for age as a covariate (b = 0.111, SE = 0.054, P < 0.05). From the full model in Table 3, the variable sex categories became, however, non-significant. Hence, its relative contribution to the model was fully mediated by all other explanatory factors (i.e. work and non-work, WFC, individual), therefore confirming the presence of gendered pathways to burnout. Precisely, results from direct effects indicate that women compared to men reported lower levels of skill utilization, of decision authority, and of physical demands and were less prone to work long hours or irregular shifts. Women also tended compared to men to more frequently head single-parent families, experience child-related strains, invest time in domestic tasks, and have lower self-esteem.

Results for indirect effects are shown in Table 4. Women reported higher levels of burnout than men because of their exposure to lower levels of decision latitude (i.e. skill utilization, decision authority) at work as well as to higher WFC levels. Differences between men and women in burnout also indirectly resulted from the ways with which work contract (i.e. working hours, irregular shifts) and WFC associated. In fact, men tended to work longer hours than women, which distinctively prevented them from reporting high levels of burnout. However, as men worked longer hours, they were also more likely to experience higher WFC levels and incidentally higher levels of burnout than women. A similar associative pattern was observed for irregular shifts: as men tended to work more frequently in irregular shifts, they were also more likely to experience higher WFC levels and incidentally higher levels of burnout than women. As for non-work factors, women reported lower levels of burnout than men indirectly due to their greater involvement in domestic tasks. Lastly, women expressed lower levels of self-esteem than men, which in turn contributed to their higher levels of burnout.

Table 4.

Results of the indirect effects from multilevel path analysis for burnout.

Indirect effects Burnout
Sex–working hours–WFC −0.016***
Sex–working hours 0.016*
Sex–work schedule (irregular)–WFC −0.008***
Sex–skill utilization 0.053***
Sex–decision authority 0.010*
Sex–WFC 0.048***
Sex–domestic tasks −0.009**
Sex–self-esteem 0.015**
Open in a new tab

Listwise n = 2026. Unstandardized coefficients reported. Analyses were adjusted for age.

*P < 0.05.

**P < 0.01.

***P < 0.001.

Discussion

In line with studies adopting a gender-based approach to occupational mental health, the aim of the study was to examine the specific nature of gendered pathways to burnout. We put to contribution theoretical and empirical work on gender and work to conceive features of the life environments (i.e. work and non-work factors, WFC) and of individuals’ personality (i.e. self-esteem, locus of control) as being gendered (Campos-Serna et al., 2013; Marchand et al., 2016). Following from this, we hypothesized that differences among men and women in burnout levels resulted from (i) a differential exposure to resources and demands from the work and non-work environments and (ii) a differential access to personal resources from the personality that are mobilized in dealing with demands from the work and non-work environments. Results from multilevel path analysis supported both hypotheses as gendered pathways were evidenced for some of the environmental and individual pathways to burnout tested in this study.

With regard to gendered environmental pathways, women expressed higher levels of burnout because of their lower levels of decision latitude (i.e. skill utilization, decision authority). Although evidence of a gendered effect of decision latitude on burnout is limited and equivocal (Seidler et al., 2014), our results nevertheless support the view that larger, societal differentiation processes associated with women’s limited access to greater control opportunities in the work and non-work domains may be at play (Whitehead et al., 2016). To illustrate, representative data from the Québec workforce collected at the same period of the SALVEO study indicate that on average, women relative to men reported lower levels of decision latitude (Vézina et al., 2011) and were overqualified in greater proportions (Gagnon, 2008). Overqualification is a key differentiation process particularly unfavorable to women (Luksyte and Spitzmueller, 2011). In the work domain, overqualification may express a glass ceiling effect where women are less likely to be promoted to higher-level positions characterized by greater levels of decision latitude. In the non-work domain, overqualification may reflect the fewer alternatives women have in accessing positions that can satisfyingly balance demands from the work and non-work domains.

