Correlates of turnover intent among primary (N = 580) and secondary (N = 675), male (N = 254) and female (N = 999) teachers, were examined through the lens of the job demands-resources (JD-R) model. Multigroup structural equation modelling indicated that job demands (workload, student misbehaviour), and the personal demand of work–family conflict, were positively associated with emotional exhaustion—the core dimension of burnout. All demands indirectly related to turnover intent via emotional exhaustion. Among all teacher groups, no significant differences were found in level of emotional exhaustion or turnover intent, and only mild stress was reported as a result of student misbehaviour. Work–family conflict was the strongest predictor of emotional exhaustion for male and female teachers. Results suggest the JD-R as a promising theory for use in explaining job-related outcomes among Australian teachers, and that personal demands should be examined in addition to job demands within it.
Teaching is an emotionally taxing and stressful occupation (Kyriacou 2001) with empirical studies confirming that stress brought about by demands of the job is predictive of emotional exhaustion, burnout (see Demerouti et al. 2001; Schaufeli and Bakker 2004), and other negative outcomes such as attrition (Borman and Dowling 2008). Teacher attrition rates worldwide are generally high (e.g. Craig 2017). Although in Australia definitive figures are hard to gauge (Weldon 2018), an estimated 30% to 40% of teachers leave the profession within their first five years (Paris 2010). This early departure from the profession can partly be attributed to burnout (Goddard and Goddard 2006; Ingersoll 2012) which has detrimental implications for teachers, students, and the wider school community. Teacher shortages themselves may trigger a cycle wherein the shortage aggravates other teachers’ stress and burnout (Farber 1991). Retaining quality teachers is central to student achievement, as teacher effects impact children reaching their full potential (e.g. Sanders and Rivers 1996).
The core dimension of burnout is emotional exhaustion (Taris et al. 2005), which is a chronic state of emotional and physical depletion that results from excessive job demands and continuous stress (Wright and Cropanzano 1998). It describes a feeling of being emotionally overextended and exhausted by one's work and is manifested by both physical fatigue and a sense of feeling psychologically and emotionally "drained" (Zohar 1997). Generally, women tend to report significantly higher levels of emotional exhaustion than men (Lorente et al. 2008; Skaalvik and Skaalvik 2017), perhaps because of their roles outside work such as caring for the family.
It is well established that motivation to leave the teaching profession is in part predicted by stressful working conditions (e.g. Klassen and Chiu 2011). For example, a correlational study of 2569 Norwegian elementary and middle school teachers (Skaalvik and Skaalvik 2011) revealed that time pressure and discipline problems indirectly predicted motivation to leave the profession as well as job satisfaction, via emotional exhaustion. Although many aspects of the teaching profession contribute to teachers’ level of emotional exhaustion, two factors intrinsic to the profession known to facilitate high levels of stress are workload and student misbehaviour (Greenglass and Burke 2003). In a comprehensive study of 20 sources of teacher stress, job satisfaction, and career commitment, workload and student misbehaviour accounted for the most variance in predicting teacher stress among a sample of 710 full-time Mediterranean primary school teachers (Boyle et al. 1995). Among Australian secondary school teachers, although both workload and student misbehaviour were sources of stress, student misbehaviour was found to be more central to predicting burnout (McCormick and Barnett 2011). And, in a Canadian sample of 951 elementary and secondary teachers, women reported significantly higher levels of stress from workload and student misbehaviour than men (Klassen 2010).
Based on international figures in the TALIS 2013 report which details the average amount of time spent on certain tasks, Australian teachers spent comparatively large amounts of time on paperwork, general administrative work, and meetings with colleagues within the school (OECD 2014). Further, more than 20% of each lesson’s allocated time was spent on other things besides actual teaching and learning: lower secondary teachers spent an average of 14.5% of time maintaining order in the classroom (Freeman et al. 2014). The extra tasks that teachers need to do over and above their teaching highlight time pressure as a great source of teachers’ stress. Unequivocal empirical evidence shows a positive link between time pressure and emotional exhaustion or burnout (Kokkinos 2007; Skaalvik and Skaalvik 2011, 2016).
Teachers’ in-school workload necessitates working long hours outside of work (e.g. marking/lesson plans), which in turn can impact time with their family. This work–home interference (work–family conflict) can cause additional stress (e.g. Evans et al. 2013) adding to the stressors teachers experience. Stress from one domain of life cannot be segmented from having an influence on another (Kanter 1977). Concordantly, we argue that any theory dealing with occupational stress needs to consider what impact stress deriving from family life has on job stressors, since people cannot always suppress family-related thoughts, feelings, and behaviours in the work domain, and vice versa. Although there is abundant literature suggesting that increased stress in personal or family life extends into, and negatively affects the roles held by individuals in the workplace (Cohen et al. 2008; Kiger et al. 2007), and that work–family conflict is positively related to burnout and often associated with greater inclination to leave a job (Cohen et al. 2008; Frone et al. 1997; Hang-yue and Loi 2005; Lee and Ashforth 1996), there is a paucity of research on how it impacts Australian teachers.
In other professions, work–family conflict is positively related to time pressure (Brayfield et al. 2005; Voydanoff 2005), burnout, and turnover intention (Greenhaus et al. 2001; Thanacoody et al. 2009). For example, a study of 114 cancer clinicians (comprising radiotherapists, radiographers, and oncologists) found that burnout mediated the relationship between work–family conflict and intention to leave the job. High levels of work–family conflict increased levels of burnout (emotional exhaustion and disengagement) among cancer clinicians which, in turn, increased turnover intention (Thanacoody et al. 2009). Similarly, a study of 1000 public accountants revealed that work–family conflict significantly affected intent to leave the job (Greenhaus et al. 1997). This ‘spillover effect’ of stress from one domain to another seems to affect men and women differently, with women reporting higher spillover (Erdamar and Demirel 2014; Voydanoff 2005), perhaps due to their traditional roles within the family (i.e. assuming responsibility for the household and children). However, this finding has not always been consistent. For example, Cook and Minnotte (2008) found no significant difference in level of spillover between men and women, in their analysis encompassing the range of professions from the U.S. 2002 National Study of the Changing Workforce (N = 2810).
