Teacher burnout during COVID-19: associations with instructional self-efficacy but not emotion regulation

ABSTRACT

Teachers face a range of exhausting job demands which contribute to burnout. These demands may be particularly acute during the COVID-19 pandemic, with lockdowns forcing rapid shifts to remote teaching. Yet during times of stress and upheaval, personal resources such as teaching self-efficacy and emotion regulation may protect teachers against burnout. Drawing on the Job Demands-Resources Model, this study aimed to examine the roles of self-efficacy (classroom management, student engagement, and instructional strategies self-efficacy) and emotion regulation (cognitive reappraisal and expressive suppression) as predictors of teacher burnout during the first wave of the COVID-19 pandemic. Using an online survey, 210 Australian school teachers completed measures of teaching self-efficacy, emotion regulation, and burnout while teaching remotely. Burnout was measured using the Copenhagen Burnout Inventory, which categorises burnout into personal, student-related and work-related sources. As hypothesised, lower instructional strategies self-efficacy predicted higher work-related burnout during COVID-19. Teaching experience also predicted higher burnout. Counter to the hypotheses, however, emotion regulation was unrelated. The findings indicate that different types of teaching self-efficacy may be beneficial in protecting teachers from burnout from different sources.

KEYWORDS:

• Burnout
• exhaustion
• teacher
• COVID-19
• self-efficacy
• emotion regulation

Teaching is both rewarding and stressful (Bernard, 2016; Bottiani et al., 2019; Saeki et al., 2018; Skaalvik & Skaalvik, 2016). Numerous workplace stressors place teachers at high risk of emotional exhaustion: a core component of burnout (Kristensen et al., 2005; Maslach et al., 2001). Burnout in turn has negative consequences for teachers’ work performance and wellbeing (Ghanizadeh & Royaei, 2015), and makes it more likely that teachers will report intentions to quit (Goddard & Goddard, 2006).

The COVID-19 pandemic has caused unprecedented upheaval to teachers and schools. Many educational jurisdictions moved to online or home-based schooling, with teachers required to rapidly learn new pedagogies and modes of interaction (e.g. Flack et al., 2020). According to the Job Demands-Resources Model, excessive job demands—including those of the kind created by the COVID-19 pandemic—are likely to cause stress, exhaustion, and burnout (Bakker et al., 2005; Demerouti et al., 2001). Consistent with this possibility, preliminary research on the pandemic suggests increases in teachers’ self-reported stress that are attributable to the imposition of new remote teaching methods and the loss of social interactions (Flack et al., 2020; Fray et al., 2022; L. J. Sokal et al., 2020; L. Sokal et al., 2020). As stress is a precursor to emotional exhaustion, the core component of burnout, these findings are particularly important.

Although the additional upheaval and stress brought on by the COVID-19 pandemic may place teachers at heightened risk of burnout, the Job Demands-Resources Model predicts that teachers’ personal resources should buffer against these effects. While research is limited, various forms of teaching self-efficacy (Dicke et al., 2014; Schwarzer & Hallum, 2008; Wang et al., 2015) and emotion regulation (Chang, 2020; Ghanizadeh & Royaei, 2015; Yin et al., 2016) have each separately been found to protect against burnout in everyday classroom contexts. These personal resources may be particularly important when teachers are responding to significant stressors and upheaval (O’toole & Friesen, 2016). Two lines of emerging research also show that teachers’ self-efficacy for instructing remotely (Košir et al., 2020), and their emotional resources (MacIntyre et al., 2020) were beneficial in protecting teachers against stress during COVID-19. Drawing on these findings, the aim of the current study was to examine how different forms of teaching self-efficacy and emotion regulation predict teachers’ burnout during the COVID-19 pandemic.

Understanding teacher burnout: the job demands-resources model

Burnout is commonly defined as the psychological and physical symptoms of chronic exhaustion and fatigue, typically resulting from job stressors (Kristensen et al., 2005; Maslach et al., 2001). Burnout has been operationalised in terms of general or occupational exhaustion (Kristensen et al., 2005; Shirom, 2005), or more broadly as comprising of exhaustion, depersonalisation or cynicism, and lowered accomplishment (Maslach & Leiter, 2016; Maslach et al., 2001). Exhaustion is emphasised across both operationalisations, and those in people-oriented professions such as teaching are considered especially susceptible (Maslach & Leiter, 2016). Focusing specifically on exhaustion, Kristensen et al. (2005) considered three perceived burnout sources. Personal burnout refers to general exhaustion, client-related burnout is exhaustion from client interactions, while work-related burnout is exhaustion from work generally (Kristensen et al., 2005; note similarities between work-related burnout and work stress). For teachers, client-related burnout is referred to as student-related burnout (Milfont et al., 2008; Platsidou & Daniilidou, 2016).

According to the Job Demands-Resources model, there are two factors that predict work-related exhaustion and burnout: job demands and job resources (Demerouti et al., 2001; Llorens et al., 2006). Job demands are physical, social or organisational aspects of the job that require sustained effort. For teachers, such demands may include managing student behaviour, planning and delivering curriculum, attending school meetings, and other tasks that require time or expertise (Hakanen et al., 2006). Exhaustion and burnout occur when these demands are excessive (Bakker et al., 2005).