Relatedly, women were specifically disadvantaged relative to men in reported burnout levels as they experienced more WFC. In accordance with gender role theory, women would be more likely to suffer from the consequences of WFC such as burnout, as WFC is deemed to impede on the recovery process necessary to the performance of non-work roles traditionally assigned to women (Reichl et al., 2014). Of note, men compared with women were more exposed to long working hours and irregular shifts, resulting in higher levels of WFC and burnout subsequently. Under such circumstances, the recovery process for men is also a challenge; long working hours and irregular shifts clash with time allocated to non-work activities, when not suppressing it completely. The fact that women saw their burnout levels significantly reduced compared with men by increasing time invested in domestic tasks may indicate that more time outside work may be necessary to recover from work demands. Part-time work arrangements may also represent a temporary strategy deployed by women to increase their participation in non-work activities such as childcare and domestic responsibilities (Walsh, 1999). Although such withdrawal strategies may offer short-term advantages with regard to burnout, in the long term, women may be missing out on other sources of beneficial resources to their mental health from work (e.g. career opportunities). More studies adopting a life course perspective are thus warranted to clarify the contributive role of a gendered time allocation (work and non-work) in explaining inequalities in occupational mental health. Overall, our study extends previous work (Blanch and Aluja, 2012) by confirming the relevance of investigating associative patterns between work and non-work demands and resources, WFC, and burnout from a gendered perspective. One gendered individual pathway to burnout associated with lower self-esteem in women was evidenced by our study, replicating preliminary trends for work-specific, performance-based self-esteem (Blom et al., 2014).

Direct pathways between the work and non-work environments and burnout were also found, although they were not gendered. Negative social interactions at work (e.g. interpersonal conflicts, abusive supervision) were associated with increased burnout levels in our sample and conversely for the possibility of experiencing good career perspectives. These results are in line with past studies concluding to the presence of environmental pathways linking these work factors (Norlund et al., 2010; Martinko et al., 2013) to burnout. As with non-work factors, weekly caregiving hours contributed to reducing burnout levels. Although the detrimental effects of cumulating long work and non-work hours on burnout have been documented (Paris and Hoge, 2010; DePasquale et al., 2017), our results alternatively suggest that not all workers add up the hours at the cost of their own mental health. Being a caregiver is a role one embarks on with outstanding dedication. For workers who also become caregivers for their close ones, pacing the amount of residual resources directed toward the work domain may prove to be a healthy strategy. Lack of cumulative evidence however regarding the role of caregiving in the occupational mental health literature warrants further examination (Beauregard et al., 2011). Of note, work factors related to job insecurity, psychological demands, and lack of recognition emerged as environmental pathways negatively impacting both burnout and WFC, which corroborates findings reported elsewhere (Leiter and Maslach, 2003; Michel et al., 2011; Sperlich et al., 2012; Seidler et al., 2014; Mauno et al., 2017). In terms of individual pathways, results showed that locus of control was a personal resource associated to lower burnout and WFC levels. This is consonant with meta-analyses confirming the protective role played by locus of control on WFC (Michel et al., 2011), while it further helps clarifying more mitigated trends for burnout (Wang et al., 2010).

Research and practical implications from our study are 2-fold. First, our results supporting direct pathways to burnout that are not gendered reinforce the rationale for primary organizational interventions in occupational mental health emphasizing a more balanced exposure to work demands and resources to all (Egan et al., 2007). The absence of gendered pathways thus expresses a shared experience by men and women of the demands/resources equation. Alternatively, the presence of gendered environmental and individual pathways sustaining differences between men and women in burnout levels supposes that men and women have a distinctive experience of the demands and the resources they juggle with on a daily basis. Indirect pathways to burnout evidenced by our study suggest that interventions aiming to alleviate gender inequalities in exposure to work demands and resources may lead to greater health benefits for one group (men or women) in particular. Nevertheless, studies conceptualizing gender as a structuring principle for work redesign intervention remain scarce (Egan et al., 2007), with similar conclusions extending to secondary interventions aiming at strengthening individual resources to better cope with environmental demands. Second, our results show that several common explanatory pathways characterized WFC and burnout, WFC being also a significant risk factor for burnout. As such, future studies on either phenomenon should pay close attention to their interrelations in the workforce (Reichl et al., 2014). In a nutshell, as suggested by LaMontagne et al. (2014), more tailored-based and context-specific interventions focusing on a better understanding of occupational mental health determinants are warranted. Following from this, our study advises that increasing knowledge about gendered environmental/individual pathways to burnout may be a key to that reflection.

