Reflections on English as a Foreign Language Teacher Burnout Risk Factors: The Interplay of Multiple Variables

It is widely acknowledged that teaching is a really demanding job. It is reported that up to 40% of teachers quit the teaching profession within their first five years (Borman & Dowling, 2008). For the past four decades, the concept of ‘burnout’ has turned out to be a matter of overriding concern in the United States and across the globe as a universal phenomenon (Aloe, Amo, & Shanahan, 2014) reflecting an individual’s perceptions of unmet needs and expectations turning into a triple syndrome of ‘exhaustion, depersonalization, and diminished personal accomplishment’ accompanied by growing disillusionment and negative overtones of debilitation, fatigue, indifference, depression, and low self-esteem (Gold & Roth, 2005; Maslach, Jackson & Leiter, 1996; Steinhardt, Smith Jaggers, Faulk & Gloria, 2011). It is however argued that emotional exhaustion alone is complex enough to exclusively represent burnout experience (Malach-Pines, 2005). The most commonly used burnout measure thus far has been the Maslach Burnout Inventory (MBI) assessing all the threefold components with an added version developed for those working in educational settings known as the ‘Educators Survey (ES)’. According to Maslach (2003), burnout is physical, mental, and emotional exhaustion following chronic job attrition, and this is most reported in professions relative to human services especially in ‘teaching’ with the highest burnout level as a demanding job (Jennett, Harris & Mesibov, 2003; Pietarinen, Pyhältö, Soini, & Salmela-Aro, 2013; Richardson & Watt, 2006). 

There are commonly known symptoms for burnout such as chronic fatigue, weariness, cynicism, negative thought patterns, low self-image, feelings of incapability, and low self-efficacy (Eggen & Kauchak, 2008; Troman & Woods, 2001). However, the major risk factors of burnout as a multifactorial progressive syndrome are little known. While recent studies within the teaching profession sought to explore possible sources of burnout, they failed to come up with consistent results and only made tentative references to a multitude of predictors ranging from individual factors (age, experience, level of education, etc.), and working conditions (job strain, work hours and setting, student misbehavior, lack of shared decision-making, workload, pedagogical barriers, role overload, work pressure, classroom environment, disruptive behavior, etc.) to personality traits, emotional intelligence (EI) and self-regulatory factors like self-efficacy (Abkari & Eghtesadi, 2020; Chang, 2009; Doménech & Gómez,  2010; Dorman, 2003; Lee, Seo, Hladkyj, Lovell  & Schwartzmann, 2013; Liaw, 2009; Lim, Bogossian  & Ahern, 2010; Santavirta, Solovieva & Theorell, 2007; Skaalvik  & Skaalvik, 2010; Toker, 2011). 

Studies producing apparently convincing evidence of burnout risk factors are not uncommon either. Brown (2012) argues that burnout arises out of one’s failure to cope with protracted, work-related stress, and the inability to live up to one’s expectations and job demands. Stevens (2007) found statistically significant differences in sex, teaching experience, personality, and school type with regard to burnout. Skaalvik and Skaalvik (2007) found a strong relationship between self-efficacy and teacher burnout. Within the Iranian EFL context, Amirian and Behshad (2016) reported significant positive relationships among teacher self-efficacy, EI, and years of experience. This argument was reiterated by Motallebzadeh, Ashraf, and Tabatabaee Yazdi (2014) who found strong negative correlations between the self-efficacy of English teachers and their reports of burnout. Additionally, significant correlations were reported between English language teachers’ sense of efficacy, burnout, teaching style, and EI (Akbari & Tavassoli, 2011). Brouwers and Tomic (2000) showed how burned-out teachers left their jobs as a result of low self-efficacy in their classrooms. In the same way, Pas, Bradshaw, and Hershfeldt, (2012) argued that teachers with higher efficacy levels experienced lower levels of burnout and that teachers’ burnout grows at a faster rate than their feeling of efficacy.

Studying the relationship between EI and burnout among 307 Chinese teachers, Ju, Lan, Li, Feng, and You (2015) found EI and workplace social support can protect teachers from teacher burnout. Taylor, McLean, Bryce, Abry, and Granger (2019) also found that life stress predictors were related to increased emotional exhaustion, which was related to decreased career optimism at the end of the first year. Self-efficacy was also the subject of a study by Malinen and Savolainen (2016) who explored how perceived school climate affects teachers’ job satisfaction and burnout. They found that school climate had a positive effect, partly mediated by self-efficacy on job satisfaction. However, job satisfaction and burnout were not explained by collective efficacy in student discipline which indicates the need for further studies on this topic.