Studies of teacher burnout and occupational outcomes have had criticism levelled against them for not investigating secondary school teachers as much as primary school teachers and, more importantly, for not being grounded in nuanced theories which consider changes in the reasons for burnout and attrition, and that the rewards of teaching are likely to change across teachers’ career paths (see Borman and Dowling 2008). This study utilises the job demands-resources (JD-R) theoretical model to investigate Australian teachers’ emotional exhaustion and turnover intent.
Job demands-resources theory (JD-R)
The JD-R theory proposes that employee well-being relates to a range of workplace conditions/characteristics that can be conceptualised as either job demands (i.e. the physical, social, or organisational aspects of the job that require sustained physical or psychological effort) or job resources (i.e. those aspects of the job that may reduce job demands, are instrumental to achieve work goals, or promote personal growth, learning, and development) (Bakker and Demerouti 2007). Excessive job demands, and lacking job resources, exert an energy-draining effect on employees through a stress process (burnout) which leads to negative work outcomes such as turnover, while high levels of job resources relate to positive work outcomes through a motivational process, and can buffer the relationship between job demands and burnout (Rothmann and Joubert 2007). This study’s focus is on the burnout pathway.
Extensive empirical support exists for the stress and motivational processes of the JD-R model, establishing that working in a demanding job and having few job resources associates with burnout (Llorens et al. 2006), sickness absences (Schaufeli et al. 2009), ill-health (Hakanen et al. 2006), and health complaints (Korunka et al. 2009). The current study focuses on the main tenet of the JD-R theory—namely, that emotional exhaustion as the core dimension of burnout (Taris et al. 2005), mediates the relationship between job demands and occupational outcomes. Further, we extend current research by incorporating a personal demand (work–family conflict) to the model, proposing that the JD-R would be strengthened by incorporating personal stressors outside the work domain, as stressors for most people cannot be compartmentalised, leading to a spillover effect and compounding stressors at work.
While there are known differences in the teaching environments of primary and secondary teachers, there are relatively few Australian teacher burnout studies examining both simultaneously. The current study compares primary and secondary teachers’ experienced demands, emotional exhaustion, and turnover intent.
The present study
In the present study, we explored relationships between primary versus secondary, and male versus female teachers’ perceptions of stressors caused by both job (workload, student misbehaviour) and personal (work–family conflict) demands, their levels of emotional exhaustion, and intent to leave the teaching profession. Based on the JD-R theory and the reviewed studies, it was hypothesised that individual stressors/demands would interrelate (H1); would predict emotional exhaustion (H2); and that emotional exhaustion would predict greater intent to leave the teaching profession (H3). We expected demands would indirectly relate to intent to leave the teaching profession, mediated by emotional exhaustion (H4). Finally, we expected secondary teachers to report higher levels of emotional exhaustion than primary teachers (H5). We also explored differences in level and association, between primary versus secondary, and male versus female teachers’ workload, stress from student misbehaviour, work–family conflict, emotional exhaustion, and turnover intent.
Participants and procedure
Participants were 580 primary (83 men, 495 women, 2 unspecified gender) and 675 secondary school teachers (171 men, 504 women) working across NSW and Victoria. Women ranged in age from 21 to 73 years (M = 40.24, SD = 11.34), and had a range of teaching experience from < 1.00 to 50.30 years (M = 14.11, SD = 11.12). Men ranged in age from 22 to 74 years (M = 41.08, SD = 11.28), and had from < 1.00 to 52.50 years’ experience (M = 14.60, SD = 11.40). Participants were Australian Education, and Independent Education, union members who were invited to participate in a 20-min, online anonymous survey via a link, advertised in 2nd term (May 2017) on the Unions’ websites and Facebook pages.
Workload was assessed using the workload inventory (Spector and Jex 1998), a psychometrically sound five-item inventory, designed to assess the amount of work and work pace. An example item is “How often is there a great deal to be done?” Each item is rated using a five-point response format, where 1 = less than once per month or never, 2 = once or twice per month, 3 = once or twice per week, 4 = once or twice per day, and 5 = several times a day. Higher scores represent a higher level of workload, or greater work demands. Good reliability and evidence of validity was reported by the authors, with an average internal consistency (Cronbach’s coefficient α) of 0.82 across 15 studies. Reliability in the current subsamples was good (see Table 1).
Student misbehaviour was assessed using the pupil misbehaviour subscale from the Teacher Stress Scale (Boyle et al. 1995). This five-item scale asks respondents to rate how great a source of stress items such as “noisy pupils” are to them. All responses were scored on a five-point Likert-type ordinal scale, with response options ranging from 'No stress', 'Mild stress', 'Moderate stress', 'Much stress', to 'Extreme stress' (with corresponding scores 0, 1, 2, 3, and 4). Higher scores correspond with greater stress from perceived student misbehaviour. The scale has been successfully used in a sample of elementary and secondary Canadian school teachers (Klassen 2010). Adequate reliability (i.e. α > 0.70) and evidence of validity were found for this subscale in previous work (e.g. Boyle et al. 1995; Klassen 2010). Reliability in the current subsamples was good (see Table 1).
Work–family conflict was assessed by the work–family subscale from the Work–Family Conflict and Family–Work Conflict Scale (Netemeyer et al. 1996). This subscale (5 items) assesses conflict in which the general demands of time devoted to, and strain created by the job, interfere with performing family-related responsibilities. Each item is rated on a 7-point Likert scale, with 1 indicating ‘strongly disagree’ and 7 ‘strongly agree’. An example item is “My job produces strain that makes it difficult to fulfil family duties”. Higher scores correspond with greater conflict. The psychometrically sound scale has been validated on various samples, including a sample of elementary and high school American teachers (Netemeyer et al. 1996) with good reliability reported by the authors (α = 0.88). Reliability in the current subsamples was good (see Table 1).
Burnout was measured using the emotional exhaustion subscale from the Maslach Burnout Inventory–Educators’ Survey (MBI–ES: Maslach et al. 1996). The emotional exhaustion subscale was designed to measure feelings of being emotionally overextended and exhausted by one’s work. The nine-item subscale asks respondents to rate how frequently they experience such feelings as fatigue, frustration, and interpersonal stress in their job. Each item (e.g. “I feel emotionally drained from my work”Footnote 1) is rated on a 7-point response scale, where 1 = never, 2 = a few times a year, 3 = once a month or less, 4 = a few times a month, 5 = once a week, 6 = a few times a week, and 7 = every day. Higher scores correspond with greater levels of emotional exhaustion. Previous research using the MBI-ES to estimate Australian teachers’ level of emotional exhaustion has consistently reported good reliability levels in the range of α = 0.89 to 0.92 (Goddard and Goddard 2006; Goddard and O'Brien 2003; Goddard et al. 2006; Richardson et al. 2013). Reliability in the current subsamples was good (see Table 1).