Job resources, in turn, buffer the impact of job demands, promote the achievement of work goals, and encourage personal growth (Demerouti et al., 2001). In this way, job resources can protect against exhaustion and burnout (Bakker et al., 2005). Situational job resources such as supervisor support, adequate funding, and a positive workplace environment (Hakanen et al., 2006) are valuable, yet largely outside individual teachers’ control. This makes it important for school systems and school leaders to also consider how best to support teachers in developing relevant personal resources, such as self-efficacy (Schwarzer & Hallum, 2008) emotion regulation (Yin et al., 2016), and other related forms of coping (McCarthy et al., 2009).

Research has supported the applicability of the Job Demands-Resources model to teaching (Hakanen et al., 2006; Roslan et al., 2015; Yin et al., 2016), with findings that teachers’ personal resources can protect against the adverse effects of job demands. For example, one line of research shows that teaching self-efficacy may protect teachers against burnout (Klassen & Chiu, 2010; Schwarzer & Hallum, 2008; Skaalvik & Skaalvik, 2007, 2010; Wang et al., 2015), and this is particularly true for classroom management self-efficacy (Aloe et al., 2014; Brouwers & Tomic, 2000; Dicke et al., 2014). A separate line of research shows that emotion regulation (Ghanizadeh & Royaei, 2015; Yin et al., 2016) protects teachers against exhaustion and burnout.

Self-efficacy as a personal resource for teachers

Self-efficacy, a domain-specific judgement of one’s own ability, is a potentially important personal resource for teachers: determining responses when faced with obstacles and challenges (Bandura, 1977, 1986). Those with higher self-efficacy are more likely to perceive the environment positively, leading to positive emotions and increasing the likelihood of task persistence, while those with lower self-efficacy are more likely to perceive the environment negatively: leading to negative emotions and increasing the likelihood of disengagement (Bandura, 2006).

Previous research in everyday teaching contexts has found that teachers with high general teaching self-efficacy report lower burnout and exhaustion (Boujut et al., 2017; Jeon et al., 2018; Schwarzer & Hallum, 2008; Skaalvik & Skaalvik, 2010). Some types of teaching self-efficacy may be more important as personal resources than others, however (Aloe et al., 2014; Dicke et al., 2014; Wang et al., 2015). Tschannen-Moran and Hoy (2001) identified three distinct elements of teachers’ self-efficacy: classroom management self-efficacy, which captures self-efficacy for managing students’ behaviours; student engagement self-efficacy, which captures self-efficacy for motivating students to engage in learning; and instructional strategies self-efficacy, which captures self-efficacy for skilfully delivering content. Drawing on this distinction, Wang et al. (2015) found that student engagement self-efficacy and classroom management self-efficacy, but not instructional self-efficacy, were each related negatively to emotional exhaustion. Thus, the interpersonal aspects of teaching may be more stressful or exhausting than instructional aspects.

Emotion regulation as a personal resource for teachers

When examining the personal resources that may protect against exhaustion and burnout, it is also important to also consider teachers’ emotion regulation. Teaching requires considerable emotional labour (McGrath & Van Bergen, 2019). Teachers are regularly required to respond to emotional classroom demands, to skilfully utilise emotional expression (Sutton et al., 2009), and to manage negative emotions as they arise. Emotions such as anger and frustration may be particularly prominent when encountering challenging student behaviours (McGrath & Van Bergen, 2019), but may also emerge in response to other demands too: for example, tight time pressures, unsupportive leadership, or worry about one’s own skillset. Adaptive emotion regulation techniques may buffer teachers from exhaustion and burnout by promoting positive emotions, facilitating rewarding relationships with students, and improving job performance (Yin et al., 2016).

Gross and John (2003, 2001) identify two key emotion regulation strategies: cognitive reappraisal and expressive suppression. Cognitive reappraisal is antecedent-focused and involves cognitively reconstructing the incoming emotional trigger to change the resulting emotions. If one is facing a large or threatening work challenge with a high risk of failure, for example, cognitive reappraisal may mean reconceptualising the challenge so that it is seen as an opportunity for growth. Expressive suppression, in turn, is response-focused and involves purposefully modulating an emotional response to inhibit its expression (Gross & John, 2003). If facing that same large or threatening work challenge, expressive suppression might instead mean trying not to think of the challenge until necessary so as to avoid feeling negative emotion.

Research on teachers’ use of emotion regulation to protect against exhaustion and burnout is equivocal. Chang’s (2020) survey of 561 school teachers in the United States found that cognitive reappraisal was negatively associated with burnout, whereas expressive suppression was positively associated. Similarly, Yin et al. (2016) found that emotional job demands were negatively related to teachers’ emotional exhaustion when mediated by cognitive reappraisal, and positively related to exhaustion when mediated by expressive suppression. Chang’s (2020) argued that suppression is more likely to lead to exhaustion as it does not remove the experience of negative emotions, yet still consumes cognitive resources. Contrary to these previous studies, however, Ghanizadeh and Royaei (2015) found that cognitive reappraisal and expressive suppression were both negatively associated with burnout among tertiary educators. One explanation for these equivocal findings is that school teachers might be required to regulate emotions more frequently than tertiary educators, in response to greater student conflict. They therefore experience greater demand on their emotional resources, and become exhausted more easily. Where job demands are greater, the specific emotion regulation strategy used may become more important.

