Burnout among senior teachers: Investigating the role of workload and interpersonal relationships at work

Teaching and Teacher Education 43 (2014) 99e109

Contents lists available at ScienceDirect

Teaching and Teacher Education journal homepage: www.elsevier.com/locate/tate

Burnout among senior teachers: Investigating the role of workload and interpersonal relationships at work Filip Van Droogenbroeck a, *, Bram Spruyt a, Christophe Vanroelen b a b

TOR, Sociology Department, Vrije Universiteit Brussel, Belgium Interface Demography, Sociology Department, Vrije Universiteit Brussel, Belgium

h i g h l i g h t s
Interpersonal relationships are differentially related to the burnout dimensions.
Relations with students are most strongly related to the three burnout dimensions.
Supervisors play a crucial role in giving teachers more autonomy.
More autonomy leads to less dissatisfaction with non-teaching-related workload.
Our data support the sequential process model of burnout.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 9 October 2013 Received in revised form 7 July 2014 Accepted 10 July 2014 Available online 30 July 2014

According to the ‘intensification’ thesis, external pressures from policymakers, supervisors, parents, and experts fuel an ever-expanding teaching role and associated workload. Against that background, we examined how four interpersonal relationships (students, colleagues, supervisors, and parents), teaching-related and non-teaching-related workload (e.g., paperwork), and autonomy are related to teacher burnout (emotional exhaustion, depersonalization, personal accomplishment). The participants were 1878 Flemish teachers 45 years of age or older. The data were analyzed by means of structural equation modelling. Interpersonal relationships were differentially related to burnout. Teaching-related and non-teaching-related workload were both related to emotional exhaustion. Autonomy was most strongly related to non-teaching-related workload. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Teacher burnout Teacher autonomy Workload Social support Intensification

1. Introduction Burnout is a persistent, negative, work-related state of psychological exhaustion that results from a misfit between personal intentions and motivations on the one hand and actual on-the-job experiences on the other (Schaufeli & Enzmann, 1998). It includes a wide range of psychological (e.g., chronic fatigue, low self-esteem, lack of self-confidence, depression) and physiological (e.g., headaches, muscle pains, hypertension) symptoms. Burnout impedes the attainment of professional goals (Maslach, 2003), depletes

* Corresponding author. Vrije Universiteit Brussel, Sociology Department, Pleinlaan 2, 1050 Brussel, Belgium. Tel.: þ32 474481384; fax: þ32 26148140. E-mail addresses: [email protected] (F. Van Droogenbroeck), [email protected] (B. Spruyt), [email protected] (C. Vanroelen). http://dx.doi.org/10.1016/j.tate.2014.07.005 0742-051X/© 2014 Elsevier Ltd. All rights reserved.

coping resources (Park, Jacob, Wagner, & Baiden, 2013), and consequently becomes self-perpetuating and hard to rehabilitate. The prevalence of burnout has been investigated in a wide range of occupational categories (Alarcon, 2011; Schaufeli & Enzmann, 1998). In particular, contact professions (e.g., nurses, physicians, social workers, and teachers) turn out to be quite susceptible to burnout (Maslach, 2003). Within this occupational category teachers have repeatedly been shown to report the highest levels of burnout (Heus & Diekstra, 1999; Schaufeli & Enzmann, 1998). Not surprisingly, an extensive body of literature studies burnout among teachers (Burke & Greenglass, 1995; Byrne, 1991; Dorman, 2003; Grayson, Alvarez, & Grayson, 2008; Hakanen, Bakker, & Schaufeli, 2005; Kinman, Wray, & Strange, 2011; Kokkinos, 2007; Skaalvik & Skaalvik, 2009, 2010, 2011; Van Der Doef & Maes, 2002). This paper engages with this literature by studying the impact of (1) the quality of the interpersonal relationships teachers maintain with

100

F. Van Droogenbroeck et al. / Teaching and Teacher Education 43 (2014) 99e109

pupils, colleagues, parents and principals, (2) teaching and nonteaching related workload, and (3) perceived autonomy on burnout among senior teachers (aged 45e65) in Flanders (the Dutch-speaking part of Belgium, N: 1878). The focus on senior teachers is warranted, because the consequences of suffering from burnout for older employees in terms of their (re)integration in the labour market prove to be dramatic. Teachers suffering from burnout have a higher chance of leaving the profession and retiring early (Leung & Lee, 2006; Martin, Sass, & Schmitt, 2012; Rudow, 1999; Skaalvik & Skaalvik, 2011; Van Droogenbroeck & Spruyt, 2014). This is especially problematic, because in several countries, including Belgium, the teacher work force is ageing (OECD, 2005, 2012). According to the OECD (2012) this tendency will put increasing pressure on the teacher labour market and may lead to increasing teacher shortages. Since policymakers attempt to motivate employees e and more specifically teachers e to work longer, it becomes crucial to understand the determinants of burnout among senior teachers. The remainder of this article is organized as follows. First, we situate our research question against the background of changes the teaching profession went through during the last two decades. This section justifies our focus on the impact of the quality of different interpersonal relationships, types of workload, and autonomy. Next, we review the existing literature on burnout among teachers and locate our own study against that background. This section is followed by a description of our data and the methods we used. Subsequently, we present the results of our study. In the concluding section we elaborate on the implications of our findings.

1.1. Intensification of the teaching profession: interpersonal relationships and workload Our investigation of burnout is inspired by the multitude of changes the teaching profession has undergone during the past decades, often referred to as ‘intensification’ and ‘deprofessionalization’ (Apple, 1988; Ballet, Kelchtermans, & Loughran, 2006; Woods, 1999). The intensification thesis holds that teachers are increasingly subjected to external pressures from policymakers, supervisors, parents, and experts. This intensification results in an ever-expanding teaching role, a significant increase in nonteaching-related (largely administrative) workload, and less time for social contact with colleagues and in private life. The increasing demands and expectations towards education in general and teachers in particular by policymakers and society at large, results from the combination of on the one hand the way education (has to) function(s) and the broadening of education's objectives on the other hand. First of all, already in the 1980s Apple (1988), pointed to the increasing adoption of a more economic and management-oriented perspective on education among policymakers. Educational policy increasingly follows a market logic in which education is thought of as a product produced by schools. Thus, efficiency and effectiveness become the central evaluative parameters together with maximal freedom of choice for parents and pupils as consumers of education (Ball, 2003; Ballet & Kelchtermans, 2009). In that view, quality becomes a synonym for the effective and efficient achievement of standards that are objectively measurable. This implies that highlevel tasks become routinized, the sphere of work narrows (e.g., pre-specified competencies), and high-stakes accountability demands and standardization, such as prescribed and pre-specified curriculums, textbooks, methods, and standardized tests, increase (Valli & Buese, 2007; Woods, 1999). These externally imposed demands reduce the autonomy and creativity in the classroom and appear to lead towards “deskilling” and “deprofessionalization”.