Intent to leave the teaching profession (turnover intent)
Intent to leave the teaching profession was assessed using the Intention to Turnover subscale, from the Michigan Organisational Assessment Questionnaire (Cammann et al. 1979). This three-item subscale was designed to measure the perceptions of organisational members about their psychological state relevant to quality of work life issues in the workplace. Good reliability was reported by the authors, α = 0.85. The first item (“What is the likelihood of you looking for a different job, other than teaching, in the next year?”) was rated (1) ‘not likely’ to (7) ‘extremely likely’; the second and third items were rated (1) ‘strongly disagree’ to (7) ‘strongly agree’. To increase the possible range in scores, the original 5-point response format of the scale was altered to a 7-point scale.Footnote 2 Reliability in the current subsamples was good (see Table 1).
Based on the empirical studies reviewed, gender, teaching experience (in years), and school level (primary/secondary) were included as control variables that could be expected to relate to turnover intent.
Analyses were conducted with Mplus (version 8: Muthén and Muthén 1998–2017). Full information maximum likelihood estimation was used to account for the low amount of missing data (< 0.5%).Footnote 3 Confirmatory factor analyses (CFAs) were first used to evaluate the fit of the measurement model for four teacher groups, in two modelsFootnote 4: the first comparing primary and secondary, the second comparing men and women. Invariance tests were conducted within each of these two (school level and gender) measurement models. Multigroup structural equation modelling (MG-SEM) then tested the posited mediation model across school-level and gender groups. Goodness of fit for all models was assessed using the comparative fit index (CFI; Bentler 1990), the root mean square error of approximation (RMSEA; Bentler 1990; Steiger and Lind 1980), the standardised root mean square residual (SRMR), and the Tucker-Lewis Index (TLI; Tucker and Lewis 1973). CFI and TLI values > 0.90 and RMSEA and SRMSR < 0.08 describe adequate model fit (Byrne 2012).
Fit of measurement model
CFA was used to confirm the hypothesised factor structure of the combined variables—i.e. turnover intent (3 items), emotional exhaustion (9 items), work–family conflict (5 items), workload (5 items), and student misbehaviour (5 items) for each teacher group. Results of the initial CFAs revealed that the data did not fit the model sufficiently for primary or secondary, or for male or female teachers (see “Appendix A”). Adequate fit of the data to the model was achieved (see Table 2) by freeing the error covariance between items 4 and 8 (due to parallel wording) on the emotional exhaustion scale, in all subsamples, confirming the factors corresponded to the hypothesised structure and providing evidence of construct validity.
All measures had good internal consistencies that were similar across subsamples. Cronbach’s α ranged from 0.78 to 0.95 for primary, versus 0.81 to 0.94 for secondary teachers. It ranged from 0.82 to 0.95 for male, versus 0.79 to 0.94 for female teachers (see Table 1).
Next, measurement invariance across each of primary/secondary, and men/women for the 5-factor model was examined, to confirm that the latent constructs had similar psychometric properties across groups. This involved a stepped review of (a) a configural invariance model which indicated the items tapped the same constructs across groups, (b) metric invariance which indicated same item loadings across groups, and (c) scalar invariance which indicated same item intercepts across groups (Cheung and Rensvold 2002). Each model is compared to the prior model for any loss of model fit due to additional equality constraints. For all 3 invariance tests (configural, metric and scalar), a change of 0.01 or lower in CFI, supplemented by either a change of 0.015 or lower in RMSEA or a change of 0.03 or lower in SRMR when sample sizes are uneven, indicates invariance across groups (Cheung and Rensvold 2002).
Fit indices for the three invariance tests across the groups presented in Table 3 indicated that scalar invariance between groups was achieved for the measurement models. This satisfied the condition that each comparison group (i.e. primary vs. secondary, and male vs. female teachers) interpreted measures in the same way, providing the basis for the following multigroup SEM modelling (see “Appendix A”, Figs. 3, 4).
Table 4 shows latent correlations, means, and standard deviations for primary and secondary teachers across the study measures. There were significant positive relationships among all measures. For both samples, teachers who reported high levels of turnover intent also reported high levels of both job and personal demands, and emotional exhaustion, supporting hypotheses H2–H4.
Table 5 shows latent correlations, means, and standard deviations for male and female teachers across the study measures. Except for student misbehaviour which, contrary to prediction (H1), was not associated with workload for males, there were significant positive relationships among all other measures for male and female teachers. Male and female teachers who reported high levels of turnover intent also reported high levels of demands (both job and personal) and emotional exhaustion, supporting hypotheses H2–H3.
Testing the mediation model across primary and secondary teachers
Invariance tests of the mediation model were next performed. First, the mediation model was examined for primary and secondary teachers, controlling for length of teaching experience and gender, and inspecting potential significant mean-level differences between groups. We then repeated these analyses in a second model for male and female teachers, controlling for length of teaching experience and school level. In both models, measurement invariance constraints were retained.
Based on the JD-R model, it was hypothesised that work–family conflict, workload, and student misbehaviour would indirectly relate to turnover intent through emotional exhaustion. Independent variables were work–family conflict, workload, student misbehaviour, and emotional exhaustion. Control variables were gender and teaching experience, and the outcome was turnover intent. The non-bias-corrected bootstrap was used to account for non-normality in the data which is known to occur when testing for intervening variable effects (Hayes 2009). This approach generally also produces preferable confidence limits and standard errors for the indirect effect test (Fritz et al. 2012). The indirect effect is significant when the range between the lower and upper limit does not include zero.