Exploring teacher burnout within the EFL context is a significant issue because research findings showed that teacher burnout may have adverse effects on student learning and motivation.  For example, Shen, McCaughtry, Martin, Garn, Kulik, and Fahlman (2015) showed that teachers’ status of burnout is a key environmental factor correlated with students’ quality of motivation. On the other hand, many factors that contribute to teacher burnout derive from student-related issues. In a recent study within the Iranian EFL context, Akbari and Eghtesadi (2020) interviewed 15 teachers with high burnout levels. Teachers reported the following factors as the major reasons for their burnout: “students’ low proficiency, lack of support from administrators, student misbehavior, students’ lack of interest in learning English, time limitation, and class oversize” (p. 31). Also, Buric, Sliskovic, and Penezic (2019) employed structural equation modeling to conduct a study on 941 school teachers at two points in time with a time lag of approximately 6 months. Their findings highlighted the adverse effect of burnout in predicting teachers’ subsequent emotions and psychopathological symptoms.

Brittle (2020) acknowledges that research on burnout has primarily concentrated on demographic, environmental, and situational factors, with few studies investigating individual characteristics. To fill this gap, the present study attempts to reflect on the interplay of multiple individual factors that may predict English teacher burnout in a foreign language context.

The multifaceted nature of teacher burnout makes it a persistently strong case for the researchers to work on. In English as a Foreign Language (EFL) settings where there is a paucity of research with so many intervening variables at play exploring the issue is more like opening Pandora’s Box that further complicates what is already complex. For instance, due to the high unemployment rate in Iran, neither the education system nor even the EFL teachers themselves are concerned about job attrition as a direct result of burnout while it has long been taken as a primary systemic concern elsewhere. Instead, teacher burnout in Iran seems to have psychological overtones in that it may otherwise affect teachers’ motivation and the quality of their professional skills. This present study, therefore, represents an investigation into the problem form a pragmatic perspective reconsidering EFL teacher burnout together with a host of variables (EI, personality traits, teaching experience, self-efficacy, school type, gender, academic degree or education level, and age) relative to ‘teaching’ as one of the lowest-paying professions in Iran. The key research questions are as follows:

1. What is the relationship among teachers’ burnout level, EI, personality traits, self-efficacy, teaching experience, and age?  
2. How well can self-efficacy, EI, personality traits, teaching experience, and age predict teacher burnout independently and altogether?

Methodology

This study was conducted on a sample of Iranian EFL junior and senior high school teachers. It was launched in early 2017 using convenience sampling as a plausible approach in case of exploring correlational considerations (McMillan & Schumacher, 2006). Prior to the study, informed consent was obtained for participation in burnout research so that all prospective contributors could be aware of the aim, the processes, the potential risks and benefits of their involvement, and their general right to anonymity and withdrawal. Case report data were planned to be confidential and participants were assured to receive a copy of the results at the end of the study. The final sample consisted of 124 EFL teachers who voluntarily took part in the study. Table 1 shows the concise demographic data of participants. 

Table 1. Demographic Data

This study utilized a battery of questionnaires and scales as testing instruments: The Maslach Burnout Inventory, The Educators Survey (MBI-ES), Bandura's Teacher Self-Efficacy Scale (TSES), Wang and Law Emotional Intelligence Scale (WLEIS), and the Short Big Five Inventory (BFI-S) of Personality.

The MBI-ES is a version of the original MBI which is matched with educational settings such as schools, colleges, and universities. It includes 22 items with a three-subscale division: emotional exhaustion (EE) (9 items), depersonalization (DEP) (5 items), and personal accomplishment (PA) (8 items). In fact, the PA items were reverse coded for consistent scoring. Using confirmatory factors analysis, Lee and Ashforth (1993) validated the MBI’s constructs as holding three subscales. Further, significant internal and test-retest reliability coefficients were reported beyond the .001 level for the subscales (.90 for EE, .79 for DEP, and .71 for PA) and (.82 for EE, .60 for DEP, and .80 for PA) (Maslach et al., 1996). Not all items of MBI-ES are negatively worded. The last section, PA, is positively worded, whose lower mean scores would mean higher degrees of burnout. Thus, the MBI-ES developers (Maslach & Jackson, 1981a, 1981b) adopted a reverse scoring process for this section.