A second e and related e issue concerns the scope of education's tasks within larger society which continued to broaden. Today, education is expected to contribute to the solution of numerous social problems (e.g., drug abuse prevention, civic education, health education, etc.). This so called ‘educationalization of society’ implies that schools are expected to take over tasks that were once carried out by parents (Smeyers & Depaepe, 2008). Some of these are experienced by teachers as improper to their assumed job description (Kelchtermans, 1999). Moreover, the broadening of education's scope also increased the number of other professionals (e.g., equal opportunities teachers, student counsellors, special needs teachers, education inspectors) that work with pupils and teachers. Teachers thus find themselves in a contradictory situation in which they face greater responsibility for their work but have less control over the way they wish to work (Ball, 2003). Although the literature often emphasizes the negative impact of intensification, different authors stress that it may also have positive effects. Being confronted with intensification, for example, some teachers look for further training or collaborative lesson planning as coping strategies which may result in re-skilling and re-professionalization (e.g., action research, teacher-as-researcher movements) (Ballet & Kelchtermans, 2009; Woods, 1999). More generally, intensification does not directly flow from educational policy into the classroom. Intensification is a nuanced, mediated and not completely deterministic process (Ballet et al., 2006). For example, teachers filter imposed policy changes through their personal ideologies and perspectives and have a certain freedom in how they implement them in the classroom. Ballet et al. (2006) distinguished several working conditions in schools that mediate the impact of intensification. First, allowing teachers to co-decide how work is organized can induce job satisfaction and fosters feelings of collegiality and a sense of community. Participation in decision-making is crucial to mitigate the impact of intensification by enabling teachers to identify problematic external pressures and tackling them collectively. At the same time, however, participating in school policy-decisions can also increase the feeling of being overburdened especially when a lot of time has to be spent on making decisions without direct benefit for the pupils. Secondly, supervisors play a central role in translating policy arrangements and demands from the local community or parents to the specific school context. Supervisors, for example, can act as a buffer by postponing the implementation of a new curriculum if they feel that this would cause too much strain on the teachers. This implies that supervisors also face intensification which makes their buffering role more difficult. They have to deal with more (administrative) tasks and have to focus on the more managerial aspects of their job. Thirdly, positive interpersonal relationships at school can also mediate the impact of intensification. Positive relations with colleagues are very important in building shared values and norms and create a collective goal orientation. In this regard, Hargreaves (1994) noticed significant changes in teacher culture. He points to a shift from individualism to collaboration, from hierarchies to teams, from supervision to mentoring, from in-service training to professional development and from authority towards parents to a contract with parents. Indeed, teachers have to interact with and are made accountable to an increasing numbers of actors. Not only do they have to maintain relationships with students, but also with colleagues, supervisors, parents, and experts. On the one hand, positive interpersonal relationships can function as a buffer against the impact of intensification. On the other hand, Hargreaves identified “contrived collegiality”, which entails enforced collaboration resulting from a drive towards managerialism and accountability, rather than the interests of the students. This creates stress and contributes to the alienation of individuals who are opposed to such developments.

F. Van Droogenbroeck et al. / Teaching and Teacher Education 43 (2014) 99e109

It is important to note that the issue we discuss above, does not concern policy changes or accountability themselves. It is, after all, reasonable that teachers are accountable for the work they perform with students as part of their responsibility as a professional. In addition, policy changes and innovation are necessary to keep up with research on teaching, student performance, and inequality. Research also shows that teachers are ready to accept changes, which have a positive impact on the learning process and wellbeing of pupils (Smylie, 1994). What matters here concerns the type of policy context teachers are working in and the way that accountability is defined (for example high-stakes accountability with a focus on summative assessments versus formative assessments which are more focused on feedback to modify the teaching and learning activities). The preceding arguments lead to two insights. First, the discussion about the ultimate or general consequences of the intensification of the teaching profession has not been settled yet. Secondly, all processes associated with intensification point towards an increased importance of (a) the quality of (an increased multitude of) interpersonal relationships, (b) different types of workload and (c) experienced autonomy. In this paper we do not investigate the intensification process itself, but focus on some of the main consequences of intensification on the prevalence of burnout among senior teachers. The antecedents of burnout have often been investigated by discriminating between job demands and job resources as latent variables comprising a wide array of job conditions (e.g., Alarcon, 2011; Elovainio et al., 2005; Schaufeli & Bakker, 2004). The refinements of the original intensification thesis, however, argue against that practice. Indeed, as the experience of intensification is mediated, it follows that it is important to distinguish different aspects of the working conditions and interpersonal relationships with relevant others for there is no reason to assume that their impact is identical. In their everyday work, teachers interact with students, parents, colleagues, and supervisors. It is the quality of these relationships that determines if the relationship creates stress e job demand e or instead is experienced as positive and serves as a buffer e job resource e against stress. In addition, the experience of intensification is mediated by the participation in school policy-decisions and how administrative duties are dealt with. Instead of grouping variables in latent factors, such as job demands or resources, we investigate several aspects of the working conditions separately. Regrouping variables into latent constructs that only reflect a distinction between demands and resources, may lead to a loss of information, certainly since previous research has indicated that burnout is differentially related to different working conditions (e.g., Lee & Ashforth, 1996; Skaalvik & Skaalvik, 2009). 1.2. Teacher burnout Burnout is a multidimensional construct that is often considered the end stage in a chain of reactions to chronic occupational stress. Maslach and Jackson (1986, p. 1) define burnout as “a syndrome of emotional exhaustion, depersonalization, and reduced personal accomplishment that can occur among individuals who do ‘people work’ of some kind”. Emotional exhaustion refers to feelings of fatigue that develop as one's emotional energies become drained. Cynical depersonalization points to the development of indifferent and negative attitudes towards others; it refers to the impersonal and dehumanized perception of (service) recipients. Finally, reduced personal accomplishment is the negative evaluation of personal accomplishments in working with other people and dissatisfaction with one's achievements, often accompanied by feelings of insufficiency and poor professional self-esteem. Initially,

101

burnout was restricted to those individuals working with recipients (e.g., student, clients, patients), but the concept was later broadened to include all occupations (Maslach, Schaufeli, & Leiter, 2001). Burnout is distinguished from other syndromes, such as depression and chronic fatigue syndrome (CFS), because burnout is considered to be mainly job-related, whereas depression and CFS are more “context-free” (Schaufeli & Enzmann, 1998). Unlike burnout, CFS is pervasive and includes mainly physical symptoms, next to psychological symptoms. By contrast, burnout symptoms are primarily psychological, although accompanying physical symptoms are not uncommon. Although burnout and depression share similar workrelated antecedents, research has concluded that they are conceptually and empirically distinct (Schaufeli & Enzmann, 1998; Toker & Biron, 2012). Studies have confirmed the multidimensional nature of burnout and indicated that the three dimensions cannot be added up to a single measure (Byrne, 1994; Langballe, 2006; Schaufeli, Daamen, & Van Mierlo, 1994). The multidimensional nature, however, does not imply that there are no relationships between the burnout dimensions. Several models have been proposed to explain the associations among the three dimensions. One of the best known models assumes that the three dimensions follow a sequential process starting with emotional exhaustion resulting from the emotional demands of dealing with recipients (Leiter & Maslach, 1988). In an inappropriate attempt to cope with high levels of exhaustion, workers then tend to withdraw psychologically from others, which leads to dehumanized and depersonalized contact. This dysfunctional coping strategy can further deteriorate the relationship with recipients. Finally, this process leads to the development of feelings of diminished personal accomplishment, as more failures are experienced and workers become more dissatisfied with their achievements. Empirical evidence about the causal relations between the burnout dimensions is scarce. One exception is a study from Taris, Le Blanc, Schaufeli, and Schreurs (2005), which reviewed the evidence of several models and tested these models in a new longitudinal study. Their study supported the Leiter and Maslach (1988) model but also found that in their teacher sample higher levels of depersonalization were associated with higher levels of exhaustion across time. Although in our cross-sectional study it is impossible to investigate the causal nature of the burnout dimensions, we expect to find the same sequential process as proposed by Leiter and Maslach (1988). 1.2.1. Interpersonal relationships Interpersonal relationships play a central role in the work of teachers. Not surprisingly, then, previous research has revealed that the support teachers receive from their supervisor and the positive relationships they maintain with colleagues, parents, and students are negatively related with emotional exhaustion and cynical depersonalization and are positively related with personal ~ oz, & Carrascoaccomplishment (e.g., Cano-García, Padilla-Mun Ortiz, 2005; Dorman, 2003; Friedman, 1995, 2003; Gavish & Friedman, 2010; Grayson et al., 2008; Greenglass, Burke, & Konarski, 1997; Leung & Lee, 2006; Russell, Altmaier, & Vanvelzen, 1987; Skaalvik & Skaalvik, 2009, 2011). Our data allow us to investigate four interpersonal relationships: students, colleagues, supervisors, and parents. Based on previous research we expect direct paths from all the interpersonal relationships to the three burnout dimensions (see Fig. 1). However, there is little research that effectively takes into account all these components separately and at the same time. Often only one or two burnout dimensions or only one or two interpersonal relationships are investigated (e.g., CanoGarcía et al., 2005; Kinman et al., 2011; Skaalvik & Skaalvik, 2011). In addition, the relations between the burnout dimensions