The impact of COVID-19 as a job demand

Sudden changes in the work and personal lives of teachers can cause significant stress and emotional exhaustion (Flack et al., 2020; L. J. Sokal et al., 2020). Although past research highlights potential roles for self-efficacy and emotion regulation in protecting teachers against emotional exhaustion and burnout, there has been little research into the experiences of teachers when faced with significant stressors and upheaval (O’toole & Friesen, 2016). In the current study, the COVID-19 pandemic therefore provides a unique opportunity.

The COVID-19 pandemic had a major impact on schools, with many making rapid shifts to remote teaching (Flack et al., 2020). Stressors may include not just the adjustment to remote teaching itself, but also concurrent issues related to work-life balance, concerns about students, and technology issues (L. J. Sokal et al., 2020). Personal resources such as self-efficacy (Haverback, 2020; Košir et al., 2020) and emotion regulation (Restubog et al., 2020) are each important during periods of stress or rapid change, and may therefore be particularly important in protecting teachers against burnout during the pandemic.

Haverback (2020) argued that self-efficacy is crucial for protecting teachers against burnout during COVID-19. Importantly, however, the specific self-efficacy that is relevant may differ from classroom contexts. Teachers engaged in remote teaching must utilise new technologies and new pedagogies to deliver content (L. J. Sokal et al., 2020), and those who do not believe they can do so effectively may have been at additional risk of disengagement, negative emotions, and exhaustion. For this reason, we predict that self-efficacy for instructional strategies will be more important than self-efficacy for classroom management and student engagement during the shift to remote teaching. To date, relatively few studies have considered this possibility. In a survey of 1626 Canadian teachers, L. Sokal et al. (2020) found that all three forms of self-efficacy correlated negatively with emotional exhaustion during COVID-19. Promisingly, however, self-efficacy for instructional strategies had a higher correlation coefficient (.198) than did classroom management (.123) or student engagement (.135). We extend this work by entering all three forms of self-efficacy into the same statistical model: thus allowing us to examine each form of self-efficacy independently while controlling for other forms. Košir et al. (2020) similarly sampled 964 Slovenian school teachers and counsellors: finding that lower self-efficacy to deliver content remotely during the first COVID-19 wave was associated with higher stress. This may be due to teachers with low self-efficacy for remote teaching disengaging and experiencing negative emotions from the loss of self-efficacy, leading to burnout.

Initial studies on teachers’ responses to COVID-19 also provide some insights into the role that emotion regulation may play in protecting teachers against burnout. While no pandemic research has yet considered cognitive reappraisal and emotional suppression, L. J. Sokal et al. (2020) found that emotional self-care was associated with lower burnout in teachers. Similarly, an international survey of 600 teachers by MacIntyre et al. (2020) found that teachers’ use of positive reframing, a coping strategy similar to cognitive reappraisal, was related to lower stress during remote teaching. Together, these findings suggest that strategies which promote emotional health, particularly antecedent-based strategies similar to reappraisal, may be effective in reducing burnout during the pandemic.

The current study

The aim of the current study was to test the relationship between teachers’ personal resources (self-efficacy and emotion regulation) and burnout during the COVID-19 pandemic. To date, no studies to our knowledge have examined how different forms of self-efficacy and emotion regulation may uniquely and differentially predict burnout from different sources. Using Kristensen et al. (2005)’s Copenhagen Burnout Inventory, which has been used successfully with teachers previously (e.g. Milfont et al., 2008; Platsidou & Daniilidou, 2016), we operationalised burnout narrowly: as a sense of general or occupational exhaustion. We captured this exhaustion in three domains: personal, work, and students. In addition, no studies to our knowledge have yet examined the relative influence of personal resources related to both self-efficacy and emotion regulation during the COVID-19 pandemic. Through the lens of the Job Demands-Resources model, we predict that relevant personal resources will be particularly important when teachers are faced with additional stress and upheaval.

Interestingly, classroom-based research suggests that self-efficacy and emotion regulation are related not just to burnout but to each other (Chan, 2004; DiFabio & Palazzeschi, 2008; Mouton et al., 2013). Ghanizadeh and Royaei (2015) proposed that the ability to regulate emotions may aid in more effective decision making in the classroom, which in turn improves actual outcomes and leads to positive self-efficacy beliefs protecting against burnout. Schwarzer and Hallum (2008) suggested that high efficacy beliefs to handle the challenges of everyday teaching may increase motivation to use effective emotion regulation techniques. Finally, Tsouloupas et al. (2010) examined self-efficacy, emotion regulation and exhaustion together. In their model, cognitive reappraisal and self-efficacy for managing student misbehaviour (similar to classroom management) were each negatively associated with teacher exhaustion, whereas expressive suppression was positively associated. We extend this research by examining teachers’ burnout from multiple perceived sources, during the COVID-19 pandemic.

