The impact of student self-disclosure on the stress and wellbeing of tertiary educators during the COVID-19 pandemic

ABSTRACT

Workplace stress, burnout, and fatigue are commonplace amongst tertiary educators, and are compounded by the ongoing challenges of teaching and learning during a global pandemic. Amid efforts to identify and understand contributors to educator stress, student-teacher interactions have received relatively little attention. However, educators are often expected to engage in pastoral care when students disclose academic and personal problems. Receiving and responding to self-disclosure can be emotionally taxing, particularly in professional contexts of care, and therefore contribute to educator experiences of stress and burnout. In this study, we examined the relations between student self-disclosure and educator stress and wellbeing during the COVID-19 pandemic in New Zealand. Almost all of the 318 tertiary educators received COVID-19 related disclosures from students. Findings show that educators whose students had shared personal problems during COVID-19 were more likely to report high stress and poor wellbeing.Such communication was also associated with higher rates of workplace presenteeism, suggesting that these teachers were likely to push themselves to a level that risks illness. Fortunately, these negative impacts were ameliorated when educators also reported a sense of support in the workplace. The implications for educators and tertiary institutions are discussed, including the provision of educator training and well-resourced student support services.

KEYWORDS:

• Self-disclosure
• higher education
• teacher stress
• teacher wellbeing
• presenteeism

Introduction

Evidence regarding stress and burnout amongst tertiary educators has prompted discussion about the factors that may manifest in declining wellbeing amongst academics (Sabagh, Hall, and Saroyan 2018). Student interactions are understood to be a potential contributor (Watts and Robertson 2011), possibly due to the cognitive and emotional demands stemming from pastoral care provided by educators to support students’ mental, emotional, and psychological wellbeing (Laws and Fiedler 2012). Educators may be privy to self-disclosure from students struggling academically or personally. While self-disclosure is generally framed as a positive interpersonal interaction (Altman and Taylor 1973; Tardy and Smithson 2018), this is not always the case. Self-disclosure can harm individuals and relationships due to over-sharing or overstepping boundaries (Kelly and McKillop 1996; Omarzu 2000). In this article, we explore how student self-disclosure may contribute to the stress and wellbeing of tertiary educators in New Zealand. Undertaken during the COVID-19 pandemic, we examine the impact of student self-disclosure on educators and factors that may alleviate the emotional impact of student self-disclosure on educators.

Teacher stress and the workplace

Educators experience high levels of work-related stress on a daily basis (e.g. Montgomery and Rupp 2005). Some of this stress is natural to teaching, as cognitive, metacognitive, and emotional demands deplete one’s resources and coping mechanisms. Teaching is also a social context whereby interpersonal exchanges with students add complexity, flux, and a need for educators to adjust accordingly. The tertiary sector has seen considerable rises in work-related stress among faculty, due to increased managerial control, workload demands, job insecurity and work-family conflict (e.g. Mudrak et al. 2018; Shin and Jung 2014). Meta-analytic research indicates that stress is higher among teachers in universities when compared to the compulsory sector (Ozamiz-Etxebarria et al. 2021). Furthermore, during a public health crisis, teachers experience new and unexpected challenges alongside their existing stressors, reflected in reports of frustration, anxiety, and stress (Schmidt-Crawford, Thompson, and Lindstrom 2021).

Regarding work-related stress among tertiary educators, one particular concern is the emotional labour involved in supporting distressed students. Schmidt-Crawford, Thompson, and Lindstrom (2021) explain that during the COVID-19 pandemic, tertiary educators are dedicating more time to meet their students’ academic and personal needs. They observe that higher prevalence rates of mental health difficulties among tertiary students may result in an increased demand for staff to serve as teachers, counsellors, and supporters. It is plausible that students disclosing their problems to their teachers and the nature of those respective problems would contribute to greater work-related stress among tertiary educators. Teachers may be vulnerable to unmanageable stress that lowers their physical and psychological wellbeing (Spilt, Koomen, and Thijs 2011). When faced with sensitive self-disclosure from distressed students, teachers may push themselves to work despite feeling burned out, fatigued, or unwell (i.e. presenteeism). Indeed, studies have shown that sectors that involve the welfare of people tend to report more presenteeism (Aronsson, Gustafsson, and Dallner 2000).