102

F. Van Droogenbroeck et al. / Teaching and Teacher Education 43 (2014) 99e109

Fig. 1. Theoretical structural model. The variables RELCOL, RELSTU, and RELPAR were grouped together to increase the clarity of the image. Note: SPC ¼ support with policy changes; RELSUP ¼ relationship with supervisor; RELCOL ¼ relationship with colleagues; RELSTU ¼ relationship with students; RELPAR ¼ relationship with parents; AUT ¼ autonomy; EE ¼ emotional exhaustion; DP ¼ depersonalization; PA ¼ personal accomplishment.

are often not investigated (e.g., Grayson et al., 2008; Kokkinos, 2007; Skaalvik & Skaalvik, 2009; Van Der Doef & Maes, 2002). In this article we try to add to the existing literature by differentiating between four interpersonal relationships and investigating how these are separately related to the three burnout dimensions. We also take into account the relations between the burnout dimensions. This is important because, first, it can be expected that not all interpersonal relationships at work have the same influence. Previous research revealed that teachers are quite likely to accept changes when they believe they will have a positive effect on pupils (Smylie, 1994). This ‘pupil-centrism’ suggests that the quality of the studenteteacher relationships will be of particular importance for teachers' well-being. Second, an inverse relationship can be expected between the amount of direct paths from the interpersonal relationships to the burnout dimensions and the strength of the relationships between the burnout dimensions. If the relationships between the burnout dimensions are strong, we expect that many of the postulated direct paths will be mediated by one of the dimensions of burnout. 1.2.2. Teachers' autonomy and teaching-related and non-teachingrelated workload Teaching is considered by many teachers to be a vocation (Roness, 2011). When teachers are asked why they want to teach they primarily refer to intrinsic and altruistic reasons, such as the desire to teach, to work with children, or to make a contribution to society (Brookhart & Freeman, 1992; OECD, 2005; Struyven, Jacobs, & Dochy, 2012). Since the desire to teach and to help

children is so central in the motivation of teachers, it should not surprise us that externally imposed demands for high-stakes summative assessments, record keeping, and accountability are experienced as a burden, as these tasks distract teachers from their perceived core job of teaching students. Often teachers feel that the time investment is disproportionate to the benefits for the student (Ballet et al., 2009). Several authors have indicated that these administrative burdens are considered an important stressor and a main cause of job dissatisfaction among teachers (Byrne, 1991; Dinham & Scott, 1998; Scott, Stone, & Dinham, 2001; Van Droogenbroeck & Spruyt, 2014; Varlaam, Nutall, & Walker, 1992). For this reason we argue that it is important to distinguish between teaching-related (e.g., class preparation, teaching) and nonteaching-related workload (e.g., administration, meetings). It is plausible and even likely that teachers, trained and motivated to teach, react differently to the workload that is directly related to teaching and the well-being of their students than to the workload many of them perceive as being generated by high-stakes accountability demands. Previous research has indicated that time pressure and workload are significantly related to burnout (Hakanen et al., 2005; Schaufeli & Bakker, 2004). When burnout dimensions are analyzed separately, time pressure and workload are positively correlated with emotional exhaustion but not with the other dimensions (Dorman, 2003; Kokkinos, 2007; Mazur & Lynch, 1989; Skaalvik & Skaalvik, 2009, 2010, 2011; Van Der Doef & Maes, 2002). We expect teaching-related and nonteaching-related workload to be directly related to emotional exhaustion. Following the intensification theory, we also

F. Van Droogenbroeck et al. / Teaching and Teacher Education 43 (2014) 99e109

hypothesize that non-teaching-related workload is negatively related to personal accomplishment. Ballet et al. (2006) argued that intensification has multiple sources and is experienced differently by teachers. Supervisors, for example, play a central role in how policy changes are implemented. Through the support that supervisors provide and the extent of teacher involvement in school policy-decisions they allow, supervisors can influence teachers' experienced job autonomy (Skaalvik & Skaalvik, 2009). Autonomy, in turn, is known to be negatively related to emotional exhaustion and depersonalization and to be positively related with personal accomplishment (Skaalvik & Skaalvik, 2009, 2010). Following the literature, we expect that supervisor support is positively related to feelings of autonomy. More autonomy can lead to less teaching-related and non-teaching-related workload, as teachers have more chances to regulate the way they perform these tasks. We also expect autonomy to be negatively related to emotional exhaustion and cynical depersonalization and to be positively related to personal accomplishment. 1.2.3. Gender, age, and burnout We also examined the impact of age and gender on burnout. Age differences have been found for the three dimensions, although the effects have been inconsistent. In previous studies younger teachers reported significantly higher emotional exhaustion than older teachers (Anderson & Iwanicki, 1984; Byrne, 1991; Russell et al., 1987). Noor and Zainuddin (2011), on the other hand, in their sample of female teachers found that older female teachers experienced more emotional exhaustion than younger teachers did. For the other dimensions the findings are less consistent. In some studies younger teachers reported higher depersonalization scores (Lau, Yuen, & Chan, 2005; Pierce & Molloy, 1990; Van Der Doef & Maes, 2002), while in other studies reverse or non-significant age differences for cynical depersonalization were found (Anderson & Iwanicki, 1984; Byrne, 1991; Leithwood, Menzies, Jantzi, & Leithwood, 1999). In terms of personal accomplishment, several studies (Byrne, 1991; Lau et al., 2005) show that younger teachers experience lower personal accomplishment than older teachers. Others reported reverse (Heus & Diekstra, 1999) or non-significant associations (Anderson & Iwanicki, 1984; Van Der Doef & Maes, 2002). The different age ranges of the studies considered and the differential exposure of their samples to the “healthy worker effect” may provide an explanation for the inconsistencies (Schaufeli & Enzmann, 1998, p. 74). The healthy worker effect refers to a selection effect in which those teachers who burnout early in their careers are likely to quit their jobs, leaving behind those who exhibit low levels of burnout. This healthy worker effect leads to an underestimation of the effects of burnout, because only working and healthier employees are investigated. With regard to senior teachers this becomes interesting because previous research indicates that teachers suffering from burnout have a higher chance of leaving the profession and retiring early (Leung & Lee, 2006; Martin et al., 2012; Rudow, 1999; Skaalvik & Skaalvik, 2011). We hypothesize that a healthy worker effect will be visible at the end of the career, since those teachers who feel burned out will leave the profession sooner. Thus, we expect that older teachers, compared with younger teachers, will report less emotional exhaustion, cynical depersonalization, and more personal accomplishment. Gender differences have been found in all three burnout dimensions. In most studies women report higher emotional exhaustion than men do (e.g., Byrne, 1991; Grayson et al., 2008; Lau et al., 2005); only a few studies reported a reverse association (Anderson & Iwanicki, 1984). Byrne (1991) and Lau et al. (2005) reported significantly lower perceptions of personal accomplishment for female teachers when compared with male teachers,