Our first hypothesis was that teachers’ self-efficacy for instructional strategies and self-reported use of cognitive reappraisal as an emotion regulation strategy would each negatively and uniquely predict teachers’ burnout during the COVID-19 pandemic. Previous research has found negative associations between teachers’ burnout and the interpersonal domains of teaching self-efficacy, both prior to COVID-19 (Klassen & Chiu, 2010; Wang et al., 2015) and in the COVID-19 era (L. J. Sokal et al., 2020), suggesting that feelings of competence can buffer teachers from exhaustion when facing everyday classroom challenges (i.e. classroom management and student engagement). Self-efficacy for instructional strategies was expected to be particularly important for burnout during shifts to remote learning, however.

With regards to emotion regulation, research also consistently identifies a negative relationship between cognitive reappraisal and exhaustion: the core component of burnout (Chang, 2020; Ghanizadeh & Royaei, 2015; Yin et al., 2016). Based on these findings from everyday classroom teaching, our second hypothesis was that cognitive appraisal would protect against burnout during the pandemic. Consistent with this possibility, we note that emotional self-care (L. J. Sokal et al., 2020) and the use of antecedent-based coping strategies (MacIntyre et al., 2020) have each been associated with lower stress in teachers during COVID-19. While self-efficacy should support teachers to confidently face new COVID-19 related teaching challenges, therefore, cognitive reappraisal should support teachers to frame these challenges in a positive way. To our knowledge, no research has yet considered this possibility.

Finally, our third hypothesis was that emotional suppression would predict burnout. While there are mixed findings for expressive suppression in everyday contexts (Chang, 2020; Ghanizadeh & Royaei, 2015; Yin et al., 2016), COVID-19 has placed large job demands on teachers. We suggest that expressive suppression may be insufficient for protecting against negative emotion with such demands.

Materials and methods

Participants

Two hundred and ten participants completed the online survey (183 female, 27 male). Participants had an average of 13.9 years (SD = 11.00) teaching experience, with a range of 0–58 years. The mean age of participants was 39.67 years (SD = 11.14), with a range of 20–69 years. Age and teaching experience were strongly correlated, r(208) = .82, p < .001.

Participants were recruited through social media and snowball sampling, using Facebook groups targeted towards Australian teachers. The advertisement specified that participants should be primary or high school teachers currently employed in an Australian school. All participants met the inclusion criteria, with 77.62% teaching in high schools, 17.62% in primary schools, and 4.76% in special education. The higher percentage of high school teachers may reflect recruitment methods, as a larger number of high school groups granted permission to advertise the study on their pages.

Measures

Demographics

Participants were asked to provide their gender (female, male, or other), age, years of teaching experience, level of Australian teacher accreditation (graduate, proficient, highly accomplished, or lead teacher), year groups they currently teach (primary, high school, or other), and type of school that they were employed at (public, independent or catholic).

Teachers’ Sense of Efficacy Scale – Long Form (TSES)

The TSES (Tschannen-Moran & Hoy, 2001) is a 24-item self-report measure used to measure teachers’ self-efficacy for classroom management (8 items), student engagement (8 items), and instructional strategies (8 items). Responses are measured on a 9-point scale from nothing (1) to a great deal (9). Higher scores indicate higher self-efficacy. Each subscale is scored separately, with participants’ average score across items calculated for each subscale.

In past research, the TSES has demonstrated strong internal consistency (Tschannen-Moran & Hoy, 2001), split-half reliability (O’Neill & Stephenson, 2013), and concurrent validity with job satisfaction (Klassen et al., 2009). In this study, the internal consistency for the TSES was strong on all subscales: classroom management self-efficacy (α = .91), student engagement self-efficacy (α = .91), and instructional strategies self-efficacy (α = .89).

Copenhagen Burnout Inventory (CBI)

The CBI (Kristensen et al., 2005) is a 19-item scale measuring burnout (further operationalised as fatigue or emotional exhaustion). It has three subscales: personal burnout/exhaustion (6 items), client-related burnout/exhaustion (6 items), and work-related burnout/exhaustion (7 items). The word ‘client’ can be replaced depending on participants’ professions, and in this case was replaced with ‘student’. Personal burnout refers to general exhaustion, and an example of a personal burnout item is; ‘How often do you feel tired?’. Student-related burnout refers to exhaustion which is perceived as specifically resulting from one’s work with students, and an example of a student-related burnout item is ‘Are you tired of working with students?’. Work-related burnout refers to exhaustion which is perceived as specifically resulting from one’s work generally, and an example of a work-related burnout item is ‘Does your work frustrate you?’. Responses are measured on a 5-point scale ranging from never/almost never or to a low degree (1) to always or to a very high degree (5), with higher scores indicating higher exhaustion. Total scores are obtained by averaging participants’ scores across items on each subscale.

In past research, the CBI has demonstrated acceptable factorial and criterion-related validity (Milfont et al., 2008), strong internal consistency, and strong convergent validity with the emotional exhaustion subscale of the Maslach Burnout Inventory (Platsidou & Daniilidou, 2016). In this study, the internal consistency for the CBI was strong on all subscales: personal burnout (α = .91), student-related burnout (α = .91) and work-related burnout (α = .86).