This study nevertheless has limitations. First, the cross-sectional design of the SALVEO study precluded us from drawing causal inferences. Longitudinal studies investigating the dynamic nature of the associative patterns supported by our analytical model are thus required. Second, common method variance is a known methodological caveat encountered in studies based solely on self-reported measures from respondents, which may yield overestimated associations between outcomes and explanatory factors. Relying on multilevel path analysis adjusted study results for objective workplace variations in burnout levels, therefore minimizing small common method variance present in our sample (Marchand et al., 2015). Third, while in our analyses we did not seek to describe the nature of contextual variations in burnout through workplace-level indicators (e.g. healthy organizational policies) (Perry‐Jenkins and Wadsworth, 2017), it is plausible that inclusion of such indicators may relevantly further improve model fit. Fourth, results from the SALVEO study cannot be generalized to the overall workforce as the sample was derived from client companies of a single insurance company. Nevertheless, the data showed a diversified range of economic sectors, firm sizes, and unionization profiles. Fifth, while we cannot fully rule out the possibility that a selection bias at the company level may be present in the data—companies dealing with workers’ mental health issues might have been more inclined to participate—no statistical differences based on incidence insurance claims rate (2009–2012) for mental health problems could distinguish participating versus non-participating companies in the SALVEO sample.

Conclusion

The main contribution of this study highlights the importance of adopting a more balanced view of the role played by gender in shaping occupational mental health, as both gendered and non-gendered environmental and individual pathways to burnout were supported. By examining gendered pathways to burnout, our study provides additional insights about the circumstances by which health differences between working men and women come to unfold (Campos-Serna et al., 2013; Malach-Pines and Ronen, 2016). Accordingly, to avoid known pitfalls of discrimination and stigmatization surrounding mental health issues in the workforce (Purvanova and Muros, 2010), results from this study invite us to consider more thoroughly the issue of gender inequalities in occupational mental health. From an OHS prevention standpoint, our study thus reinforces the key message that high-quality work conditions are a sine qua non condition for a healthy and balanced workforce and that strategic actions in that sense could relevantly begin with an in-depth understanding of the specificity of working men and women’s reality in the first place.

Declaration of Publication

This study was supported by the Canadian Health Research Institutes (Grant number: 201309MOP-309828-GSH-CFCA-155960). The authors also thank Manulife for their help in workplace recruitment, and Marie-Eve Blanc and Julie Dextras-Gauthier for the field work.The authors declare no conflict of interest relating to the material presented in this Article. Its contents, including any opinions and/or conclusions expressed, are solely those of the authors.

Conflict of Interest

No conflicting interests are declared.