103

whereas Unterbrink et al. (2007) found the reverse. Men generally report higher cynical depersonalization than women do (Byrne, 1991; Lau et al., 2005; Russell et al., 1987; Stoeber & Rennert, 2008). Following the literature we expect female teachers would experience less cynical depersonalization and personal accomplishment and higher levels of emotional exhaustion than male teachers would. 2. Method 2.1. Data This study focuses on senior teachers 45e65 years of age. The threshold age of 45 aligns with previous studies on older employees (e.g., Elovainio et al., 2005; Schreurs, Van Emmerik, De Cuyper, Notelaers, & De Witte, 2010). The data are based on the “End of career survey” conducted in 2011 among a random sample of the population of active and non-active teachers between 45 and 65 years of age in Flanders (Belgium). We contacted 5200 working and retired Flemish teachers using their private addresses. Because of the sensitive information in the survey, a higher response rate could be achieved by contacting teachers directly instead of using the school to distribute the surveys. To optimize the response rate we followed the Dillman procedure (Dillman, Smyth, & Christian, 2009) and used a mixed methods design. The teachers were contacted in four waves using mail surveys and reminder cards. Respondents were invited to fill out the questionnaire on the Internet if they preferred so (44% of those who participated in the survey did so). With a response rate of 62% or 3124 filled-out questionnaires, the survey was very well received (in comparison, the European Social Survey 2010 had an average response rate of 61% using faceto-face interviews). Our sample only differed slightly from the population of teachers. The data were weighted for gender, age, and educational level (highest weight factor: 2.32). Of the 3124 respondents 1246 were retired and 1878 were working at the time of the survey. Because burnout is primarily work related, only the data from the 1878 working teachers were used in the empirical part of the study. 2.2. Measures 2.2.1. Teacher burnout Teacher burnout was assessed with the Dutch version of the Maslach Burnout Inventory e The Dutch Educators Survey (MBI-NLES; Horn & Schaufeli, 1998) adapted for teachers. Emotional exhaustion was measured with five items (e.g., “I feel emotionally drained from my work.” a ¼ .89), cynical depersonalization was measured with five items (e.g., “I doubt the significance of my job” a ¼ .82) and personal accomplishment with six items (e.g., “I can effectively solve the problems that rise in my work” a ¼ .81). Responses were given on an eight-point Likert scale ranging from 0 “Never” to 7 “Every day”. Higher scores indicate more emotional exhaustion, cynical depersonalization and personal accomplishment. 2.2.2. Interpersonal relationships Four dimensions of interpersonal relationships at work were analyzed. We measured the relationships with students (four items e.g., “Most students treat me with respect” a ¼ .70), colleagues (five items e.g., “When I have a hard time at school I seek support from colleagues” a ¼ .80), supervisor (five items e.g., “The principal supports me on an adequate way during conflicts with students and parents” a ¼ .90), and parents (five items e.g., “I have good contacts with parents” a ¼ .75). All items were rated on a five-point Likert scale ranging from 1 “Strongly disagree” to 5 “Strongly agree”. Higher scores indicate more positive relationships.

104

F. Van Droogenbroeck et al. / Teaching and Teacher Education 43 (2014) 99e109

2.2.3. Autonomy and support with policy changes We also analyzed the autonomy of teachers and the support teachers receive when policy changes occur. The perception of autonomy was assessed by three items (e.g., “I have enough say in school policy” a ¼ .78). The perceived support teachers receive when policy changes occur was also assessed by three items (e.g., “I receive adequate support for the implementation of policy changes” a ¼ .70). All items were rated on a five-point Likert scale ranging from 1 “Strongly disagree” to 5 “Strongly agree”. Higher scores indicate more autonomy and more perceived support with the implementation of policy changes.

2.2.4. Teaching-related and non-teaching-related workload As argued in the theoretical Section 1.2.2 we analyze teachingrelated and non-teaching-related workload separately instead of using a general workload indicator. We asked teachers how satisfied they were with (1) workload because of teaching-related activities such as class preparation, teaching itself and (2) workload because of non-teaching-related activities such as paperwork and meetings. Both items were rated on a five-point Likert scale ranging from 1 “Very satisfied” to 5 “Very dissatisfied”.

2.3. Data analysis Confirmatory factor analysis (CFA) and structural equation n & modelling (SEM) as implemented by MPLUS 6.2 (Muthe n, 1998e2012) were used to analyze the data. We used Muthe SEM instead of other multivariate approaches, because SEM follows a confirmatory approach, corrects for measurement error, and allows for a clear distinction between direct and indirect effects (Byrne, 2011). The goodness-of-fit of the models was evaluated by fit indices such as the c2 statistic, root mean square error of approximation (RMSEA), comparative fit index (CFI), TuckereLewis fit index (TLI), and standardized root mean square residual (SRMR). For the RMSEA and SRMR indices, values less than .05 indicate a good fit, and values as high as .08 represent an acceptable fit (Browne & Cudeck, 1993). As a rule of thumb for the TLI and CFI indices, values greater than .95 are indicative of good fit, whereas values greater than .90 are considered acceptable (Hu & Bentler, 1999). Our preliminary CFA analysis of the burnout dimensions indicates the presence of non-normality in the variables. This does not directly affect the estimates, but it could lead to biased standard errors and an inflated number of statistically significant parameters. As suggested by Byrne (2011), we used the robust maximum likelihood estimator MLM that estimates RMSEA, CFI, TLI, standard errors, and mean-adjusted c2 test statistics that are robust to nonnormality. To evaluate our research questions, two measurement models were created using CFA to test the multidimensionality of the theoretical constructs. The first measurement model specified the three Maslach burnout dimensions, whereas the second specified autonomy, support with policy changes, and the interpersonal relationships with students, colleagues, supervisors, and parents. In the next step we combined both measurement models and tested a theoretical model by means of SEM where gender, age, and teaching-related and non-teaching-related workload were added. To discriminate between significantly better fitting models we used the c2 difference test for the SatorraeBentler1 scaled c2.

1 The MLM method estimates the SatorraeBentler scaled c2 and a scaling correction factor instead of the regular c2. See http://www.statmodel.com/chidiff. shtml for more information on the adjusted c2 difference test when using the MLM estimator.