The WLEIS is a 16-item scale with four subscales: self-emotion appraisal, others' emotion appraisal, the use of emotion, and the regulation of emotion. Each subscale includes four items. It is a 7-point scale anchored with the notations (1 strongly disagree, 2 disagree, 3 moderately disagree, 4 neither agree nor disagree, 5 moderately agree, 6 agree, 7 strongly agree). For validation, Aslan and Erkus (2008) reported that the four subscale factors were loaded between 0.83-0.85, 0.74-0.89, 0.76-0.82, and 0.66-0.83, respectively, with Cronbach's alpha reliability coefficients of these factors ranging between 0.83 and 0.90. The results of confirmatory factor analysis also proved that these four factors were coherent with the data, with high internal consistency reliability of .83 (Foo, Elfenbein, Tan, & Aik, 2004). Additionally, Libbrecht, Beuckelaer, Lievens, and Rockstuhl (2014) were interested to know whether WLEIS was as invariant across Singapore and Belgium as it was in its country of origin (i.e. China) and the results of their measurement invariance (MI) showed identical ratings, factor loadings, and interconnections for both countries.

The short Big Five Inventory of the Personality (BFI-S) is a short 15-item version of personality dimension whose constructs are firmly rooted in the psychology of individual differences. It reduces personality traits to five universal domains of individual differences including neuroticism (N), extraversion (E), openness to experience (O), agreeableness (A), and conscientiousness (C) (John & Srivastava, 1999; McCrae & Costa, 1997). Lang, John, Lüdtke, Schupp, and Wagner (2011) assessed MI and age robustness of BFI–S of personality dimensions in three different conditions, using exploratory structural equation modeling, and found strong robustness of self-report measures of personality dimensions for young and middle-aged people. Last but not least amongst instruments is Bandura’s TSES, a questionnaire formulated for teachers for offering insights into problematic areas of school activities such as decision-making, resources, instruction, discipline, parent and community involvement, and school climate. TSES is a 9-point, 30-item interrogative self-efficacy scale with seven subscales anchored at ‘nothing, very little, some influence, quite a bit, and a great deal’.

All the instruments were with an attached demographic data request section comprising age, gender, teaching experience (experience), level of teaching (school type), and academic degree (education level). Correlations were run among teachers’ burnout level, EI, self-efficacy and personality traits, experience, and age. Sequential multiple regression analysis was also run to assess how much of the relationship between the response variable (teachers’ burnout) and the set of explanatory variables (teachers’ EI, self-efficacy, personality traits, age, and teaching experience) can be accounted for by explanatory variables and the nature and size of this relationship.

Results and Discussion

At first, a preliminary analysis was run to examine the assumptions of correlation and multiple regression. The results ensured that no violation of the assumptions including normality, linearity, multicollinearity, and homoscedasticity was noticed. In order to answer the first research question, correlations analysis was run (Table 2). As seen in Table 2, it can be said that at the .05 level only personality factors and teacher’s self-efficacy are negatively correlated although the strength of the association is approximately small (r < .3). It means that teachers’ personality traits and self-efficacy are not mutually exclusive and can adversely affect each other. Not surprisingly, though, at the .01 level, there is a strong positive relationship between age and experience. However, the only remaining strong positive association (r >.5) at this level is between teachers’ self-efficacy and EI which is not counter-intuitive as it suggests that with an increase in EI, there is a rise in self-efficacy and vice versa.

Table 2. Correlation among Burnout and Other Variables

In fact, it is an important finding. At the .01 level, EI and self-efficacy plausibly have moderate negative correlations with burnout. Other significant correlations include the moderate positive relationship between EI and age, self-efficacy and experience, and self-efficacy and age.

As for the second question: “How well can Self-efficacy and EI, personality traits, teaching experience, and age predict teacher burnout independently and altogether?” sequential multiple regression was run to assess the unique and collective effects of five explanatory variables on teacher burnout with the regression assumptions having been met already. The burnout model summary table below determines how well the sequential regression models fit the data. The approximated 0.36 value of R (Table 3) indicates a modest level of burnout prediction by the independent variables. The five models presented in Table 3 indicate how much of the variance in the response variable can uniquely and incrementally be accounted for by the predictors. R² Change values suggest that Model 1 with just ‘experience’ accounts for a negligible amount of variance (R²= 0.001%).  R² for the second model, the predictor of age, also amounts to an insignificant value of 0.006 %, which adds just 0.005 % of explanation for the variation in scores on the burnout test. In fact, the model with five explanatory variables can explain about 13 % of the variance in scores on the burnout test. This is a small amount.