The adverse effects of work-related stress on wellbeing may be mitigated by social support. Social support refers to the provision of psychological and material resources that potentially help an individual’s capacity to cope with stress (Amason, Allen, and Homes 1999), including quality relationships with colleagues such as supervisors and co-workers (Stinglhamber and Vandenberghe 2003). Indeed, the institution’s atmosphere, sense of community, and interpersonal dynamics with colleagues can make or break an educator’s job satisfaction and experiences with stress (Lacy and Sheehan 1997). For example, a perception of support from one’s direct supervisor may alleviate employee strain because they are in a position to provide protection, encouragement, and motivation to employees (Charoensukmongkol, Moqbel, and Gutierrez-Wirsching 2016). Co-workers also have the potential to serve as a resource that aids employees through listening, empathy, and advice. In tertiary institutions, co-workers can contribute to more positive views on teaching, lower levels of job stress, greater work-related motivation, and better staff commitment (e.g. Ahmad, Bibi, and Majid 2017; Tschannen-Moran 2001). At a broader level, a sense of organisational empowerment plays important structural and psychological roles on employees’ feeling of power and agency (Laschinger and Read 2017). Faculty members who can accomplish their work (structural role) and feel a sense of meaning, self-determination, and impact (psychological role) often experience workplace support that may be instrumental to effective coping with stress. Collectively, social support is critical to employees during times of crisis when unexpected changes occur (Cole, Bruch, and Vogel 2006). Based on the extant literature, we posit that tertiary educators’ perceptions of social support from their supervisors, co-workers, and organisation may lessen the degree to which work-related stress impairs wellbeing and increases presenteeism.

Self-disclosure in tertiary education

Self-disclosure in professional settings can be subject to unique complexities, particularly in those contexts whereby professionals act in a helping or supportive role for a client. Issues regarding self-disclosure and power dynamics, professional boundaries, role expectations, and other ethical considerations have been explored in relational, professional contexts like psychotherapy and social work (e.g. Hill and Knox 2002; Knight 2012). Literature on self-disclosure in such settings illustrates the intricacies of managing self-disclosure, not only to protect clients’ needs, but also those of the professional. Although these professionals are usually highly trained, there can be an emotional burden associated with receiving self-disclosure that must be managed (Newell and MacNeil 2010). Consequently, some professionals have access to institutional or sector support (e.g. supervision in social work settings) to help reduce the incidence of burnout and fatigue associated with the emotional impact of their job.

Tertiary education settings, such as universities and polytechnics, similarly represent highly relational contexts that may involve self-disclosure. If students are experiencing academic problems, self-disclosure to teaching staff is a critical mechanism for receiving support (Venville, Street, and Fossey 2014). Student disclosures may invoke academic or pastoral guidance from educators, such as referral to academic support services or the provision of assignment extensions or other compassionate considerations (Becker et al. 2002; Martin 2010). Providing pastoral support for students likely presents educators with more complex conditions that can impact educator wellbeing. While many tertiary institutions have some expectations regarding the educator provision of pastoral care to students (Laws and Fiedler 2012), this is often inaccurately accounted for in educator workloads (McAllister et al. 2014). This may be significant for educators teaching courses where they are the primary point of contact for hundreds of students, and may therefore be subject to receiving student self-disclosure at scale. Research has shown educators with higher teaching loads are more prone to burnout (Watts and Robertson 2011).

Educators often lack professional supervision support to help them manage the emotional burden of their job, including receiving self-disclosure from their students (Huyton 2009; McAllister et al. 2014). They are more likely to rely on collegial support from other educators or managers to help, but this can raise issues around student confidentiality if their colleagues interact with the same students (McAllister et al. 2014). Furthermore, educators may not have the professional expertise to know how to deal with issues that are raised by student self-disclosure (McAllister et al. 2014; Stanley and Manthorpe 2001). Some educators would like training on supporting psychological distress (Margrove, Gustowska, and Grove 2014). When educators have had professional development on mental health, they may be more confident in responding appropriately (McAllister et al. 2014).