3. Results Table 1 shows the zero order correlations between the variables and the Cronbach's alphas. The correlations between the three burnout dimensions varied from moderate to strong. The strongest correlation was found between emotional exhaustion and depersonalization (.74).2 The correlations between the four interpersonal relationships were positive and varied from weak to moderate. The strongest correlations were found between relationships with the supervisor and the relationships with colleagues (.42) and the relationships with the parents and the relationships with the students (.48). Furthermore, autonomy was positively related to the four interpersonal relationships and negatively related to emotional exhaustion (.32) and depersonalization (.43). Autonomy was most strongly correlated with relationships with the supervisor (.71). Teaching-related and non-teaching-related workload had a moderate negative correlation (.39). Table 2 summarizes the results for the CFA and SEM analyses. We first tested a measurement model for burnout. Results of the CFA revealed an acceptable fit for the hypothesized threedimensional MBI structure. The incorporation of one crossloading and two error-covariances significantly improved the fit (model 2). The second measurement model specified six aspects of the school context, four interpersonal relationships (students, colleagues, supervisors, and parents), autonomy, and the support with policy changes. The baseline model had a good fit with the data. Fit indices suggested a significantly better fit with one additional error-covariance (see model 4). In the next step the two measurement models were combined, also revealing a good fit of data (model 5). Based on previous empirical findings, a theoretical model (Fig. 1) was specified. Since previous research predicts direct paths between all interpersonal relationships and the three burnout dimensions, we grouped the interpersonal relationships with students, colleagues, and parents in the figure for the sake of clarity. As argued in the theoretical section (see Section 1.2.1) we expect several of these direct paths to become non-significant because we take into account four interpersonal relationships at the same time and we also specify the relationships between the three burnout dimensions. This should cause a mediation of several of the direct paths from the interpersonal relationships to the burnout dimensions. For the SEM analysis we followed the method suggested by Byrne (2011). The best fitting model is determined first, then the most parsimonious model. Once the final model is established, the remaining non-significant parameters are deleted. The SEM analysis of the postulated model (Fig. 1) revealed a satisfactory fit to the data (model 6). The modification indices suggested that a better fit could be achieved (model 7) if one path was added (relationships with students / teaching-related workload). The c2 difference test for the SatorraeBentler scaled c2 (Dc2(1) ¼ 18.07, p < .001) and other fit indices indicated that the addition of this path leads to a statistically significant improvement of the model. In addition, the modification indices suggested adding one further path (relationships with colleagues / autonomy), which equally resulted in a statistically significant improvement of the model (model 8). In the final step non-significant paths were removed. The final model revealed an acceptable fit (model 9). Fig. 2 presents the standardized regression estimates. Interestingly, the relationship with parents was not significantly related to any indicator in the final model. When this latent factor was removed,

2 Because of these strong correlations we controlled for multicollinearity between all the study variables. The highest Variance Inflation Factor (VIF) was 2.4, with a threshold of 10, multicollinearity can be excluded.

F. Van Droogenbroeck et al. / Teaching and Teacher Education 43 (2014) 99e109

105

Table 1 Zero order correlations and Cronbach's alphas.

1 EE 2 DP 3 PA 4 RELSUP 5 RELCOL 6 RELPAR 7 RELSTU 8 AUT 9 SPC 10 Teaching-related workload 11 Non-teaching-related workload 12 Age 13 Gender

1

2

3

4

5

6

7

8

9

10

11

12

13

e

.74 e

.29 .38 e

.26 .31 .17 e

.24 .33 .18 .42 e

.26 .35 .24 .25 .25 e

.28 .36 .26 .23 .19 .48 e

.32 .43 .22 .71 .43 .30 .25 e

.31 .35 .20 .40 .30 .24 .18 .48 e

.40 .35 .19 .25 .17 .19 .23 .25 .24 e

.36 .31 .16 .33 .16 .21 .16 .39 .29 .39 e

.13 .05 .10 .01 .03 .02 .03 .00 .00 .13 .07 e

.06 .08 .11 .02 .11 .05 .03 .00 .05 .13 .06 .17 e

.89

.82

.81

.90

.80

.75

.78

.70

e

e

e

e

Cronbach's alpha

Note: all correlations are significant (p < .001); EE ¼ emotional exhaustion; DP ¼ depersonalization; PA ¼ personal accomplishment. RELSUP ¼ relationship with supervisor; RELCOL ¼ relationship with colleagues; RELPAR ¼ relationship with parents; RELSTU ¼ relationship with students; AUT ¼ autonomy; SPC ¼ support with policy changes.

the model fit further improved to a good fit (c2 (712) ¼ 1546.541, RMSEA ¼ .033, CFI ¼ .952, TLI ¼ .947, SRMR ¼ .041). As expected, the three burnout dimensions influenced one another. We found that higher levels of emotional exhaustion were associated with higher levels of cynical depersonalization. The relationship between emotional exhaustion and cynical depersonalization was very strong (b ¼ .71). Further, higher levels of cynical depersonalization were associated with lower levels of personal accomplishment. We expected that the strong relationships between the burnout dimensions would cause several direct effects of interpersonal relationships to become non-significant. To verify this we tested the final model with and without associations between the burnout dimensions. In the model without relations between the burnout dimensions there are eight direct effects from interpersonal relationships to the burnout dimensions (not displayed). In the model with relations between the burnout dimensions the amount of direct effects was reduced to five (see Fig. 2). The latter model had a significantly better fit (Dc2(1) ¼ 56.51, p < .001) than the model that did not specify relations between the burnout dimensions.

Also in accordance with our expectations, both teaching-related and non-teaching-related workload had a direct positive but modest effect on emotional exhaustion. In a similar vein, our hypothesis assuming that teaching-related and non-teaching-related workload mediated the relationship between autonomy and emotional exhaustion was supported. Autonomy was most strongly related to non-teaching-related workload. Autonomy itself was most strongly predicted by positive relationships with the supervisor and in lesser extent by positive relationships with colleagues and feelings of being supported when policy changes occur. For the interpersonal relationships, the central role of relationships with students in the experience of burnout is noteworthy. Relations with students were related to all three burnout dimensions. Positive relationships with students were negatively related to emotional exhaustion and cynical depersonalization but positively related to personal accomplishment. Relationships with students also had an indirect effect on emotional exhaustion through teaching-related workload. Higher levels of positive relationships with students were related to lower levels of teachingrelated workload. In contrast, relationships with parents were not

Table 2 Summary of goodness-of-fit for the confirmatory factor analyzes and structural models. Actions Measurement model 1. Burnout - Model 1 baseline - Model 2 adapted

1 Cross-loading þ 2 Error covariances

2. School context - Model 3 baseline - Model 4 adapted

1 Error covariance

Model 5 combined

Structural model - Model 6 baseline - Model 7 adapted - Model 8 adapted Model 9 final model

þRELSTU on teaching-related workload þRELCOL on AUT þDeletion non-significant paths

MLM c2 (df, scaling correction)

RMSEA

Upper limit RMSEA 90%

CFI

TLI

SRMR

483.98 (101, 1.292) 319.629 (98, 1.280)

.053

.057

.954

.945

.039

.041

.046

.973

.967

.031

.039

.042

.952

.945

.038

.037

.040

.958

.951

.037

.031

.033

.955

.950

.039

.034

.036

.940

.935

.045

.033

.035

.942

.936

.043

.033

.035

.942

.937

.042

.033

.035

.942

.937

.043

761.689 (260, 1.201) 702.166 (259, 1.200) 1508.918 (739, 1.151)

1965.025 (906, 1.124) 1934.125 (905, 1.124) 1920.608 (904, 1.124) 1934.937 (913, 1.123)

106

F. Van Droogenbroeck et al. / Teaching and Teacher Education 43 (2014) 99e109

Fig. 2. Final structural model of four interpersonal relations, teaching-related and non-teaching-related workload, autonomy, support with policy changes, age, and gender with teacher burnout. Only significant standardized regression weights are included in the figure. Note: SPC ¼ support with policy changes; RELSUP ¼ relationship with supervisor; RELCOL ¼ relationship with colleagues; RELSTU ¼ relationship with students; RELPAR ¼ relationship with parents; AUT ¼ autonomy; EE ¼ emotional exhaustion; DP ¼ depersonalization; PA ¼ personal accomplishment.