What we were interested in examining with this sequential regression was if other predictors besides ‘personality’ had an independent ability to account for the variance. The output in Table 4 showed that they did not. The F Change column of the sequential regression model summary gives the result of an ANOVA comparing the current model with the previous model to decide whether the two models are statistically different. Model 3, for example, is statistically different from Model 1 (F = 14.22, p = 0.000), but it is only responsible for 1% change in burnout score. All other models are not statistically different from one another, as seen by the insignificant p values. There is not a very good fit for predictors.

Table 3. Burnout Model Summary for Sequential Regression

The ANOVA test for sequential regression (Table 4) shows that models for EI, self-efficacy, and personality fit the data well, although this is not the case for models of experience and age.

Table 4. ANOVA for the Sequential Regression Model

Unstandardized coefficients in Table 5 indicate how much the dependent variable varies with the independent variable when all other independent variables are held constant. The data show that EI is by far the most important factor in the total model and that with an increase in EI, there is a moderate decrease in burnout. Likewise, compared to the statistical significance of the independent variables, none of the coefficients are significantly different from zero except for EI which is highly correlated with burnout although this is hardly an unexpected finding.

In sum, through multiple regressions to predict burnout from experience, age, EI, self-efficacy, and personality, it can be concluded that although the predictive power of the model is not surprisingly equal to zero F (5, 118) = 3.502, p < .005, none of the independent variables except for EI uniquely significantly added to the burnout prediction and that the overall model was just minimally statistical (R² = .129).  Although the R2 from models 2 to 3 is suddenly going up, the change from models 3 to 5 shows that there is not much predictive power added to the model by the addition of another variable in these steps.

Table 5.Coefficients for the Sequential Regression

The findings of the study suggested that the average Iranian EFL teacher exhibited a moderate level of burnout, a relatively high EI, collectively positive orientation to such personality traits as emotional stability, extraversion, agreeableness, openness to new experience and conscientiousness, and a medium-range of self-efficacy. Of the five independent variables used in the regression models and correlations, the EFL teacher’s EI has proved to be the only significant predictor of burnout. This finding is well supported by previous studies (Akbari & Tavassoli, 2011; Cano- Garcia, Padilla-Munoz & Carrasco-Ortiz, 2005; Colomeischi, 2015; Mérida-López & Extremera, 2017; Pishghadam & Sahebjam, 2012).

The subtle implication for the education system is that whereas IQ is predominantly genetic-based, EQ can be learned, developed, and refined (Slaski & Cartwright, 2003) and accordingly pre- and in-service EFL programs can compromise high levels of burnout and job turnovers through fostering high emotional intelligence in (student) teachers in their (early and) mid-careers. The fact that the EFL learner’s EI could be developed was supported by Abdolrezapour and Tavakoli (2012), who found that the participants who received training in EI showed greater achievement in reading comprehension and they scored higher on EI tests. Also, there was a positive relationship between EI and age which means that as teachers grow older, they seem to come to be more emotionally intelligent. In line with this finding, evidence for a positive correlation between emotional intelligence and age is presented in the manuals for the Mayer Salovey Caruso Emotional Intelligence Test (MSCEIT) (Mayer, Salovey & Caruso, 2002) and EQ-i:S (Bar-On, 1997); however, effect sizes for age appear to be very small (Atkins & Stough, 2005). Moreover, the result of the study is consistent with that of other studies (Brouwers & Tomic, 2000; Motallebzadeh et al., 2014; Pas et al., 2012; Skaalvik & Skaalvik, 2007) wherein there was a negative correlation between teacher burnout and self-efficacy. It means that poor investment in EFL teacher perception and self-efficacy dynamics runs the risk of high burnout rates for EFL practitioners and educators.

However, unlike many studies that see younger or older teachers more likely to suffer from burnout (Motallebzadeh et al., 2014; Schwab & Iwanicki, 1982), this study, in line with that of Dennis (2008), documents no significant relationship between teacher burnout and age/experience, denoting that teachers at any age levels are prone to burnout. Additionally, whereas Motallebzadeh et al. (2014) reported a negative relationship between self-efficacy and age, the present study found a strong positive correlation between them inferring that as EFL teachers grow older, they feel more professionally efficient. This is a good news indicative of the importance of raising self-efficacy at the advent of teaching experience and in pre-service training. Amirian and Behshad (2016), also, found a positive direct relationship between teacher self-efficacy and EI which were in turn positively correlated with years of experience. Whereas in several studies personality traits are taken as significant predictors of burnout levels (Cano-Garcia, et al., 2005; Colomeischi, 2015; Pishghadam & Sahebjam, 2012), this study like Dennis (2008) found no significant relationship between the two variables. It may suggest that none of the personality types should be potentially more receptive to burnout experience.