For many educators, the primary course of action is to refer students to university services to provide professional pastoral care (Laws and Fiedler 2012). For some, however, institutions lack clear guidance regarding when and how such referrals should occur (McAllister et al. 2014), and who assumes responsibility for attending to the issues raised by students during self-disclosure. This confusion may influence how educators respond to student self-disclosure, perhaps resulting in inconsistent approaches where some students receive better support than others (McAuliffe et al. 2012). Moreover, if educators know student support services are struggling to meet demand, this can be another layer of emotional burden, knowing that a student requires support that is unlikely to be fulfilled by the services educators are advised to direct students to (McAllister et al. 2014). As Huyton (2009) argues, providing emotional support to students is not inherently stressful for educators; rather, the conditions characterised by lack of preparedness and a large volume of such work create the burden. Collectively, these strands of research show that educators can experience significant emotional ramifications, particularly when the disclosures are of a sensitive nature.

The current study

Like the rest of the world, New Zealand experienced significant upheaval as a result of the COVID-19 pandemic throughout 2020. In March 2020, approximately one month into the academic year, the country was moved into a strict lockdown and all tertiary institutions were physically closed, with learning conducted online for two months. From June to December 2020, most of the country was no longer in lockdown and tertiary institutions could be open for face-to-face teaching. There was widespread concern about the impact of lockdown and pandemic-associated disruption across society, including education. It is in this context that we undertook a study regarding the impact of student self-disclosure on the wellbeing of tertiary educators across the country. COVID-19 provides a unique context for exploring this phenomenon as it represents a time of greater stress for students and educators, coupled with greater awareness of the associated academic and wellbeing repercussions. While a significant amount of attention has been rightly directed towards the experiences of students during this time, we acknowledge that the effects of COVID-19 are not limited to students, and educators may bear the responsibility of receiving student disclosures about these impacts.

Materials and methods

Participants and procedures

Our sample comprised 318 tertiary educators in New Zealand. Participants were from 16 tertiary institutions across the country, including universities (64.6%) and institutes of technology (35.4%). Most participants reported their role at their institution as involving both teaching and research responsibilities (61.9%), with some individuals involved in teaching only (32.1%). Other roles included working directly with students in a student support capacity (e.g. learning advisors, librarians). Participant gender was self-reported as: Female (63.8%), Male (33.0%), Gender diverse (0.6%), and Prefer not to say (2.5%). Participants’ ages ranged from 26 to 72 years (M = 49.33, SD = 10.88).

Individuals were recruited via email through the support of each institution, asking them to complete an anonymous, online Qualtrics questionnaire about their experiences with students upon returning to the physical classroom after one semester of online teaching. Participants self-reported as having taught a course between June and December 2020. The questionnaire was open from September to December 2020 and no reminders were sent after the initial invitation to participate in the study. Participants provided informed consent electronically at the start of the survey. Ethical approval was received from the University of Canterbury Human Ethics Committee.

Instruments

In this section, we present instrument descriptions and psychometric results according to confirmatory factor analysis (CFA) with the lavaan package in R (Rosseel 2012). Before CFA was performed, we conducted preliminary assumption testing with all scale items to check for violations of normality using the Shapiro-Wilk statistic with Lilliefors Significance Correction. Although no values signalled severe non-normality, we follow recommendations by Finney and DiStefano (2006). We chose the Robust Maximum Likelihood (MLR) estimator in the lavaan package of R as the most suitable estimation method for CFA with this sample. Criteria for good CFA model fit were compared against absolute and relative indexes (Schreiber et al. 2006).

Student Self-Disclosure (SSD)

We used two items to represent student self-disclosure regarding (a) learning and coursework (SSD-Academic, or SSD-A), and (b) personal support (SSD-Personal, or SSD-P). Items were: ‘Since June 2020, [fill in the blank] have told me about the impact COVID-19 has had on their classwork or assignments’ (SSD-A) and ‘Since June 2020, [fill in the blank] have told me about the impact COVID-19 has had on their physical, emotional, and mental wellbeing’ (SSD-P). Participants responded to items using a three-point Likert-style scale, including (1) no students, (2) a few students, and (3) many students. Because we wanted to distinguish rates of SSD based on academic and personal support, we chose to maintain these items as separate, manifest variables.