Emotion Regulation Questionnaire (ERQ)

The ERQ (Gross & John, 2003) is a 10-item scale measuring engagement in two emotion regulation strategies: cognitive reappraisal (6 items) and expressive suppression (4 items). Responses are measured on a 7-point scale from strongly disagree (1) to strongly agree (7), with higher scores indicating greater use of the strategy. Each subscale is scored separately, with participants’ average scored across items. In past research, the ERQ has demonstrated high internal consistency and concurrent validity with measures of psychological distress (Preece et al., 2019). In this study, the internal consistency was strong for cognitive reappraisal (α = .92) and acceptable for expressive suppression (α = .68). We suggest that the lower internal consistency for expressive suppression relative to cognitive reappraisal may reflect the smaller number of items on the expressive suppression subscale. Given that Cronbach's alpha is particularly sensitive to item number, and may indeed be lower than .70 when working with only a small number of items (see Taber, 2018), we therefore retained the full expressive suppression subscale with caution.

Procedure

IRB approval was provided by the Macquarie University Human Research Ethics Committee (Approval no: 52020631115152). Participants completed the survey during the first COVID-19 wave in April 2020, when 70% had recently begun remote teaching and 28% were currently transitioning to remote teaching following the Easter holidays (Flack et al., 2020). All participants provided informed consent by completing the IRB-approved participant information and consent form within Qualtrics. The survey then began on a new webpage.

Participants were shown a statement acknowledging that “The nature of many teachers’ work has changed in response to COVID-19”. The statement further specified that ‘some items in the survey will ask about everyday experiences, before COVID-19, while others will ask about experiences during COVID-19’. Participants were then asked to answer the following items to the best of their ability, according to their personal teaching experiences at the time of survey completion. Following this, participants completed the demographic questions, TSES, CBI, and ERQ in order. Because both teaching self-efficacy and emotional exhaustion from different sources are likely to be context-specific (e.g. self-efficacy for instruction may differ in regular offline classroom contexts vs online context, as might the sense of exhaustion due to students in an offline vs online teaching context), participants were given specific prompts prior to the TSES and CBI asking them to complete the scale according to their teaching experiences during the current pandemic. Finally, participants completed two open-ended questions about specific changes to their teaching roles during COVID-19. Responses to this question are beyond the scope of the current study, and are reported separately in Van Bergen and Daniel (2022). The average time for survey completion was 19.48 minutes (SD = 12.60; note that four outliers were removed who had opened the survey and returned to it on a later day).

Results

Preliminary analyses

Descriptive statistics are displayed in Table 1, and correlations in Table 2. Personal and work-related exhaustion/burnout were highly correlated, suggesting redundancy. Drawing on Kristensen et al. (2005) conceptualisation of burnout as emotional exhaustion from multiple distinct sources, our a priori intention was to examine the predictors of exhaustion on each burnout subscale: personal, work-related, and student-related. Given our finding of redundancy between personal and work-related burnout, however, we omitted personal burnout from the subsequent regression analyses.

Table 1. Descriptive statistics for teachers’ burnout, self-efficacy, emotion regulation, and demographics.

Variable	M/Count	SD	Min	Max
Personal Burnout	61.5	23.3	4.2	100
Student-Related Burnout	41.3	21.3	0	100
Work-Related Burnout	58.8	22.4	10.7	100
Classroom Management Self-Efficacy	5	1.8	1	9
Student Engagement Self-Efficacy	4.7	1.3	1.5	8.5
Instructional Strategies Self-Efficacy	5.5	1.5	1	9
Cognitive Reappraisal	4.9	1.1	1	7
Expressive Suppression	3.5	1.1	1	6.25
Years of Teaching Experience	13.9	11	0	58

Table 2. Correlations between teacher burnout, self-efficacy, and emotion regulation.

Variable	1	2	3	4	5	6	7	8	9
1. Personal Burnout	-
2. Work-Related Burnout	.87**	-
3. Student-Related Burnout	.46**	.54**	-
4. Student Engagement Self-Efficacy	−.03	−.08	−.29**	-
5. Classroom Management Self-Efficacy	−.01	−.07	−.23**	.71**	-
6. Instructional Strategies Self-Efficacy	0.12	−.17*	−.26**	.81**	.68**	-
7. Cognitive reappraisal	−.10	−.11	−.14*	.15*	.07	.07	-
8. Expressive suppression	−.02	−.12	.04	.08	.05	.05	−.05	-
9. Teaching Experience	0.13	.11	.16*	.10	.07	.07	−.02	.09	-
10. Gender (F-0/M = 1)	−.13	−.09	.03	.02	.10	.10	−.03	.03	−.03

Note: * p < .05. **p < .01.

Work-related burnout was negatively correlated with self-efficacy for instructional strategies, while student-related burnout was negatively correlated with all three forms of self-efficacy, negatively correlated with cognitive reappraisal, and positively correlated with teaching experience.

Regression analyses

Two multiple regression analyses were conducted to measure the extent to which different types of self-efficacy and emotion regulation uniquely predict burnout. In each regression the predictors were self-efficacy (classroom management self-efficacy, student engagement self-efficacy, instructional strategies self-efficacy) and emotion regulation (cognitive reappraisal, expressive suppression), and the covariates were teacher gender and experience. The first regression included work-related burnout as a dependant variable, and the second included student-related burnout.