The triple interconnection among age, self-efficacy, and EI, on the one hand, and lack of relationship among burnout and age as well as the negative interrelationship between burnout, self-efficacy, and EI, on the other hand, allude to a counter-argument here. Surprisingly enough, the authors found peculiar cases of clinically depressed experienced teachers who claimed to have high self-efficacy and EI. It seems that as EFL teachers grow more experienced, they come to adapt themselves better to their working conditions and they get conditioned to suffer less emotional exhaustion though they are aware that academically they are way behind their young educated colleagues newly joining the profession. In fact, based on the authors’ informal field observations, this is not good news at least for some EFL practitioners in Iran as they might be affected by a special kind of ‘learned helplessness’. They learn over time that their professional job is a Hobson’s choice from which there is no escape and that things are not going to change. They well know that as soon as they leave their unrewarding job, they are most likely to remain unemployed forever with a bleak outlook ahead. Therefore, they gradually move from their state of low self-efficacy in their midcareer into that of learned helplessness so as to be able to put up with the rest of their careers. They do not manage to live up to their expectations and they use their so-called high self-perceived efficacy and EI as a compensation strategy to mask their extreme feelings of burnout.

By the same token, it seems to be counterintuitive that age and experience had no significant correlation with burnout. Apparently, growing older and more experienced can neutralize the interconnections between burnout and self-efficacy or EI but there must be something wrong with that. To the authors’ view, experience and age are used by teachers as a filler strategy for their low or missing self-efficacy and EI. Unofficial or evidence-based reports of the in-service institutions of the Ministry of Education show an increasing rate of program attrition amongst seasoned EFL practitioners after the first session, suggesting that despite their auto-suggestion, older teachers have by far the lowest self-efficacy.

Conclusions

As there was a paucity of research in EFL teacher burnout integrated with multiple variables (Kim, Jörg & Klassen, 2019), this study sought to reflect on teacher burnout as a unitary construct with its associated risk factors to arrive at a panoramic rather than a piecemeal view of the problem as few studies investigated the possible interactions of many variables using different testing instruments in EFL contexts without mixed and inconsistent results. The findings of the current research suggest that it should be premature to try to apply a neat and easy solution to a multidimensional problem by investigating the effect of one or two variables.

All in all, the findings of the present study in the EFL context showed that emotional intelligence is the strongest predictor of EFL teacher burnout. It was also found that emotional intelligence and self-efficacy have a negative moderate correlation with teacher burnout. Strong correlations were also found between emotional intelligence and self-efficacy, age and self-efficacy, and age and experience. No significant relationship was found between burnout and age, experience, or personality. Moderate correlations were found between emotional intelligence and age, efficacy and emotional intelligence, personality and experience, personality and efficacy, and experience and age.

This study offers considerable implications for the education system in Iran and also for the international audience. Some EFL practitioners with chronic burnout symptoms in the study tried to conceal their real burnout levels by answering the questions falsely owing to their indeterminacy about the misuse of the case reports by the official monitoring institutions. However, the authors tried to correct the defect by increasing the number of participants and removing the possibly misleading results. The nonsignificant relation between burnout and age or experience repudiates the results of several studies outside the EFL context arguing that the risk of job attrition becomes considerably higher in the first five years of service. With this in mind and considering the unique impact of EI as a proven risk factor of teacher burnout in several studies, the technically-oriented pre- and in-service teacher education programs for teachers at all age levels should attach more importance to the acquisition of EI skills by practitioners. This includes programs in which EFL teachers can share ideas and coordinate their efforts. Building a pleasant and motivating school atmosphere for teachers to show their voice, agency, and expertise cooperatively and to exchange feedbacks by drawing on their experience and professional skills can also be beneficial for deferring burnout experience.

EFL teacher development programs should offer individual and group counseling services for teachers at risk of burnout and monitor teachers periodically to check if there is a change in their lifestyle and quality of teaching. The goal of these programs should be to enhance teachers' interpersonal communication skills and to develop their coping strategies (Akbari & Eghtesadi, 2017) to enable them to cope with on-the-job difficulties and avoid teacher burnout. In the case of acute teacher burnout, it is appropriate as part of the educational quality assurance protocols to make a referral to a career counselor. While considering the findings of the study, caution should be exercised since the study is not without sampling limitations because convenience sampling was employed. Future studies can replicate this study with larger sample sizes drawn randomly from the population of EFL teachers in Iran.