related to any dimension of burnout. Positive relationships with colleagues were negatively related to emotional exhaustion and cynical depersonalization. Finally, the effect of positive relationships with the supervisor on exhaustion was fully mediated by autonomy. For the demographic variables, female teachers reported less cynical depersonalization and less personal accomplishment compared with male teachers. Older teachers reported less emotional exhaustion and more personal accomplishment, although the latter relation was weak (.08). Several expectations were not confirmed. We did not find a significant effect for non-teaching-related workload, autonomy, and relationships with colleagues and supervisor on personal accomplishment. Moreover, contrary to expectations, age and relationships with the supervisor were not related to cynical depersonalization, nor was gender related to emotional exhaustion. 4. Discussion and conclusion In this paper we assessed the impact of four interpersonal relationships (students, colleagues, supervisors, and parents) and teaching-related and non-teaching-related workload on burnout among senior teachers. Burnout among senior teachers deserves special attention, because research has indicated that burnout is related to attrition and early retirement (Leung & Lee, 2006; Martin et al., 2012; Rudow, 1999; Skaalvik & Skaalvik, 2011). Together with the coming retirement of the baby boom generation (Ekerdt, 2009),

the early exit culture among teachers (Eurydice Network, 2012), looming teacher shortages in several countries, and legislators aiming to motivate employees to work longer (OECD, 2012), studying burnout among senior teachers is highly relevant for society at large. According to the intensification thesis the working conditions of teachers changed significantly during the last decades. Teachers experienced a noticeable increase in high-stakes accountability demands, administrative tasks and standardization which may cause stress and feelings of deprofessionalization. This intensification of the profession is the result of teachers being increasingly subjected to external pressures and demands from policymakers, supervisors, parents, and experts. Maslach and Jackson (1984, p. 134) stated that the root cause of burnout comes from being emotionally overextended and drained by one's contacts with other people. In the case of teachers, the question may be asked, “Contact with whom?” In accordance with previous research, our results showed that the interpersonal relationships that teachers maintain at work contribute significantly to the prediction of burnout (e.g., Cano-García et al., 2005; Dorman, 2003; Friedman, 2003; Gavish & Friedman, 2010; Grayson et al., 2008; Greenglass et al., 1997; Leung & Lee, 2006; Skaalvik & Skaalvik, 2009, 2011). In line with the recent refinements of the intensification thesis we found, however, that not all interpersonal relationships at work have the same impact on burnout. The relationship with students, for example, appears to play the most important role and relates directly to each of the burnout dimensions. This is not surprising, since research

F. Van Droogenbroeck et al. / Teaching and Teacher Education 43 (2014) 99e109

shows that teachers are mainly driven by intrinsic and altruistic motivations, such as the desire to teach and help children (Brookhart & Freeman, 1992; OECD, 2005; Struyven et al., 2012). It is often taken for granted, but teachers find themselves in the peculiar situation where most of their time is spent with nonadults. When this time spent with students is experienced in a positive and pleasant way this feeling may directly spill-over to their job satisfaction and increase their perception of personal accomplishment. In such a situation these teachers are accomplishing what they are mainly driven to do. In contrast, the relationship with the parents was not significantly related to any of the burnout dimensions. Teachers encounter parents only a couple of times every year. This might explain why this relationship has no impact on burnout. The strongest influence appears to come from the relationships that are encountered on a daily basis. Relationships with colleagues were directly related to emotional exhaustion and cynical depersonalization. Although teachers spend most of their time alone with children, the support that teachers receive from colleagues can function as a buffer against burnout. In this regard, Hargreaves (1994) points to the need to develop “collaborative cultures” where teachers can engage in collaborative inquiry to collect and analyze local data so that data based decisions can be made to improve teaching and learning (e.g., finding joint solutions to shared problems, exchanging practices, co-teaching). Collaborating and talking about problems at work may induce feelings of solidarity and empathy and prevent indifferent and negative attitudes towards others. Finally, also the relationship with the supervisor plays an important role. The effect of this relationship on emotional exhaustion was fully mediated by autonomy. Supervisors play a crucial role in giving teachers more autonomy. They have the power to involve teachers in school policy-decisions and enable teachers to take initiative. The importance of autonomy is illustrated by the direct effect on depersonalization and the indirect effect on emotional exhaustion through teaching-related and non-teachingrelated workload. Although the relationship with both types of workload were significant, it is interesting to find that the one with non-teaching related workload was strongest. This is particularly interesting, because the intensification thesis states that, due to the rise of accountability demands and standardization over the last decades, the autonomy of teachers is diminishing (Ballet et al., 2006). Our results show that school leaders play an important role in mediating the experience of intensification. By increasing the autonomy of teachers they can enable teachers to better cope with teaching-related and especially non-teaching-related workload, such as paperwork and accountability demands. If teachers are involved in school policy-decisions and maintain positive relations with their supervisor and colleagues it is probable that external pressures and demands can be better identified and mitigated. More autonomy may enable teachers to regulate these tasks in ways that are more correspondent to their own views. In the same line our results showed that teachers experience more autonomy when they receive detailed information about policy changes and experience support with policy change implementation. More information about policy changes helps teachers to have a better understanding of the context of the imposed changes. That finding also supports the notion of Ballet et al. (2006): in the experience of intensification, teachers react in complex ways and do not just become “technical professionals” who blindly follow policy dictates. When teachers experience autonomy they are enabled to cope actively and creatively with changes in the workplace. Our results indicate, however, that non-teaching-related workload and especially teaching-related workload are positively related to emotional exhaustion. This illustrates that the total

107

workload is as important as the kind of workload experienced. That is, teaching-related workload together with the relationships maintained with the students has the strongest effect on emotional exhaustion. This finding confirms the central role of the quality of relationships in the experience of burnout. It might indicate that the most draining and also buffering aspect is the workload coming from the teaching activity itself and the quality of the relationships experienced therein. Nevertheless, non-teaching-related workload also plays an important role. Excessive paperwork and high-stakes accountability demands are among the most important stressors and causes of teacher job dissatisfaction (Byrne, 1991; Scott et al., 2001; Van Droogenbroeck & Spruyt, 2014; Varlaam et al., 1992). When teachers feel that they are accountable at all times and constantly have to ‘prove’ their competence, than this undermines their professional self-confidence (Kelchtermans, 2005). Senior teachers may be especially vulnerable to this kind of pressure, since they have experienced the rise of high-stakes accountability demands firsthand. It would help teachers if school leaders and policymakers try to align the non-teaching demands with the professional standards teachers set for themselves, namely, the well-being and achievements of their pupils. Today the added value of administrative tasks for the well-being of their pupils is often unclear or perceived as disproportionate to the time investment required. Research by Smylie (1994) indicated that teachers are more willing to accept changes which have a positive impact on the learning process of pupils. This illustrates that for teachers the problem probably does not lie in dealing with accountability or policy changes themselves but in how accountability is defined and how it is put to use. It is reasonable to expect that teachers account for their work with students as part of their responsibility as a professional and to some extent administrative tasks are of course part and parcel of the contemporary teaching profession. In this context one could raise the question whether teachers in training are sufficiently prepared for the administrative component of the job. In several countries the job of teachers has traditionally been defined in terms of classroom teaching hours (OECD, 2005). Broader job profiles may be needed; these should include the complex and wide array of teachers' tasks and responsibilities. For senior teachers, burnout can be a factor that makes them leave the profession or retire early (Leung & Lee, 2006; Rudow,1999; Skaalvik & Skaalvik, 2011; Van Droogenbroeck & Spruyt, 2014). As expected, older teachers reported less emotional exhaustion and more personal accomplishment than younger teachers. Previous research indicated that a widespread early exit culture exists among teachers (Eurydice Network, 2012). In Belgium teachers make massive use of early exit schemes (Van Droogenbroeck & Spruyt, 2014). Our findings might be an indication for the healthy worker effect at the end of the career where those teachers who feel burned out leave the profession as early as possible, while motivated, happy, and healthy teachers are those who keep teaching. We found a differential relation between the burnout dimensions and interpersonal relationships, autonomy, and teaching-related and non-teaching-related workload. That finding supports the multidimensional conceptualization of burnout; that is, the three dimensions of burnout should not be added up to a single measure (Byrne, 1994; Langballe, 2006; Skaalvik & Skaalvik, 2009). Although burnout is a multidimensional construct, this does not mean that its dimensions are unrelated to each other. Often in teacher burnout studies, the three burnout dimensions are assumed to be unrelated (e.g., Grayson et al., 2008; Kokkinos, 2007; Skaalvik & Skaalvik, 2009; Van Der Doef & Maes, 2002). This may cause an overestimation of direct paths, because mediated effects are not taken into account. We confirmed our expectation that there indeed exists an inverse relation between the amount of direct paths from interpersonal relationships to the burnout