The first model for primary and secondary teachers revealed the data showed an adequate fit for both groups: primary—χ2(363, N = 580) = 1041.562, p < 0.001, CFI = 0.936, TLI = 0.928, RMSEA = 0.058 (CI 0.054–0.062) and SRMR = 0.057; secondary—χ2(363, N = 675) = 1214.811, p < 0.001, CFI = 0.932, TLI = 0.924, RMSEA = 0.061 (CI 0.057–0.065) and SRMR = 0.065. As shown in Fig. 1, for both groups, all relationships were positive, and all demands (work–family conflict, workload, student misbehaviour) were significantly interrelated. Job demands (workload and student misbehaviour) were significantly related, such that higher levels of workload were associated with higher perceived levels of student misbehaviour. Both job and personal demands were significantly related such that higher levels of student misbehaviour and workload associated with higher levels of work–family conflict. Work–family conflict, workload, and student misbehaviour significantly predicted emotional exhaustion. In turn, emotional exhaustion significantly predicted turnover intent.
The indirect effects tested using bootstrapped standard errors, across both primary and secondary teachers (Table 6), were significant as predicted—work–family conflict, workload, and student misbehaviour were each related to turnover intent via emotional exhaustion.
These findings support the hypothesised mediation model for both teacher groups. The models explained 40.3% versus 43.2% of the total variance in emotional exhaustion, and 24.6% versus 22.3% of the total variance in turnover intent, for primary and secondary teachers, respectively.
The model for primary and secondary teachers, controlling for gender and teaching experience, revealed the predictive paths (Table 7) were not significantly different across the measures predicting turnover intent. There were no significant mean differences between primary and secondary teachers for turnover intent (z = 0.734, p = 0.463), emotional exhaustion (z = 0.375, p = 0.707), or work–family conflict (z = 14.993, p = 0.360). However, there were significant differences for workload (z = 11.302, p < 0.001; Cohen’s dFootnote 5 = 0.17, 95% CI (0.06–0.29), and student misbehaviour (z = 12.152, p < 0.001, Cohen’s d = 0.23, 95% CI (0.12–0.34). Although the effect sizes were small (Cohen 1988), primary teachers reported higher levels (Table 4).
Testing the mediation model across male and female teachers
The theoretical model was next tested for men and women, controlling for length of teaching experience and strand. Results revealed adequate fit of the data to the model for both groups: men—χ2(363, N = 254) = 764.697, p < 0.001, CFI = 0.921, TLI = 0.912, RMSEA = 0.066 (CI 0.060–0.073) and SRMR = 0.070; women—χ2(363, N = 999) = 1582.956, p < 0.001, CFI = 0.934, TLI = 0.927, RMSEA = 0.059 (CI 0.056–0.062) and SRMR = 0.059.
As shown in Fig. 2, the hypothesised model was supported for women, as all demands were interrelated and indirectly predicted turnover intent via emotional exhaustion. For men, the model was partially supported; workload and work–family conflict demands were positively related, and indirectly predicted turnover intent via emotional exhaustion. For men, student misbehaviour was unrelated to the other demands, but still indirectly related to turnover intent via emotional exhaustion. The indirect effects (Table 8) were significant as predicted, both for male and female teachers. For men and women, work–family conflict had the strongest indirect effect on turnover intent across the three predictors. The models explained 40.3% versus 42.8% of the total variance in emotional exhaustion, and 28.0% versus 22.0% of the total variance in turnover intent, for male and female teachers, respectively.
The modelling of male versus female teachers, controlling for teaching experience and school level, revealed that predictive paths (Table 9) similarly predicted turnover intent across both gender groups. While there were no significant mean differences between male and female teachers on either turnover intent (z = 1.768, p = 0.077) or emotional exhaustion (z = 0.595, p = 0.552), there were differences in mean work–family conflict (z = 2.745, p = 0.006, Cohen’s d = 0.21, 95% CI (0.08–0.35), workload (z = 2.549, p = 0.011, Cohen’s d = 0.21, 95% CI (0.07–0.35), and student misbehaviour (z = 2.679, p = 0.007, Cohen’s d = 0.20, 95% CI (0.07–0.34)) (see Table 5); female teachers reported higher levels compared to males.
Findings from the present study extend previous research investigating teachers’ burnout and turnover intent. Relationships between primary versus secondary, and male versus female teachers’ perceptions of stressors caused by both job (workload, student misbehaviour) and personal (work–family conflict) demands, as well as their experience of emotional exhaustion and intent to leave the teaching profession, were explored through the lens of the JD-R model. Findings revealed the JD-R model evidenced good fit for teachers across school-level (primary/secondary) and gender groups (men/women), and with the inclusion of the work–family conflict personal demand.
Comparison of primary and secondary teachers
For both primary and secondary teachers, all demands were positively interrelated (H1) and positively predicted emotional exhaustion (H2), which in turn promoted turnover intent (H3). All demands (job and personal) indirectly related to intent to leave the teaching profession, mediated through emotional exhaustion (H4). These findings support the numerous studies indicating that burnout mediates the relationship between job demands and turnover intent (e.g. Skaalvik and Skaalvik 2011, 2017; Wright and Cropanzano 1998). The finding that the personal demand of work–family conflict made a significant contribution to the model highlights the need to also include personal demands in investigation of burnout among school teachers, particularly as their work is not confined to their work site, and as teaching is a female-dominated occupation. Working women who have families, and/or those having to care for parents, potentially face the issues of juggling work and family, particularly if they are responsible for managing the home.
Although primary teachers reported significantly higher levels of workload and student misbehaviour than secondary teachers, contrary to expectation (H5), there was no significant difference in their level of emotional exhaustion. Inconsistent findings exist in the literature as to whether primary or secondary teachers suffer more burnout (e.g. Tejedor and Rigotti 2008). For example in a French study (Vercambre et al. 2009; N = 2558), researchers found that elementary teachers were more susceptible to higher emotional exhaustion than high school teachers. However, contrasting results were found in a Spanish study (Betoret 2009). No significant differences were evident in our study.
There were also no significant differences between primary and secondary teachers in their levels of work–family conflict or turnover intent. Both primary and secondary teachers strongly agreed that their job created strain which makes it difficult to fulfil family duties, that plans for family activities often needed to be changed, and neither teacher group reported high levels of turnover intent.
Comparison of men and women
The JD-R model also evidenced good fit across gender groups (again supporting H2–H4). For men, student misbehaviour did not significantly relate to the job demand of workload or to the personal demand of work–family conflict (contradicting H1), although it still indirectly affected turnover intention through emotional exhaustion (supporting H4). A premise of the JD-R theory is that job demands are positively interrelated. However, perhaps men are better at dealing with misbehaving students, in line with their lower ratings of this demand than women teachers, and at not letting student misbehaviour interfere with getting on with the job. Although student misbehaviour affected men’s emotional exhaustion, it may not alter their perception of workload. The non-significant relationship between student misbehaviour and work–family conflict for men could suggest men are less bothered by student misbehaviour and can better disregard it than women. This could also be connected to the different relational and nurturing skills expected of men and women (e.g. Spilt et al. 2012).