References

  1. Altintas E, Sullivan O (2016)Fifty years of change updated: cross-national gender convergence in housework. Demogr Res; 35: 455. [Google Scholar]
  2. Armon G, Melamed S, Shirom A et al. (2010)Elevated burnout predicts the onset of musculoskeletal pain among apparently healthy employees. J Occup Health Psychol; 15: 399–408. [DOI] [PubMed] [Google Scholar]
  3. Bährer-Kohler S. (2012)Introduction. Burnout for experts: Prevention in the context of living and working. Basel, Switzerland: Springer Science & Business Media: pp. 1–14. Available at http://www.springer.com/gp/book/9781461443902. [Google Scholar]
  4. Baruch Y, Holtom BC (2008)Survey response rate levels and trends in organizational research. Hum Relat; 61: 1139–60. [Google Scholar]
  5. Beauregard N, Marchand A, Blanc ME (2011)What do we know about the non-work determinants of workers’ mental health? A systematic review of longitudinal studies. BMC Public Health; 11: 439. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Blanch A, Aluja A (2012)Social support (family and supervisor), work-family conflict, and burnout: sex differences. Hum Relat; 65: 811–33. [Google Scholar]
  7. Blom V, Sverke M, Bodin L et al. (2014)Work-home interference and burnout: a study based on Swedish twins. J Occup Environ Med; 56: 361–6. [DOI] [PubMed] [Google Scholar]
  8. Brenninkmeijer V, VanYperen N (2003)How to conduct research on burnout: advantages and disadvantages of a unidimensional approach in burnout research. Occup Environ Med; 60(Suppl 1): i16–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Brewer EW, Shapard L (2004)Employee burnout: a meta-analysis of the relationship between age or years of experience. Hum Resour Dev Rev; 3: 102–23. [Google Scholar]
  10. Campos-Serna J, Ronda-Pérez E, Artazcoz L et al. (2013)Gender inequalities in occupational health related to the unequal distribution of working and employment conditions: a systematic review. Int J Equity Health; 12: 57. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Crawford ER, Lepine JA, Rich BL (2010)Linking job demands and resources to employee engagement and burnout: a theoretical extension and meta-analytic test. J Appl Psychol; 95: 834–48. [DOI] [PubMed] [Google Scholar]
  12. Demerouti E, Bakker AB, Nachreiner F et al. (2001)The job demands-resources model of burnout. J Appl Psychol; 86: 499–512. [PubMed] [Google Scholar]
  13. DePasquale N, Mogle J, Zarit SH et al. (2017)The family time squeeze: perceived family time adequacy buffers work strain in certified nursing assistants with multiple caregiving roles. Gerontologist, gnw191. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Egan M, Bambra C, Thomas S et al. (2007)The psychosocial and health effects of workplace reorganisation. 1. A systematic review of organisational-level interventions that aim to increase employee control. J Epidemiol Community Health; 61: 945–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Gagnon S. (2008)La surqualification: Qui la vit? Où s’observe-t-elle?Travail et rémunération; 9: 1–5. [Google Scholar]
  16. Greenhaus JH, Beutell NJ (1985)Sources of conflict between work and family roles. Acad Manag Rev; 10: 76–88. [Google Scholar]
  17. Gutek BA, Searle S, Klepa L (1991)Rational versus gender-role explanations for work family conflict. J Appl Psychol; 76: 560–8. [Google Scholar]
  18. Harvey S, Blouin C, Stout D (2006)Proactive personality as a moderator of outcomes for young workers experiencing conflict at work. Personal Individ Differ; 40: 1063–74. [Google Scholar]
  19. Hu Lt, Bentler PM (1999)Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Struct Equ Model; 6: 1–55. [Google Scholar]
  20. Karasek R, Brisson C, Kawakami N et al. (1998)The Job Content Questionnaire (JCQ): an instrument for internationally comparative assessments of psychosocial job characteristics. J Occup Health Psychol; 3: 322–55. [DOI] [PubMed] [Google Scholar]