108

F. Van Droogenbroeck et al. / Teaching and Teacher Education 43 (2014) 99e109

dimensions and the strength of the associations between the burnout dimensions. Based on our literature study, we postulated 11 direct paths from interpersonal relationships to the burnout dimensions. Eventually, only five of these direct paths remained in the final model when the relations between burnout dimensions were specified. By regarding burnout as a sequential process, researchers may better understand the factors that contribute to its development and take into account the mediation of effects. Our study has several limitations. First, the cross-sectional design based on self-reported measures does not allow causal inferences to be made between the investigated variables. Second, we measured only four interpersonal relationships at work. Other interpersonal relationships, such as studentepeer relations and parentecommunity relations, can also have an influence on burnout (Grayson et al., 2008). Third, we used two one-item scales to assess teaching-related and non-teaching-related workload. This may have led to an underestimation of the influence of these independent variables. Future research could benefit by further exploring this distinction. A multidimensional analysis of what the work of teachers entails could give more insight in which aspects influence the burnout process. Acknowledgement FundingThe author(s) disclosed receipt of the following financial support for the research of this article: This work was partly supported by the Flemish Department of Education [grant number OBPWO 09.06]. References Alarcon, G. M. (2011). A meta-analysis of burnout with job demands, resources, and attitudes. Journal of Vocational Behavior, 79(2), 549e562. Anderson, M. B. G., & Iwanicki, E. F. (1984). Teacher motivation and its relationship to burnout. Educational Administration Quarterly, 20(2), 109e132. Apple, M. W. (1988). Teachers and texts: A political economy of class and gender relations in education. New York, NY: Routledge. Ball, S. J. (2003). The teacher's soul and the terrors of performativity. Journal of Education Policy, 18(2), 215. Ballet, K., & Kelchtermans, G. (2009). Struggling with workload: primary teachers' experience of intensification. Teaching and Teacher Education, 25(8), 1150e1157. Ballet, K., Kelchtermans, G., & Loughran, J. (2006). Beyond intensification towards a scholarship of practice: analysing changes in teachers' work lives. Teachers and Teaching: Theory and Practice, 12(2), 209e229. Brookhart, S. M., & Freeman, D. J. (1992). Characteristics of entering teacher candidates. Review of Educational Research, 62(1), 37e60. Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen, & J. S. Long (Eds.), Testing structural equation models (pp. 136e162). Newbury Park, CA: Sage. Burke, R. J., & Greenglass, E. (1995). A longitudinal study of psychological burnout in teachers. Human Relations, 48(2), 187e202. Byrne, B. M. (1991). Burnout: investigating the impact of background variables for elementary, intermediate, secondary, and university educators. Teaching and Teacher Education, 7(2), 197e209. Byrne, B. M. (1994). Burnout: testing for the validity, replication, and invariance of causal structure across elementary, intermediate, and secondary teachers. American Educational Research Journal, 31(3), 645e673. Byrne, B. M. (2011). Structural equation modeling with Mplus: Basic concepts, applications, and programming (1st ed.). New York, NY: Routledge Academic. ~ oz, E. M., & Carrasco-Ortiz, M. A. (2005). Personality Cano-García, F. J., Padilla-Mun and contextual variables in teacher burnout. Personality and Individual Differences, 38, 929e940. Dillman, D. A., Smyth, J. D., & Christian, L. M. (2009). Internet, mail and mixed-mode surveys: The tailored design method (3rd ed.). Hoboken: Wiley. Dinham, S., & Scott, C. (1998). A three domain model of teacher and school executive career satisfaction. Journal of Educational Administration, 36(4), 362e378. Dorman, J. P. (2003). Relationship between school and classroom environment and teacher burnout: a LISREL analysis. Social Psychology of Education, 6(2), 107e127. Ekerdt, D. J. (2009). Frontiers of research on work and retirement. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 65B(1), 69e80. €ki, M., Sinervo, T., Sutinen, R., & Laine, M. (2005). Job Elovainio, M., Forma, P., Kivima demands and job control as correlates of early retirement thoughts in Finnish social and health care employees. Work & Stress, 19(1), 84e92.