Teacher Stress (TSTR)

To measure TSTR, we used two items to estimate participants’ perceived degree of work-related stress and temporal change in work-related stress. Participants responded to a seven-point Likert-style scale ranging from (1) strongly disagree to (7) strongly agree. Items were: ‘Since June 2020, I have experienced high levels of stress at work’ (degree of work-related stress), and ‘Since June 2020, the work-related stress I have experienced has been more than usual’ (change in work-related stress). Internal consistency was acceptable: α = .87.

Current Wellbeing (CWB)

For a comprehensive estimate of current wellbeing, we included a set of self-report measures previously validated in organisational psychology research. We included the following single-items and sub-scale: (1) Current Well-being (single item, 5-point scale from (1) terrible to (5) excellent: ‘How would you rate your wellbeing presently?’; Idler and Benyamini 1997), (2) Life Satisfaction (single item, 7-point scale from (1) strongly disagree to (7) strongly agree: ‘As a whole, I am satisfied with my life’.; Fisher, Matthews, and Gibbons 2016), (3) reverse-coded Burnout (single item, 7-point scale from (1) strongly disagree to (7) strongly agree: ‘I feel burned out’.; Fisher, Matthews, and Gibbons 2016); (4) Energy (single item, 5-point scale from (1) terrible to (5) excellent: ‘During the past 7 days, how would you rate your energy levels?’; Arnetz et al. 2008), (5) Sleep Quality (single item, 5-point scale from (1) terrible to (5) excellent: ‘During the past 7 days, how would you rate your sleep quality overall?’; Snyder et al. 2018), and (6) Mental Well-being using the six-item Short Warwick Edinburgh Mental Well-Being Scale ranging from (1) strongly disagree to (6) strongly agree (e.g. ‘Over the past 7 days, I’ve been thinking clearly’; see Stewart-Brown et al. 2009).

Initial results using CFA showed model fit concerns due to RMSEA confidence intervals that were below the recommended cut-offs; χ² (9) = 39.17, p < .001, CFI = .97, TLI = .95, and RMSEA = .10 [90% CI: .07-.14]. All standardised factor loadings ranged from .64 to .84. Evaluation of modification indices (MI) suggested that the estimation of one co-variance between measurement error terms for the items representing Life Satisfaction and Mental Well-being would improve model fit. Given the possibility for conceptual overlap, it seemed appropriate to estimate this parameter in the measurement model. The final model revealed good fit to data; χ² = 15.25 (8), p < .001, CFI = .99, TLI = .99, RMSEA = .053 [90% CI = .00-.094]. Internal consistency was acceptable: α = .88.

Workplace Support (WPS)

We included four items to estimate educators’ perceptions of workplace support. Two items were by Fisher, Matthews, and Gibbons (2016) and represented Supervisor Support (‘I can usually count on my supervisor/manager for support when I need it’) and Co-worker Support (‘I can usually count on my co-workers/work colleagues for support when I need it’) using a 7-point scale from (1) strongly disagree to (7) strongly agree:. The other two items were by Laschinger et al. (2000) and represented Work Empowerment (5-point scale from (1) strongly disagree to (5) strongly agree: ‘Overall, my current work environment empowers me to accomplish my work in an effective manner’ and ‘Overall, I consider my workplace to be an empowering environment’). We proposed that these four items could represent a latent construct of WPS.

Initial results using CFA showed poor model fit to the data due to RMSEA and its confidence intervals as below the recommended cut-offs for adequate fit; χ² (2) = 7.06, p = .03, CFI = .99, TLI = .96, and RMSEA = .11 [90% CI: .03-.20]. All standardised factor loadings ranged from .57 to .76. Evaluation of MI suggested that the estimation of one co-variance between measurement of the error terms for WorkEmpower_1 and WorkEmpower_2 would greatly improve model fit. It seemed appropriate to estimate this parameter in the measurement model. The final model revealed good fit to data; χ² = 0.41 (1), p = .52, CFI = 1.00, TLI = 1.01, RMSEA = .00 [90% CI = .00-.10]. Internal consistency was acceptable: α = .88.