Contrary to expectation, no differences were found between men’s and women’s level of emotional exhaustion (H5). This finding contrasts with numerous previous studies of teachers (Antoniou et al. 2006; Byrne 1991; Lau et al. 2005; Maslach et al. 1996; Vercambre et al. 2009), which found women reported higher levels of emotional exhaustion than men. Perhaps in Australian culture, men are better at sharing responsibilities in the home (e.g. helping with chores),Footnote 6 which would give women some relief and reduce their overall levels of emotional exhaustion. In France for instance, women are mainly responsible for housework and children, even if they are employed outside the home (e.g. Vercambre et al. 2009), which could help explain the inconsistent findings.
No differences were found in turnover intent between male and female teachers. This finding contrasts with that of Borman and Dowling (2008) who, in their meta-analysis of 34 studies, found the odds of attrition were higher among female teachers, and may reflect the mean age of our sample, which for both men and women was over 40. Teachers in this age bracket are less likely to leave teaching (Borman and Dowling 2008).
Work–family conflict was the strongest predictor of emotional exhaustion for both men and women, indicating that time devoted to, and strain created by, the job interfering with family-related responsibilities more importantly predicts emotional exhaustion than the workload itself. Level of work–family conflict was significantly higher for women than men, which concurs with the findings of Erdamar and Demirel (2014) and Voydanoff (2005), but contrasts with results from a meta-analysis (N > 250,000 workers; Shockley et al. 2017) which found no significant differences in level of work–family conflict between men and women, working in various occupations (i.e. not specifically related to teachers).
Interestingly, results from the few overseas studies on school teachers are also conflicting. While no significant differences in level of work–family conflict based on gender was found in a small study of 130 school teachers in Malaysia (Panatik et al. 2011), differences were noted in a Turkish study of 240 primary and 124 secondary teachers where women and young teachers reported higher levels of work–family conflict (Erdamar and Demirel 2014). These inconsistent findings suggest different mechanisms at play and unless studies match contextual factors across samples (e.g. in terms of age, life-stage, and culture), no general conclusions can be drawn.
Across all teacher groups, only mild stress was reported as a result of student misbehaviour and it was the least important predictor of emotional exhaustion. While this finding contrasts with that of McCormick and Barnett (2011), who found student misbehaviour was central to predicting burnout among secondary teachers, our findings may reflect the increasing workload teachers face which may overpower the stress experienced from their other job demands.
Implications and future research
Overall, findings from this investigation suggest that different strategies are needed for different teacher groups to help ameliorate burnout and reduce turnover intent. Strategies dealing more with reducing workload are needed for secondary teachers, whereas strategies dealing with reducing work–family conflict are needed for primary teachers, regardless of gender. The theoretical perspective of the JD-R model describes the relationship between demands and exhaustion as ‘energy-driven’, by which individuals suffer because their energy has been depleted by demands. Our findings highlight that the stress perceived from demands both within and outside the job domain must be jointly investigated, because the various stressors encountered in life do not disappear when one leaves home or work, but each may exacerbate the effect of stressors within the other domain. Additional personal demands could be productively examined through the JD-R lens in future research. For example, time commuting to and from work can impact job stress (Costa et al. 1988; Flood and Barbato 2005). In Australia, those working full-time spend an average of approximately 4.5 hours a week commutingFootnote 7 (Wilkins et al. 2019); such wasted time can exacerbate the negative impact of long working hours and work stress on interpersonal relationships and family lives. Role strain theory (Goode 1960) argues that inter-role conflict results in an undesirable state owing to conflicting demands on time, behaviour, and energy among roles (Greenhaus and Beutell 1985; Kahn et al. 1964). As people have a finite total amount of energy, it is logical that any occupational stress model needs to apply an holistic approach, especially when, in jobs such as teaching, work is not only completed at the workplace. The inclusion of the personal demand of work–family conflict, and the finding that for men and women it was the strongest predictor of emotional exhaustion, highlights the importance of including stressors accumulated in life domains other than at work, for investigating burnout and negative occupational outcomes.
Several limitations need to be considered when interpreting the results of this study. As with any self-report study, one needs to be cognisant of common method bias. The findings may also not be generalisable to the broader teacher population due to the use of a purposive sample. Participants were likely to be teachers who identified with feeling burned out, or who had a specific interest in burnout. Disappointingly, the small sample of men particularly among primary teachers limited our ability to make gender comparisons across school levels; the unequal gender ratio comprising mainly women participants, although typical of research involving teachers, may make the results less applicable to male teachers. Finally, although theoretically driven, the correlational data need to be interpreted cautiously in terms of the directionality of the results implied by the structural model. Longitudinal studies using more representative teacher samples are needed to clarify the directionality of relationships among the variables, and further our understanding of the causal ordering of stressors teachers endure (both personally and on the job), and how these affect burnout and turnover intent.
The study supports use of the JD-R model as a helpful theory for explaining job-related negative outcomes among teachers in an Australian sample, and highlights the importance of including personal demands when investigating how stressors (demands) impact occupational outcomes. Stressors, regardless of where they originate, can affect overall stress levels to impact work outcomes. Demands from family will play a role in teachers’ overall level of stress, as increased stress in family life can extend into and negatively affect roles held in the workplace (Kiger et al. 2007). Creating policies and a culture that supports the integration of work and family could have positive consequences for teachers. Secondary teachers especially could also benefit from attention to workload and coping strategies to reduce emotional exhaustion. One strategy which the OECD (2018) positions as promising for reducing stress is mindfulness. Mindfulness-based stress reduction courses adapted specifically for teachers have had promising results in alleviating psychological symptoms and burnout (Flook et al. 2013; Gold et al. 2010; Roeser et al. 2012; Todd et al. 2019). Attention to reducing teachers’ experienced demands should reduce their emotional exhaustion, and, in turn, intent to leave the profession.
Copyright ©1996 Christina Maslach, Susan E. Jackson & Richard L. Schwab. All rights reserved in all media. Published by Mind Garden, Inc., www.mindgarden.com.