All statistical assumptions were met for both regression analyses. The assumption of homoscedasticity was met through inspection of the residuals against predicted values plots, which revealed no trends. The assumption of normality was met through the inspection of the normal probability plots, which were linear. In addition, Shapiro-Wilk tests performed on the residuals were not significant for student-related burnout (W = .99, p = .635) or work-related burnout (W = .99, p = .979). Thus, the null hypothesis of normal distribution was accepted. The assumption of linearity was met by inspection of the plots of residuals against each independent variable, which all demonstrated no trend in residuals. Finally, examination of the variance inflation factors for each analysis revealed no multicollinearity (range: 1.02 to 3.47).

Work-related burnout

As shown in Table 3, instructional strategies self-efficacy was a significant negative predictor of work-related burnout during COVID-19. Lower self-efficacy for instructional strategies was associated with higher work-related burnout. No other variables were significant.

Table 3. Multiple regression analysis predicting work-related burnout.

Predictor	b	SE			95% CI
			t	p	LL	UL
Classroom Management Self Efficacy	0.53	1.27	0.42	.675	−1.97	3.04
Student Engagement Self Efficacy	2.67	2.19	1.22	.223	−1.64	6.99
Instructional Strategies Self Efficacy	−4.69**	1.78	−2.63	.009	−8.21	−1.17
Cognitive Reappraisal	−2.29	1.40	−1.64	.103	−5.06	0.47
Expressive Suppression	−0.79	1.47	−0.54	.593	−3.69	2.12
Teaching Experience	0.23	0.14	1.59	.112	−0.05	0.51
Gender	−4.26	4.61	−0.92	.357	−13.34	4.83

Note. CI = confidence interval; LL = lower limit; UL = upper limit.

* p < .05 **p < .01.

Student-related burnout

As shown in Table 4, teaching experience was a significant positive predictor of student-related burnout during COVID-19. Student-related burnout increased with years of teaching experience. No other variables were significant.

Table 4. Multiple regression analysis predicting student-related burnout.

Predictor	b	SE			95% CI
			t	p	LL	UL
Classroom Management Self Efficacy	−0.88	1.17	−0.75	.452	−3.18	1.42
Student Engagement Self Efficacy	−3.01	2.00	−1.50	.135	−6.96	0.95
Instructional Strategies Self Efficacy	−0.95	1.63	−0.58	.561	−4.17	2.27
Cognitive Reappraisal	−1.94	1.28	−1.51	.132	−4.47	0.59
Expressive Suppression	0.92	1.35	0.68	.497	−1.74	3.58
Teaching Experience	0.32*	0.13	2.47	.014	0.07	0.58
Gender	3.06	4.22	0.73	.469	−5.26	11.38

Note. CI = confidence interval; LL = lower limit; UL = upper limit.

* p < .05 **p < .01.

Discussion

The aim of this study was to examine how different forms of teacher self-efficacy and emotion regulation strategies would uniquely predict burnout, operationalised as emotional exhaustion, during the COVID-19 pandemic. This study provided unique insight into the role of these varied personal resources during the pandemic, when remote teaching practices were rapidly implemented and when the number of stressors faced by teachers was likely to be high.

Drawing on the Job Demands-Resources Model (see Demerouti et al., 2001; Llorens et al., 2006), we hypothesised that teachers’ instructional strategies self-efficacy and cognitive reappraisal would each be negatively associated with burnout during COVID-19, and expressive suppression would be positively associated burnout during COVID-19. While the a priori intention was to examine the predictors of exhaustion on each burnout subscale, personal burnout was omitted from the analyses due to redundancy found between personal and work-related burnout. For the two analyses conducted, the results partially supported the hypotheses. Instructional strategies self-efficacy was negatively associated with work-related burnout, but not student-related burnout. Neither emotion regulation strategy was associated with either burnout subscale.

Interestingly, and despite Kristensen et al. (2005) conceptual distinction between exhaustion from personal, work-related, and student sources, personal burnout and work-related burnout were highly correlated in the current sample. This high correlation may reflect a lack of distinction between teachers’ personal and work lives during remote teaching. For example, moving curriculum and pedagogies to remote teaching contexts may have been so exhausting that it affected teachers’ personal lives too. This possibility is supported by Flack et al. (2020), who found Australian teachers’ reported challenges balancing work and home lives during COVID-19.

Self-efficacy and burnout

As hypothesised, lower instructional strategies self-efficacy predicted higher work-related burnout among teachers during COVID-19. One potential explanation for this finding is in the way self-efficacy changes how challenges are perceived (Bandura, 1977, 2006). Teachers with high instructional self-efficacy may have been more likely to view the need for rapid remote teaching as a positive challenge rather than a negative situation, while those with lower self-efficacy may have been more likely to experience increased feelings of doubt, hopelessness, and failure: thus leading to exhaustion around work.

Our findings for instructional strategies self-efficacy are consistent with those of Košir et al. (2020), who found an association between higher self-efficacy to instruct remotely and lower stress in European teachers during COVID-19. They also are consistent with suggestions by Haverback (2020) that self-efficacy for remote instruction may be crucial in buffering teachers from burnout during the pandemic. Of course, it is possible that teachers who are otherwise efficacious might lack specific technological skills (Haverback, 2020). Although our focus was not on remote instructional strategies specifically, however, we note that more general instructional self-efficacy nonetheless captures teachers’ adaptability and flexibility in teaching situations; including those required for remote teaching. While future research is required to confirm the direction of causality, we suggest that instructional self-efficacy may act as a personal resource in buffering the negative impacts of job exhaustion and stress.