Presenteeism

To measure presenteeism, we asked participants: ‘Since June 2020, how often have you worked despite feeling unwell, tired, and/or fatigued?’ Participants responded to a five-point Likert-style scale ranging from (1) Never to (5) Most of the time. Single-item measures of presenteeism have been advocated and used in various studies (e.g. Kigozi et al. 2014).

Positive Affect (PA) and Negative Affect (NA)

We included PA and NA as latent covariates, which were theorised to influence the associations relating to student self-disclosure and work-related factors resulting from affect, which is known to influence subjective perceptions of stress (Melvin and Molloy 2000). We used the 10-item International Positive and Negative Affect Schedule (PANAS) Short Form (I-PANAS-SF; Thompson 2007). This instrument asks participants to respond to items using a five-point Likert-style scale ranging from (1) Never to (5) Always. The prompt includes: ‘Thinking about yourself and how you normally feel, to what extent do you generally feel … ’ with items such as ‘upset’, ‘hostile’, and ‘inspired’.

Results using CFA showed poor model fit to the data due to RMSEA and its confidence intervals as below the recommended cut-offs for adequate fit; χ² (34) = 90.94, p < .001, CFI = .90, TLI = .97, and RMSEA = .07 [90% CI: .06-.09]. Standardised factor loadings ranged from 37. to .71. The factor loading that was below the traditional cut-off of .40 was Hostile. Evaluation of MI suggested that the estimation of two co-variances between the measurement of the error terms for (a) Hostile-Upset and (b) Ashamed-Afraid, would greatly improve model fit. It seemed appropriate to estimate these parameters in the measurement model, as the items were theorised to represent negative affect dimensionality. The final model revealed good fit to the data; χ² = 40.45 (32), p = .15, CFI = 99, TLI = .98, RMSEA = .03 [90% CI = .00-.05]. Internal consistency coefficients were acceptable: PA (α = .88) and NA (α = .88). Thus, positive and negative affect were computed using their respective subscale’s manifest indicators (items) and the aforementioned covariances. We found a correlation of r = −0.47 between the two factors, consistent with existing literature (Thompson 2007).

Statistical analysis

We conducted our analyses in three steps. First, we examined descriptive statistics and bivariate correlations among the confirmed study variables. Second, we inspected the psychometric properties of the full measurement model using the same CFA procedures and criteria for good model fit as described in the Instruments section. Third, we tested the structural model that was designed to measure associations between Student Self-disclosure (SSD-A and SSD-P), Teacher Stress (TSTR), Current Well-being (CWB), Workplace Support (WPS), and Presenteeism while controlling for Positive Affect (PA) and Negative Affect (NA). We used structural equation modelling (SEM) procedures with the lavaan package in R. We inspected direct, indirect, and total effects and corresponding effect-size values for all endogenous variables.

Results

Descriptive statistics

We began our analyses by inspecting descriptive statistics among the study variables (see Table 1). Values were well within the recommended criteria for normality (Kline 2011). We found that tertiary educators received self-disclosure from students regarding academic (93.80%) and personal (86.20%) concerns. With regards to student self-disclosure for academic concerns (SSD-A), almost all respondents received disclosure from many (50.20%) or a few (43.60%) students. Similar trends were present for personal self-disclosures (SSD-P): many students (42.40%), a few students (43.80%), and no students (13.80%).

Table 1. Descriptive statistics among the variables of interest (n = 318).

	M	SD	Skewness	Kurtosis
1. Student Self Disclosure – Academic (SSD-A)	2.44	0.61	−0.60	−0.57
2. Student Self Disclosure – Personal (SSD-P)	2.29	0.69	−0.45	−0.87
3. Teacher Stress (TSTR)	5.64	1.40	−1.27	1.50
4. Current Well-being (CWB)	3.69	0.91	−0.17	−0.38
5. Workplace Support (WPS)	3.98	1.12	−0.43	−0.53
6. Presenteeism	3.62	1.02	−0.44	−0.29
7. Positive Affect (PA)	18.78	2.62	−0.56	0.90
8. Negative Affect (NA)	12.27	2.60	0.41	0.49