Increasing a scale’s response format from 5 to 7 does not make any appreciable difference to the distributional properties (skewness/kurtosis), nor does it destroy the comparability of historical data (Dawes 2008).
Little’s MCAR test was not significant, χ2(616) = 686.69, p > 0.05.
Due to low numbers of male primary teachers (n = 83), it was not possible to cross gender by school level.
Cohen’s d provides a metric for comparing effects between groups that is not biased by sample size.
In Australia, household tasks have become less gendered over time, and as men become more egalitarian in their attitudes, they become more willing to do what was traditionally (e.g. cooking, cleaning, laundry) known as women’s work (Chesters 2012).
The average travel time across Australian cities ranges from approximately 3 (Northern Territory) to 6 h (Sydney).
Antoniou, Α. S., Polychroni, F., & Vlachakis, Α. Ν. (2006). Gender and age differences in occupational stress and professional burnout between primary and high-school teachers in Greece. Journal of Managerial Psychology,21(7), 682–690.
Bakker, A. B., & Demerouti, E. (2007). The Job Demands-Resources model: State of the art. Journal of Managerial Psychology,22(3), 309–328.
Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin,107(2), 238–246.
Betoret, F. D. (2009). Self-efficacy, school resources, job stressors and burnout among Spanish primary and secondary school teachers: A structural equation approach. Educational Psychology,29(1), 45–68.
Borman, G. D., & Dowling, N. M. (2008). Teacher atttrition and retention: A meta-analytic and narrative review of the research. Review of Educational Research,78(3), 367–409.
Boyle, G. J., Borg, M. G., Falzon, J. M., & Baglioni, A. J., Jr. (1995). A structural model of the dimensions of teacher stress. British Journal of Educational Psychology,65(Pt 1), 49–67.
Brayfield, A., Mennino, S. F., & Rubin, B. A. (2005). Home-to-job and job-to home spillover: The impact of company policies and workplace culture. The Sociological Quarterly,46, 107–135.
Byrne, B. M. (1991). Burnout: Investigating the impact of background variables for elementary, intermediate, secondary, and university educators. Teaching and Teacher Education,7(2), 197–209.
Byrne, B. M. (2012). Structural equation modeling with Mplus: Basic concepts, applications, and programming. New York: Routledge.
Cammann, C., Fichman, M., Jenkins, D., & Klesh, J. (1979). The Michigan Organizational Assessment Questionnaire. University of Michigan. Unpublished manuscript.
Chesters, J. (2012). Gender attitudes and housework: Trends over time in Australia. Journal of Comparative Family Studies,43(4), 511–526.
Cheung, G., & Rensvold, R. (2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling,9(2), 233–255.
Cohen, J. (1988). Statistical power analysis for the behavioural sciences. Hillsdale, NJ: Erlbaum.
Cohen, T. F., DeVault, C., & Strong, B. (2008). The marriage and family experience: Intimate relationships in a changing society. Belmont, CA: Thompson Wadsworth.
Cook, A., & Minnotte, K. L. (2008). Occupational and industry sex segregation and the work–family interface. Sex Roles,59(11), 800–813.
Costa, G., Pickup, L., & Di Martino, V. (1988). Commuting — a further stress factor for working people: Evidence from the European Community. International Archives of Occupational and Environmental Health,60(5), 377–385.
Craig, C. (2017). International teacher attrition: Multiperspective views. Teacher and Teaching,23(8), 859–862.
Dawes, J. (2008). Do data characteristics change according to the number of scale points used? An experiment using 5-point, 7-point and 10-point scales. International Journal of Market Research,50(1), 61–104.
Demerouti, E., Bakker, A. B., Nachreiner, F., & Schaufeli, W. B. (2001). The job demands-resources model of burnout. Journal of Applied Psychology,86(3), 499–512.
Erdamar, G., & Demirel, H. (2014). Investigation of work-family, family-work conflict of the teachers. Social and Behavioral Sciences,116, 4919–4924.
Evans, A. M., Jamie, S. C., & Morgan, W. (2013). Work-life balance for men: Counseling implications. Jounal of Counseling and Development,91(4), 436–441.
Farber, B. A. (1991). Crisis in education: Stress and burnout in the American teacher. San. Francisco: Jossey Bass.
Flood, M., & Barbato, C. (2005). Off to work: Commuting in Australia. Canberra: The Australia Institute.
Flook, L., Goldberg, S. B., Pinger, L., Bonus, K., & Davidson, R. J. (2013). Mindfulness for teachers: A pilot study to assess effects on stress, burnout and teaching efficacy. Mind Brain Education,7(3), 186–195.
Freeman, C., O'Malley, K., & Eveleigh, F. (2014). Australian teachers and the learning environment: An analysis of teacher response to TALIS 2013: Final Report.
Fritz, M. S., Taylor, A. B., & MacKinnon, D. P. (2012). Explanation of two anomalous results in statistical mediation analysis. Multivariate Behavioral Research,47(1), 61–87.
Frone, M. R., Yardley, J. K., & Markel, K. S. (1997). Developing and testing an integrative model of the work–family interface. Journal of Vocational Behavior,50(2), 145–167.
Goddard, R., & Goddard, M. (2006). Beginning teacher burnout in Queensland schools: Associations with serious intentions to leave. The Australian Educational Researcher,33(2), 61–75.
Goddard, R., & O'Brien, P. (2003). Beginning teachers' perceptions of their work, well-being, and intention to leave. Asia-Pacific Journal of Teacher Education and Development,6(2), 99–118.
Goddard, R., O'Brien, P., & Goddard, M. (2006). Work environment predictors of beginning teacher burnout. British Educational Research Journal,32(6), 857–874.
Gold, E., Smith, A., Hopper, I., Herne, D., Tansey, G., & Hulland, C. (2010). Mindfulness-based stress reduction (MBSR) for primary school teachers. Journal of Child and Family Studies,19(2), 184–189.
Goode, W. J. (1960). A theory of role strain. American Sociological Review,25, 483–496.
Greenglass, E. R., & Burke, R. J. (2003). Teacher stress. In M. F. Dollard, A. H. Winefield, & H. R. Winefield (Eds.), Occupationalstress in the service professions (pp. 213–236). New York: Taylor and Francis.