Interestingly, classroom management self-efficacy and student engagement self-efficacy were not associated with any source of burnout. These findings contrast with those of Wang et al. (2015), who found that student engagement self-efficacy and classroom management self-efficacy (but not instructional self-efficacy) were related negatively to emotional exhaustion in everyday classroom contexts. Taken together, our results for all three forms of teaching-related self-efficacy suggest that different teaching contexts influence the relationship between self-efficacy and burnout. Instructional strategies self-efficacy may be an important personal resource during the pandemic, and particularly during the subsequent shift to remote teaching, where multiple instructional changes are managed concurrently. Interpersonal aspects of teaching, including classroom management self-efficacy and student engagement self-efficacy, may be less important during this same transition period.

While is possible that the interpersonal aspects of teaching are simply more relevant in face-to-face teaching contexts, longitudinal research is also needed to determine this pattern of findings after longer periods of remote learning. We note L. Sokal et al. (2020), for example, that classroom management self-efficacy increased across a three-month period of remote learning (from April to June 2020). Higher self-efficacy for classroom management in remote teaching contexts may therefore buffer against increasing levels of disengagement and burnout.

Emotion regulation and burnout

Contrary to our hypotheses, neither cognitive reappraisal or emotional suppression predicted higher or lower student-related or work-related burnout during COVID-19. Although no studies have yet examined teachers’ emotion regulation and burnout during the pandemic, our findings differ from those of recent studies focused on coping. For example, recent findings show that positive reframing (MacIntyre et al., 2020) can protect against teachers’ burnout during the pandemic.

There are several explanations for the non-significant findings for emotion regulation during COVID-19. First, and counter to our predictions, emotion regulation strategies may simply be less relevant for teachers’ work during remote learning than they are in everyday teaching contexts. Given that challenging student behaviours are a key emotional stressor for teachers (McGrath & Van Bergen, 2019), for example, emotion regulation may be less important when teachers are not required to respond to such behaviour face-to-face. Importantly, however, this explanation cannot fully account for the positive findings of the pandemic-related research on coping approaches (L. J. Sokal et al., 2020; MacIntyre et al., 2020).

Alternatively, the importance of emotion regulation as a personal resource for buffering against burnout may be lessened in the presence of self-efficacy. This proposition is supported by the significant zero-order correlation between cognitive reappraisal and student-related burnout, before self-efficacy is entered into the model. This explanation is also consistent with past literature on everyday classroom teaching, which has found an association between emotion regulation ability and teaching self-efficacy (Chan, 2004; DiFabio & Palazzeschi, 2008; Mouton et al., 2013). In past research examining the influence of teachers’ self-efficacy and emotion regulation on burnout, however, unique contributions of each predictor were found (Tsouloupas et al., 2010).

Finally, we highlight that past research examining emotion regulation and burnout has been conducted solely in everyday classroom contexts (Chang, 2020; Ghanizadeh & Royaei, 2015; Yin et al., 2016). It is possible that strategies such as cognitive reappraisal may be beneficial only up until a certain magnitude of stress, or only for stressors that are seen as within teachers’ control. During COVID-19, however, teachers faced several large and uncontrollable stressors that required management within a short timeframe. These include not just teaching changes but also concerns about health, family members, and other life changes. In this context of heightened stress and time pressure, teachers’ cognitive reappraisal skills may not have been sufficient to protect against exhaustion and burnout. Future studies should seek to examine further the relationship between cognitive reappraisal and burnout under stressors of different magnitudes.

Teaching experience as a risk factor

Greater teaching experience has previously been associated with higher burnout in non-pandemic contexts (Klassen & Chiu, 2010; McCarthy et al., 2009), suggesting that teachers may experience burnout over long periods of time. Consistent with these pre-pandemic findings, this study found that teaching experience was positively associated with student-related burnout during COVID-19. This study also extends past literature by highlighting that the greater sense of exhaustion in more experienced teachers is specific to work with students, and that this relationship remains even when teachers are engaged in remote learning. The finding is particularly interesting given no other variables were associated with teachers’ burnout.

While further research is needed, we highlight the possibility of a lag in the effect of teaching experience on student-related burnout: with experienced teachers continuing to express student-related exhaustion over time, even when no longer interacting face to face with students. It is possible, for example, that teachers’ perceptions of students and the degree of interpersonal satisfaction they achieve from teaching may begin to change following sufficient experiences of interpersonal conflict. Disruptive behaviour is a key stressor for teachers, and the effects of such stress may accumulate over time. With greater teaching experience, therefore, teachers may also begin to experience greater student-related fatigue.

It is also possible that teachers with greater teaching experience felt less satisfied with online engagement overall. A greater sense of exhaustion may then manifest when reaching out to students and supporting students in this way. While the association between teachers’ experience and student-related burnout did not extend to work-related burnout, and while Morley (2011) has found no difference between younger and older teachers in their use of computers in the classroom, more experienced teachers may nonetheless have hit a ‘tipping point’ in their experiences when sustaining multiple interpersonal interactions online.