  21. Karasek R, Theorell T (1990)Healthy work: stress, productivity, and the reconstruction of working life. New York: Basic Books. [Google Scholar]
  22. Kocalevent RD, Klapp BF, Albani C et al. (2014)Gender differences in a resources-demands model in the general population. BMC Public Health; 14: 902. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. LaMontagne AD, Martin A, Page KM et al. (2014)Workplace mental health: developing an integrated intervention approach. BMC Psychiatry; 14: 131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Leiter MP. (1990)The impact of family resources, control coping, and skill utilization on the development of burnout: a longitudinal study. Hum Relat; 43: 1067–83. [Google Scholar]
  25. Leiter MP, Maslach C (2003)Areas of worklife: a structured approach to organizational predictors of job burnout. In Perrewe PL, Ganster DC, editor. Emotional and physiological processes and positive intervention strategies (Research in Occupational Stress and Well-being, Volume 3). Amsterdam; London: Emerald Group Publishing Limited; pp. 91–134. [Google Scholar]
  26. Luksyte A, Spitzmueller C (2011)Overqualified women: what can be done about this potentially bad situation?Ind Organ Psychol, 4: 256–9. [Google Scholar]
  27. Malach-Pines A, Ronen S (2016)Gender differences in burnout: a literature review. In Antoniou AC, Copper CL, editors. New directions in organizational psychology and behavioral medicine. New York: Routledge; pp. 107–22. [Google Scholar]
  28. Marchand A, Bilodeau J, Demers A et al. (2016)Gendered depression: vulnerability or exposure to work and family stressors?Soc Sci Med; 166: 160–8. [DOI] [PubMed] [Google Scholar]
  29. Marchand A, Durand P, Haines V III et al. (2015)The multilevel determinants of workers’ mental health: results from the SALVEO study. Soc Psychiatry Psychiatr Epidemiol; 50: 445–59. [DOI] [PubMed] [Google Scholar]
  30. Martinko MJ, Harvey P, Brees JR et al. (2013)A review of abusive supervision research. J Organ Behav 34(Suppl 1): S120–37. [Google Scholar]
  31. Maske UE, Riedel-Heller SG, Seiffert I et al. (2016)Prevalence and comorbidity of self-reported diagnosis of burnout syndrome in the general population—results of the German health interview and examination survey for adults (DEGS1). Psychiatr Prax; 43: e1. [DOI] [PubMed] [Google Scholar]
  32. Maslach C, Schaufeli WB, Leiter MP (2001)Job burnout. Annu Rev Psychol; 52: 397–422. [DOI] [PubMed] [Google Scholar]
  33. Mauno S, Cheng T, Lim V (2017)The far-reaching consequences of job insecurity: a review on family-related outcomes. Marriage Fam Rev; 53: 717–43. [Google Scholar]
  34. Michel JS, Kotrba LM, Mitchelson JK et al. (2011)Antecedents of work–family conflict: a meta‐analytic review. J Organ Behav; 32: 689–725. [Google Scholar]
  35. Missler M, Stroebe M, van der Laan G (2013)The work–home interface: the role of home-based predictors of burnout among mothers. Family Science; 4: 148–60. [Google Scholar]
  36. Mùthen LK, Mùthen B (2006)Mplus user’s guide. 4.1 edn. Los Angeles: Mùthen, L.K. & Mùthen, B. [Google Scholar]
  37. Nahrgang JD, Morgeson FP, Hofmann DA (2011)Safety at work: a meta-analytic investigation of the link between job demands, job resources, burnout, engagement, and safety outcomes. J Appl Psychol; 96: 71–94. [DOI] [PubMed] [Google Scholar]
  38. Norlund S, Reuterwall C, Höög J et al. (2010)Burnout, working conditions and gender–results from the northern Sweden MONICA Study. BMC Public Health; 10: 326. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Paris M Jr, Hoge MA (2010)Burnout in the mental health workforce: a review. J Behav Health Serv Res; 37: 519–528. [DOI] [PubMed] [Google Scholar]
  40. Pearlin LI, Schooler C (1978)The structure of coping. J Health Soc Behav; 19: 2–21. [PubMed] [Google Scholar]
  41. Perry‐Jenkins M, Wadsworth SM (2017)Work and family research and theory: review and analysis from an ecological perspective. J Fam Theory Rev; 9: 219–37. [Google Scholar]