Greenhaus, J. H., & Beutell, N. J. (1985). Sources of conflict between work and family roles. The Academy of Management Review,10(1), 76–88.
Greenhaus, J. H., Collins, K. M., Singh, R., & Parasuraman, S. (1997). Work and family influences on depature from public accounting. Journal of Vocational Behaviour,50, 249–270.
Greenhaus, J. H., Parasuraman, S., & Collins, K. M. (2001). Career involvement and family involvement as moderators of relationships between work-family conflict and withdrawal from a profession. Journal of Occupational Health Psychology,6(2), 91–100.
Hakanen, J. J., Bakker, A. B., & Schaufeli, W. B. (2006). Burnout and work engagement among teachers. Journal of School Psychology,43(6), 495–513.
Hang-yue, F. S., & Loi, R. (2005). Work role stressors and turnover intentions: A study of professional clergy in Hong Kong. International Journal of Human Resource Management,16(11), 2133–2146.
Hayes, A. F. (2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium. Communication Monographs,76(4), 408–420.
Ingersoll, R. M. (2012). Beginning teacher induction: What the data tell us. Phi Delta Kappan,93(8), 47–51.
Kahn, R. L., Wolfe, D. M., Quinn, R. P., Snoek, J. D., & Rosenthal, R. A. (1964). Organisational stress: Studies in role conflict and ambiguity. Oxford: Wiley.
Kanter, R. M. (1977). Work and family in the United States: A critical review and agenda for research and policy. New York: Russell Sage Foundation.
Kiger, G., Mannon, S. E., Minnotte, K. L., & Stevens, D. P. (2007). Examining the 'neglected side of the work-family interface': Antecendents of positive and negative family to work spillover. Jounal of Family Issues,28(2), 242–262.
Klassen, R. M. (2010). Teacher stress: The mediating role of collective efficacy beliefs. The Journal of Educational Research,103(5), 342–350.
Klassen, R. M., & Chiu, M. M. (2011). The occupational commitment and intention to quit of practicing and pre-service teachers: Influence of self-efficacy, job stress, and teaching context. Contemporary Educational Psychology,36(2), 114–129.
Kokkinos, C. M. (2007). Job stressors, personality and burnout in primary school teachers. British Journal of Educational Psychology, 77(1), 229–243.
Korunka, C., Kubicek, B., Schaufeli, W. B., & Hoonakker, P. (2009). Work engagement and burnout: Testing the robustness of the job demands-resources model. The Journal of Positive Psychology,4(3), 243–255.
Kyriacou, C. (2001). Teacher stress: Directions for future research. Educational Review,53(1), 27–35.
Lau, P. S., Yuen, M. T., & Chan, R. M. (2005). Do demographic characteristics make a difference among Hong Kong secondary school teachers? Social Indicators Research,71, 491–516.
Lee, R. T., & Ashforth, B. E. (1996). A meta-analytic examination of the correlates of the three dimensions of job burnout. Journal of Applied Psychology,81(2), 123–133.
Llorens, S., Bakker, A. B., Schaufeli, W., & Salanova, M. (2006). Testing the robustness of the job demands-resources model. International Journal of Stress Management,13(2), 378–391.
Lorente, P. L., Salanova, S. M., Martinez, M. I., & Schaufeli, W. (2008). Extension of the Job Demands-Resources model in the prediction of burnout and engagement among teachers over time. Psicothema,20(3), 354–360.
Maslach, C., Jackson , S. E., & Schwab, R. L. (1996). Maslach Burnout Inventory-Educators (MBI-ES). In Maslach, C., Jackson, S. E. & Leiter, M. P. (Eds.), MBI Manual (3rd ed.). Palo Alto, CA: Consulting Psychologists Press.
McCormick, J., & Barnett, K. (2011). Teachers' attributions for stress and their relationships with burnout. The International Journal of Educational Management,25(3), 278–293.
Muthén, L. K., & Muthén, B. O. (1998–2017). Mplus user’s guide (Version Eighth). Los Angeles, CA: Muthén and Muthén.
Netemeyer, R. G., Boles, J. S., & McMurrian, R. (1996). Development and validation of Work - Family Conflict and Family - Work Conflict Scales. Journal of Applied Psychology,81(4), 400–410.
OECD. (2014). TALIS 2013 results: An international perspective on teaching and learning. TALIS: OECD Publishing, Paris.
OECD. (2018). The Future of Education and Skills 2030. Retrieved August 1, 2019, from https://www.oecd.org/education/2030/E2030%20Position%20Paper%20(05.04.2018).pdf.
Panatik, S. A. B., Badri, S. K. Z., Rajab, A., Rahman, H. A., & Shah, I. M. (2011). The impact of work-family conflict on psychological wellbeing among school teachers in Malaysia. Social and Behavioral Sciences,29, 1500–1507.
Richardson, P. W., Watt, H. M. G., & Devos, C. (2013). Types of professional and emotional coping among beginning teachers. In M. Newberry, A. Gallant & P. Riley (Eds.), Emotion and school; Advances in Research on Teaching (Vol. 18, pp. 229–254). Bingley: Emerald Group Publishing Ltd.
Roeser, R. W., Skinner, E., Beers, J., & Jennings, P. A. (2012). Mindfulness training and teachers' professional development: An emerging area of research and practice. Child Development Perspectives,6(2), 167–173.
Rothmann, S., & Joubert, J. H. M. (2007). Job demands, job resources, burnout and work engagement of management staff at a platinum mine in the North West Province. South African Journal of Business Management.,38(3), 49–61.
Sanders, W., & Rivers, J. C. (1996). Cumulative and residual effects of teachers on future student academic achievement. Knoxville: University of Tennessee: Value-Added Research and Assessment Center.
Schaufeli, W. B., & Bakker, A. B. (2004). Job demands, job resources, and their relationship with burnout and engagement: A multi-sample study. Journal of Organizational Behavior, 25(3), 293–315.
Schaufeli, W. B., Bakker, A. B., & Van Rhenen, W. (2009). How changes in job demands and resources predict burnout, work engagement, and sickness absenteeism. Journal of Organizational Behavior,30(7), 893–917.
Shockley, K. M., Shen, W., Denunzio, M. M., Arvan, M. L., & Knudsen, E. A. (2017). Disentangling the relationship between gender and work-family conflict: An integration of theoretical perspectives using meta-analytic methods. The Journal of Applied Psychology,102(12), 1601–1635.