Finally, teacher age may play a role. We note pre-pandemic findings from Van Droogenbroeck et al. (2014), who found that relationships with students played a key role in predicting burnout in teachers over 45 years of age. Teachers who were able to have more positive interactions with students had higher satisfaction and lower exhaustion. While it is not clear why older teachers were more strongly affected by younger teachers, Van Droogenbroeck et al. (2014) proposed that teachers often enter the profession to help students: thus, the absence of positive relationships can have adverse effects on wellbeing. For many teachers in the current study, remote learning may have removed the opportunity for positive interactions with students altogether. This in turn may have compounded teachers’ negative emotions and burnout.

Limitations and future directions

Our findings should be considered in light of four key limitations. First, given our cross-sectional design, we were only able to capture relationships between teachers’ self-efficacy, emotion regulation, and burnout at a single point in time during the onset of remote learning. Importantly, research conducted well before the COVID-19 pandemic has found evidence for a longitudinal effect of teachers’ self-efficacy on burnout (Brouwers & Tomic, 2000; Dicke et al., 2014; Schwarzer & Hallum, 2008). Given L. Sokal et al. (2020) findings of teachers’ increasing emotional exhaustion and cynicism across the first three-months of the pandemic, mirroring trends of increasing pandemic-related burnout in other intensive care professions such as nursing and health care (Abramson, 2022; Maunder et al., 2022), we consider it likely that teachers who already displayed exhaustion early on may be particularly vulnerable to negative mental health as the pandemic continues. In Australia, as in other jurisdictions, different patterns of lockdown have also been implemented across different states and territories across 2020 and 2021: thus, both individual and group variation in mental health effects are possible. We recommend that future longitudinal studies considering teachers’ burnout also consider different forms of both self-efficacy and emotion regulation.

Second, consistent with Kristensen et al. (2005), we operationalised burnout narrowly: as sense of emotional exhaustion. While exhaustion is a defining characteristic of burnout, and while Kristensen et al.’s Burnout Inventory (Kristensen et al., 2005) is valuable in that it enables exhaustion to be captured from student, work, and personal sources, our narrow focus on exhaustion meant we could not draw any additional conclusions about how teachers’ various personal resources during COVID-19 may have impacted upon their accomplishment or their feelings of cynicism and depersonalisation (see Maslach & Shaufeli, 1993).

Third, our study was entirely self-report. Other methodologies, such as observations, clinical notes, and ecological momentary assessments might be considered in future research to validate participants’ responses and provide additional insights.

Fourth, we were unable to consider the interaction of personal job resources, such as self-efficacy and emotion regulation, with situational job resources. While situational resources are beyond the control of the individual, the Job Demands-Resources model nonetheless highlights how personal and situational resources work together to buffer the individual from burnout in the face of strong job demands (Demerouti et al., 2001; Llorens et al., 2006). In the context of the COVID-19 pandemic, these situational resources might include principal and school leadership support, adequate resourcing of online and remote learning, availability of professional learning resources, government support, and a positive workplace environment. For example, emerging research by Collie (2021) has highlighted the important role of principals’ autonomy-supportive leadership in protecting teachers against a range of pandemic-related outcomes including stress, somatic burden, and emotional exhaustion. Research from Woltran et al. (2021) highlights concerns about a lack of technical equipment during the pandemic, while our own research (Van Bergen & Daniel, 2022) and that from Fray et al. (2022) each highlight a perception among teachers of being unappreciated and treated poorly by politicians and government. Drawing on these findings, large-scale hierarchical analyses should be used to consider the interaction of personal and situational resources at both the individual and the school level.

Notwithstanding the limitations of our study, there are two key implications for future research and practice. First, the finding that instructional strategies self-efficacy predicted work-related burnout during COVID-19 provides unique insight into the challenges of teachers adjusting to remote learning. Providing workplace support for teachers’ development of instructional self-efficacy during times of upheaval may help to mitigate against burnout. This may be particularly important as new pandemic waves continue to influence countries around the world, and particularly given the close connection between work-related and personal burnout in the current study. While it may not be possible to predict what instructional strategies will be useful in future, the development of robust instructional self-efficacy should offer teachers the confidence to adjust their instruction in response to a range of new and unpredicted demands.

Second, we note our somewhat surprising finding that experienced teachers demonstrated higher student-related burnout during COVID-19. This finding suggests that exhaustion may progress across careers and remain when teachers are not in direct contact with students. Future longitudinal studies should examine the processes underpinning teachers’ views on students, experience of positive emotions, and burnout across the career-span to detect such trajectories and implement support processes.

Conclusion

This study demonstrates specific personal resources which may protect against the development of burnout in teachers during the COVID-19 pandemic. Self-efficacy for instructional strategies predicted lower burnout, while teaching experience predicted higher burnout. Interestingly, and counter to our hypotheses, neither emotion regulation strategy predicted burnout in this early stage of the pandemic. While self-efficacy and emotion regulation are both regarded as important personal resources, therefore, emotion regulation may be less relevant in remote teaching contexts than it is in other challenging or difficult situations. Drawing on our significant findings for instructional strategies self-efficacy, we recommend that schools implement additional supports for teachers to feel competent with new methods of instruction during periods of upheaval. We further recommend that future research continue to monitor the changes to teachers’ efficacy, emotion regulation, coping, and burnout as the COVID-19 pandemic continues.

Disclosure statement

No potential conflict of interest was reported by the author(s).