To better understand the descriptive information, we performed a series of independent-samples t-tests and Analysis of Variance (ANOVA) to identify meaningful group differences in demographic characteristics. With cisgender self-reports only,¹ we found that female educators rated higher levels of SSD-A; t(292) = 2.89, p = .004, and for SSD-P; t(207.06) = 4.79, p < .001. We found small correlations between course enrolment size and SSD-A (r = .16, p = .008) and for SSD-P (r = .15, p = .02). We found no differences in self-disclosure rates based on age, employment status (part-time vs. full-time), or contract type (permanent vs. fixed-term contract).

Next, we inspected Pearson’s r correlations among the computed study variables (see Table 2). Noteworthy results highlight statistically significant associations among estimates of Student Self-disclosure (SSD-A and SSD-P), Teacher Stress (TSTR) and Presenteeism. We also report partial evidence relating estimates of SSD to Current Well-being (CWB) and Positive Affect (PA). We performed curve estimation procedures to test for non-linearity among correlations that were not statistically significant at the p < .05 level; however, no identifiable patterns were found.

Table 2. Bivariate correlations among the variables of interest (n = 318).

	1. SSD -A	2. SSD-P	3. TSTR	4. CWB	5. WPS	6. Presenteeism	7. PA	8. NA
1. Student Self Disclosure – Academic (SSD-A)	1		-	-	-	-	-	-
2. Student Self-Disclosure – Personal (SSD-P)	.708**	1	-	-	-	-	-	-
3. Teacher Stress (TSTR)	.255**	.293**	1	-	-	-	-	-
4. Current Well-being (CWB)	−.136*	−.102	−.534**	1	-	-	-	-
5. Workplace Support (WPS)	.016	.078	−.265**	.404**	1	-	-	-
6. Presenteeism	.254**	.205**	.537**	−.558**	−.308**	1	-	-
7. Positive Affect (PA)	.041	.122*	−.193**	.526**	.322**	−.216**	1	-
8. Negative Affect (NA)	.014	−.013	.171**	−.387**	−.197**	.204**	−.365**	1

*p < .05. **p < .01. ***p < .001.

Measurement model

Our measurement model included three manifest variables (SSD-A, SSD-P, and Presenteeism) and five, first-order latent factors (TSTR, CWB, WPS, PA, and NA). Using the same constraints and estimation method as were performed for CFA procedures, our model resulted in acceptable fit to data; χ² = 387.31 (252), p < .001, CFI = .96, TLI = .95, RMSEA = .04 [90% CI = .03, .05]. Standardised factor loadings ranged between .36 and .92.

Structural model

Our structural model includes 20 structural regression paths. We present the model with its standardised regression paths and coefficients of determination (R²) in Figure 1. Note that we do not present the structural paths relating to PA and NA for ease of interpretation. Fourteen of the 20 regression paths were statistically significant. Five of the six non-significant paths related to the control variables PA and NA. The only non-significant regression path depicted in Figure 1 (dashed line) was that from SSD-A to TSTR (β = 0.07, p = .40). This highlights that educators were more stressed when students disclosed having personal problems than when students disclosed having issues with learning and coursework.

Figure 1. Structural model predicting the standardised influences of types of Student Self-disclosure (SSD) on Teacher Stress (TSTR), Current Well-being (CWB), Workplace Support (WPS), and Presenteeism while controlling for Positive Affect and Negative Affect (n = 318). Paths with solid lines are statistically significant and dashed lines are not statistically significant at the p < .05 level. Coefficients of determination (R²) are included for all endogenous variables. Note: *p < .05. **p < .01. ***p < .001.

We summarise direct, indirect, and total effects in Table 3. As presented, SSD-A exerted no direct, indirect, or total effects on the variables of interest. Meanwhile, SSD-P exerted several noteworthy results, including a positive, direct effect on TSTR (β = .29), a negative, indirect effect on CWB (β = −.12), and a positive, indirect effect on Presenteeism (β = .14). Our results show that TSTR exerted a negative, direct effect on CWB (β = −.43) and a positive, direct effect on Presenteeism (β = .34). In terms of indirect and total effects, we found that TSTR had a negligible, indirect effect on Presenteeism (β = .05). The total effect of TSTR on Presenteeism was, therefore, moderate and positive (β = .39).