  42. Purvanova RK, Muros JP (2010)Gender differences in burnout: a meta-analysis. J Vocat Behav; 77: 168–85. [Google Scholar]
  43. Reichl C, Leiter MP, Spinath FM (2014)Work-nonwork conflict and burnout: a meta-analysis. Hum Relat; 67: 979–1005. [Google Scholar]
  44. Risman BJ, Davis G (2013)From sex roles to gender structure. Curr Sociol; 61: 733–55. [Google Scholar]
  45. Rosenberg M. (1979)Conceiving the self. New York: Basic Books. [Google Scholar]
  46. Rosse JG, Boss RW, Johnson AE et al. (1991)Conceptualizing the role of self-esteem in the burnout process. Group Organ Stud; 16: 428–51. [Google Scholar]
  47. Schaufeli WB, Leiter MP, Maslach C et al. (1996)The Maslach Burnout Inventory-General Survey. In Maslach C, Jackson SE, Leiter MP, editors. MBI Manual. 3rd edn. Palo Alto, CA: Consulting Psychologists Press. [Google Scholar]
  48. Scott J, Plagnol AC (2012)Work-family conflict and wellbeing in northern Europe. Gendered Lives: Gender Inequalities in Production and Reproduction. Northampton, MA: Edward Elgar Publishing; pp. 174–205. [Google Scholar]
  49. Seidler A, Thinschmidt M, Deckert S et al. (2014)The role of psychosocial working conditions on burnout and its core component emotional exhaustion—a systematic review. J Occup Med Toxicol; 9: 10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Sherman AC, Higgs GE, Williams RL (1997)Gender differences in the locus of control construct. Psychology & Health; 12: 239–48. [Google Scholar]
  51. Siegrist J. (2002)Effort-reward imbalance at work and health. In Perrewé P, Ganster D, editors. Historical and current perspectives on stress and health. Amsterdam: Elsevier Science; pp. 261–91. [Google Scholar]
  52. Siegrist J, Peter R (1996)Measuring effort-reward imbalance at work: Guidelines. Düsseldorf: Heinrich Heine University. [Google Scholar]
  53. Siegrist J, Peter R (2000)The effort-reward imbalance model. Occup Med; 15: 83–7. [Google Scholar]
  54. Sperlich S, Peter R, Geyer S (2012)Applying the effort-reward imbalance model to household and family work: a population-based study of German mothers. BMC Public Health; 12: 12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Springer KW, Hankivsky O, Bates LM (2012)Gender and health: relational, intersectional, and biosocial approaches. Soc Sci Med; 74: 1661–6. [DOI] [PubMed] [Google Scholar]
  56. Stansfeld S, Candy B (2006)Psychosocial work environment and mental health–a meta-analytic review. Scand J Work Environ Health; 32: 443–62. [DOI] [PubMed] [Google Scholar]
  57. Tabachnick BG, Fidell LS (2013)Using multivariate statistics. 6th edn. New Jersey: Pearson. [Google Scholar]
  58. Tepper BJ. (2000)Consequences of abusive supervision. Acad Manag J; 43: 178–90. [Google Scholar]
  59. Vézina M, Cloutier E, Stock S et al. (2011)Enquête québécoise sur des conditions de travail, d’emploi et de santé et de sécurité du travail (EQCOTESST). Québec, QC: Institut de recherche Robert-Sauvé en santé et en sécurité du travail, Institut national de santé publique du Québec, et Institut de la statistique du Québec. [Google Scholar]
  60. Walsh J. (1999)Myths and counter-myths: an analysis of part-time female employees and their orientations to work and working hours. Work Employ Soc; 13: 179–203. [Google Scholar]
  61. Wang Q, Bowling NA, Eschleman KJ (2010)A meta-analytic examination of work and general locus of control. J Appl Psychol; 95: 761–8. [DOI] [PubMed] [Google Scholar]
  62. Wheaton B. (1994)Sampling the stress universe. In Avison WR, Gotlib IH, editors. Stress and mental health: contemporary issues and prospects for the future. New York: Plenum Press; pp. 77–114. [Google Scholar]
  63. Whitehead M, Pennington A, Orton L et al. (2016)How could differences in ‘control over destiny’ lead to socio-economic inequalities in health? A synthesis of theories and pathways in the living environment. Health Place; 39: 51–61. [DOI] [PubMed] [Google Scholar]

Articles from Annals of Work Exposures and Health are provided here courtesy of Oxford University Press

RESOURCES