Skaalvik, E. M., & Skaalvik, S. (2011). Teacher job satisfaction and motivation to leave the teaching profession: Relations with school context, feeling of belonging, and emotional exhaustion. Teaching and Teacher Education: An International Journal of Research and Studies,27(6), 1029–1038.
Skaalvik, E. M., & Skaalvik, S. (2017). Still motivated to teach? A study of school context variables, stress and job satisfaction among teachers in senior high school. Social Psychology of Education,20(1), 15–37.
Spector, P. E., & Jex, S. M. (1998). Development of four self-report measures of job stressors and strain: Interpersonal conflict at work scale, organizational constraints scale, quantitative workload inventory, and physical symptoms inventory. Journal of Occupational Health Psychology,3(4), 356–367.
Spilt, J. L., Koomen, H. M. Y., & Jak, S. (2012). Are boys better off with male and girls with female teachers? A multilevel investigation of measurement invariance and gender match in teacher-student relationship quality. Journal of School Psychology,50(3), 363–378.
Steiger, J. H., & Lind, J. C. (1980). Statistically-based tests for the number of common factors. Paper presented at the Psychometric Society, Iowa City, IA.
Taris, T. W., Le Blanc, P. M., Schaufeli, W. B., & Schreurs, P. J. G. (2005). Are there causal relationships between the dimensions of the Maslach Burnout Inventory? A review and two longitudinal tests. Work and Stress,19(3), 238–255.
Tejedor, E. M., & Rigotti, T. (2008). Burnout differences between primary and secondary education teachers. International Journal of Psychology,43(3–4), 430–430.
Thanacoody, P. R., Bartram, T., & Casimir, G. (2009). The effects of burnout and supervisory social support on the relationship between work-family conflict and intention to leave. Journal of Health Organization and Management,23(1), 53–69.
Todd, C., Cooksey, R., Davies, H., McRobbie, C., & Brophy, S. (2019). Mixed-methods evaluation comparing the impact of two different mindfulness approaches on stress, anxiety and depression in school teachers. British Medical Journal Open,9(7), 1–13.
Tucker, L. R., & Lewis, C. (1973). A reliability coefficient for maximum likelihood factor analysis. Psychometrika,38, 1–10.
Vercambre, M., Brosselin, P., Gilbert, F., Nerrière, E., & Kovess-Masféty, V. (2009). Individual and contextual covariates of burnout: A cross-sectional nationwide study of French teachers. BMC Public Health,9(1), 1–12.
Voydanoff, P. (2005). Work demands and work-to-family and family-to-work conflict: Direct and indirect relationships. Journal of Family Issues,26(6), 707–726.
Weldon, P. (2018). Early career teacher attrition in Australia: Evidence, definition, classification and measurement. Australian Journal of Education,62(1), 61–78.
Wilkins, R., Laß, I., Butterworth, P., & Vera-Toscano, E. (2019). The household, income and labour dynamics in Australia survey: Selected findings from waves 1 to 17. Retrieved from Melbourne Institute: Applied Economic and Social Research.
Wright, T. A., & Cropanzano, R. (1998). Emotional exhaustion as a predictor of job performance and voluntary turnover. Journal of Applied Psychology,83(3), 486–493.
Zohar, D. (1997). Predicting burnout with a hassle-based measure of role demands. Journal of Organizational Behavior,18(2), 101–115.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Measurement model for primary versus secondary and male versus female teachers
Results of the initial CFAs revealed the data did not fit the model sufficiently well for any teacher group: primary—χ2(314, N = 580) = 1244.831, p < 0.001, CFI = 0.912, TLI = 0.902, RMSEA = 0.073 (CI 0.069–0.077) and SRMR = 0.062; secondary—χ2(314, N = 675) = 1472.285, p < 0.001, CFI = 0.907, TLI = 0.896, RMSEA = 0.076 (CI 0.072–0.080) and SRMR = 0.067; male—χ2(314, N = 254) = 741.414, p < 0.001, CFI = 0.900, TLI = 0.888, RMSEA = 0.073 (CI 0.066–0.080) and SRMR = 0.067; female—χ2(314, N = 999) = 1823.928, p < 0.001, CFI = 0.906, TLI = 0.894, RMSEA = 0.069 (CI 0.066–0.072) and SRMR = 0.064. Examination of the modification indices (MIs) revealed the measurement error covariance between EE8 and EE4 indicated large MIs in all samples: primary MI = 243.863, secondary MI = 297.232; male MI = 92.226, female MI = 419.159. Inspection of the items EE8 (“Working with people directly puts too much stress on me”) and EE4 (“Working with people all day is really a strain for me”) revealed item content overlap/parallel wording of the items relating to the negative impact that working with people can have on a person which would explain their additional relationship. Therefore, the error covariance for these two items was freed for estimation and the models were reanalysed. This resulted in acceptable fit of the data to the model, in all samples: primary, χ2(313, N = 580) = 941.271, p < 0.001, CFI = 0.941, TLI = 0.933, RMSEA = 0.060 (CI 0.056–0.064) and SRMR = 0.057; secondary, χ2(313, N = 675) = 1097.944, p < 0.001, CFI = 0.937, TLI = 0.929, RMSEA = 0.062 (CI 0.058–0.066) and SRMR = 0.063; male, χ2(313, N = 254) = 634.508, p < 0.001, CFI = 0.925, TLI = 0.916, RMSEA = 0.064 (CI 0.056–0.071) and SRMR = 0.064; female, χ2(313, N = 999) = 1298.573, p < 0.001, CFI = 0.938, TLI = 0.931, RMSEA = 0.056 (CI 0.053–0.059) and SRMR = 0.060. The measurement models of the measures used in the current study for primary and secondary teachers are shown in Fig. 3, and those for male and female teachers are shown in Fig. 4. Except for item sm5 on the student misbehaviour factor, all items showed moderate to high factor loadings (standardised), across all subsamples.
About this article
Cite this article
Rajendran, N., Watt, H.M.G. & Richardson, P.W. Teacher burnout and turnover intent. Aust. Educ. Res. 47, 477–500 (2020). https://doi.org/10.1007/s13384-019-00371-x
- Work–family conflict
- Emotional exhaustion
- Turnover intent
- JD-R model