Eurydice Network. (2012). Key data on education in Europe 2012. Education, Audiovisual and Culture Executive Agency e European Commission. Friedman, I. A. (1995). Student behavior patterns contributing to teacher burnout. The Journal of Educational Research, 88(5), 281e289. Friedman, I. A. (2003). Self-efficacy and burnout in teaching: the importance of interpersonal-relations efficacy. Social Psychology of Education, 6(3), 191e215. Gavish, B., & Friedman, I. A. (2010). Novice teachers' experience of teaching: a dynamic aspect of burnout. Social Psychology of Education, 13(2), 141e167. Grayson, J. L., Alvarez, H. K., & Grayson, J. L. A. H. K. (2008). School climate factors relating to teacher burnout: a mediator model. Teaching and Teacher Education, 24(5), 1349e1363. Greenglass, E. R., Burke, R. J., & Konarski, R. (1997). The impact of social support on the development of burnout in teachers: examination of a model. Work & Stress, 11(3), 267e278. Hakanen, J. J., Bakker, A. B., & Schaufeli, W. B. (2005). Burnout and work engagement among teachers. Journal of School Psychology, 43(6), 495e513. Hargreaves, A. (1994). Intensification. Teachers' work e better or worse?. Changing teachers, changing times: Teachers' work and culture in the postmodern age (pp. 117e140) London: Cassell. Heus, P. D., & Diekstra, R. F. W. (1999). Do teachers burn out more easily? A comparison of teachers with other social professions on work stress and burnout symptoms. In R. Vandenberghe, A. M. Huberman, R. Vandenberghe, & A. M. Huberman (Eds.), Understanding and preventing teacher burnout (pp. 269e284). Cambridge: Cambridge University Press. Horn, J. E., & Schaufeli, W. B. (1998). Maslach burnout inventory: The Dutch educators survey (MBI-NL-ES) psychometric evaluations. Manual (unpublished manuscript). Utrecht University, Department of Social and Organizational Psychology. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1e55. Kelchtermans, G. (1999). Teaching career: between burnout and fading away? Reflections from a narrative and biographical perspective. In R. Vandenberghe, & A. M. Huberman (Eds.), Understanding and preventing teacher burnout (pp. 176e191). Cambridge: Cambridge University Press. Kelchtermans, G. (2005). Teachers' emotions in educational reforms: selfunderstanding, vulnerable commitment and micropolitical literacy. Teaching and Teacher Education, 21(8), 995e1006. Kinman, G., Wray, S., & Strange, C. (2011). Emotional labour, burnout and job satisfaction in UK teachers: the role of workplace social support. Educational Psychology, 31(7), 843e856. Kokkinos, C. M. (2007). Job stressors, personality and burnout in primary school teachers. British Journal of Educational Psychology, 77, 229e243. Langballe, E. M. (2006). The factorial validity of the Maslach burnout inventorygeneral survey in representative samples of eight different occupational groups. Journal of Career Assessment, 14(3), 370e384. Lau, P. S. Y., Yuen, M. T., & Chan, R. M. C. (2005). Do demographic characteristics make a difference to burnout among Hong Kong secondary school teachers? Social Indicators Research, 71(1e3), 491e516. 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), 123e133. Leiter, M. P., & Maslach, C. (1988). The impact of interpersonal environment on burnout and organizational commitment. Journal of Organizational Behavior, 9(4), 297e308. Leithwood, K., Menzies, T., Jantzi, D., & Leithwood, J. (1999). Teacher burnout: a critical challenge for leaders of restructuring schools. In R. Vandenberghe, & A. M. Huberman (Eds.), Understanding and preventing teacher burnout (pp. 85e114). Cambridge: Cambridge University Press. Leung, D. Y. P., & Lee, W. W. S. (2006). Predicting intention to quit among Chinese teachers: differential predictability of the components of burnout. Anxiety, Stress & Coping, 19(2), 129e141. Martin, N. K., Sass, D. A., & Schmitt, T. A. (2012). Teacher efficacy in student engagement, instructional management, student stressors, and burnout: a theoretical model using in-class variables to predict teachers' intent-to-leave. Teaching and Teacher Education, 28(4), 546e559. Maslach, C. (2003). Burnout: The cost of caring. Los Altos, CA: Malor Books. Maslach, C., & Jackson, S. E. (1984). Burnout in organizational settings. In S. Oskamp (Ed.), Applied social psychology annual: Applications in organizational settings. Beverly Hills, CA: Sage Publications. Maslach, C., & Jackson, S. E. (1986). Maslach burnout inventory (2nd ed.). Palo Alto, CA: Consulting Psychologists Press. Maslach, C., Schaufeli, W. B., & Leiter, M. P. (2001). Job burnout. Annual Review of Psychology, 52(1), 397e422. Mazur, P. J., & Lynch, M. D. (1989). Differential impact of administrative, organizational, and personality-factors on teacher burnout. Teaching and Teacher Education, 5(4), 337e353. n, L. K., & Muthe n, B. O. (1998). Mplus user's guide (7th ed.). Los Angeles, CA: Muthe n & Muthe n. Muthe Noor, N. M., & Zainuddin, M. (2011). Emotional labor and burnout among female teachers: workefamily conflict as mediator: emotional labor and burnout. Asian Journal of Social Psychology, 14(4), 283e293. OECD. (2005). Teachers matter: Attracting, developing and retaining effective teachers. Education and training policy. OECD Publishing. OECD. (2012). Education at a glance 2012: OECD indicators. OECD Publishing. Park, H. I., Jacob, A. C., Wagner, S. H., & Baiden, M. (2013). Job control and burnout: a meta-analytic test of the conservation of resources model. Applied Psychology. http://dx.doi.org/10.1111/apps.12008.

F. Van Droogenbroeck et al. / Teaching and Teacher Education 43 (2014) 99e109 Pierce, C. M. B., & Molloy, G. N. (1990). Psychological and biographical differences between secondary school teachers experiencing high and low levels of burnout. British Journal of Educational Psychology, 60(1), 37e51. Roness, D. (2011). Still motivated? The motivation for teaching during the second year in the profession. Teaching and Teacher Education, 27(3), 628e638. Rudow, B. (1999). Stress and burnout in the teaching profession: European studies, issues, and research perspectives. In R. Vandenberghe, & A. M. Huberman (Eds.), Understanding and preventing teacher burnout (pp. 115e138). Cambridge: Cambridge University Press. Russell, D. W., Altmaier, E., & Vanvelzen, D. (1987). Job-related stress, social support, and burnout among classroom teachers. Journal of Applied Psychology, 72(2), 269e274. 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), 293e315. Schaufeli, W. B., Daamen, J., & Van Mierlo, H. (1994). Burnout among Dutch teachers: an MBI-validity study. Educational and Psychological Measurement, 54(3), 803e812. Schaufeli, W. B., & Enzmann, D. (1998). The burnout companion to study and practice: A critical analysis. London: Taylor & Francis Group. Schreurs, B., Van Emmerik, H., De Cuyper, N., Notelaers, G., & De Witte, H. (2010). Job demandseresources and early retirement intention: differences between blue-and white-collar workers. Economic and Industrial Democracy, 32(1), 47e68. Scott, C., Stone, B., & Dinham, S. (2001). “I love teaching but… ” international patterns of teacher discontent. Education Policy Analysis Archives, 9, 28. Skaalvik, E. M., & Skaalvik, S. (2009). Does school context matter? Relations with teacher burnout and job satisfaction. Teaching and Teacher Education, 25(3), 518e524. Skaalvik, E. M., & Skaalvik, S. (2010). Teacher self-efficacy and teacher burnout: a study of relations. Teaching and Teacher Education, 26(4), 1059e1069. 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, 27(6), 1029e1038. Smeyers, P., & Depaepe, M. (Eds.). (2008). Educational research: The educationalization of social problems. Dordrecht: Springer Netherlands.

109

Smylie, M. A. (1994). Redesigning teachers' work: connections to the classroom (Chapter 3) Review of Research in Education, 20(1), 129e177. Stoeber, J., & Rennert, D. (2008). Perfectionism in school teachers: relations with stress appraisals, coping styles, and burnout. Anxiety, Stress & Coping, 21(1), 37e53. Struyven, K., Jacobs, K., & Dochy, F. (2012). Why do they want to teach? The multiple reasons of different groups of students for undertaking teacher education. European Journal of Psychology of Education, 1e16. 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 & Stress, 19(3), 238e255. Toker, S., & Biron, M. (2012). Job burnout and depression: unraveling their temporal relationship and considering the role of physical activity. Journal of Applied Psychology, 97(3), 699e710. Unterbrink, T., Hack, A., Pfeifer, R., Buhl-Grießhaber, V., Müller, U., Wesche, H., et al. (2007). Burnout and effortereward-imbalance in a sample of 949 German teachers. International Archives of Occupational and Environmental Health, 80(5), 433e441. Valli, L., & Buese, D. (2007). The changing roles of teachers in an era of high-stakes accountability. American Educational Research Journal, 44(3), 519e558. Van Der Doef, M., & Maes, S. (2002). Teacher-specific quality of work versus general quality of work assessment: a comparison of their validity regarding burnout, (psycho)somatic well-being and job satisfaction. Anxiety, Stress & Coping, 15(4), 327e344. Van Droogenbroeck, F., & Spruyt, B. (2014). To stop or not to stop: an empirical assessment of the determinants of early retirement among active and retired senior teachers. Research on Aging. http://dx.doi.org/10.1177/ 0164027513519449, 0164027513519449. Varlaam, A., Nutall, D., & Walker, A. (1992). What makes teachers tick?. Clare market papers no. 4. London: Centre for Education Research, London School of Economics. Woods, P. (1999). Intensification and stress in teaching. In R. Vandenberghe, & A. M. Huberman (Eds.), Understanding and preventing teacher burnout (pp. 115e138). Cambridge: Cambridge University Press.