Table 3. Summary of standardised direct, indirect, and total effects for Student Self-disclosure – Academic (SSD-A), Student Self-disclosure – Personal (SSD-P), Teacher Stress (TSTR), Workplace Support (WPS), Current Well-being (CWB), Positive Affect (PA), and Negative Affect (NA)(n = 318).

	SSD – Academic			SSD – Personal			TSTR			WPS			CWB			Presenteeism
	Direct	Indirect	Total	Direct	Indirect	Total	Direct	Indirect	Total	Direct	Indirect	Total	Direct	Indirect	Total	Direct	Indirect	Total
SSD-A	-	-	-	-	-	-	.07	-	.07	-	-	-	-	−0.03	−0.03	-	.03	.03
SSD-P	-	-	-	-	-	-	.29***	-	.29***	-	-	-	-	−0.12***	−0.12***	-	.14**	.14**
TSTR	-	-	-	-	-	-	-	-	-	-	-	-	−0.43***	-	−0.43***	.34***	.05	.39***
WPS	-	-	-	-	-	-	−0.30***	-	−0.30***	-	-	-	.19***	.13***	.32***	−0.13**	−0.22***	−0.35***
CWB	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	−0.36***	-	−0.36***
PA	.08	-	.08	.18**	-	.18**	−0.10	.06	−0.04	.37***	-	.37***	.37***	.02	.39***	.12**	−.15***	−0.03
NA	.07	-	.07	.10	-	.10	.08	.03	.11*	−0.10	-	−0.10	−0.16**	−0.04	−0.21***	0.02	0.09	0.11*

*p < .05. **p < .01. ***p < .001.

WPS exerted a negative, direct effect on TSTR (β = −.30), positive, direct effect on CWB (β = .19), and a negative, direct effect on Presenteeism (β = −.13). In terms of indirect and total effects, we found that WPS exerted a small, indirect effect on CWB (β = .13), slightly weaker than its direct effect on CWB. The total effect of WPS on CWB was, therefore, moderate and positive (β = .32). We also found that WPS had a small, indirect effect on Presenteeism (β = −.22), thus stronger than its direct effect on Presenteeism. The total effect of WPS on Presenteeism was, therefore, moderate and negative (β = −.35).

We include the effects of PA and NA on all variables in Table 3. We note that PA had small-to-moderate, positive direct effects on SSD-P (β = .18), WPS (β = .37), CWB (β = .37), and Presenteeism (β = .12). PA exerted a small, negative indirect effect on Presenteeism (β = −.15), resulting in a null total effect (β = −.03). We note that NA had no direct effects on the variables of interest except for a small, negative direct effect on CWB (β = −.16) and a small, positive total effect on TSTR (β = .11). These findings result in a small-to-moderate, negative total effect on CWB (β = −.21).

To understand coefficients of determination for the endogenous variables, we consider R² values presented in Figure 1. Coefficients of determination calculated were: SSD-A (.01), SSD-P (.02), TSTR (.26), WPS (.18), CWB (.66), and Presenteeism (.43).

Discussion

In this study, we explored the impact of student self-disclosure on a sample of New Zealand tertiary educators. Our results present evidence that in the wake of nationwide COVID-19 lockdowns, there was a high prevalence of students sharing academic and personal concerns with their lecturers, with cascading negative effects on educators. We found student self-disclosure contributes to individual differences in teachers’ work-related stress, current wellbeing, and rates of presenteeism. Educators who received greater student self-disclosure tended to report higher levels of stress, yet continued working even though they were burned out, unwell, or fatigued. Put simply: some of our best teachers – those who welcome their students and encourage them to communicate – are at considerable risk.

One crucial observation was that the content of self-disclosure had differential impacts on teacher stress. Student disclosures relating to academic concerns contributed minimally to teacher stress, whereas self-disclosure relating to physical, emotional, and mental wellbeing problems contributed to higher teacher stress levels. We theorise that the discrepant impact of the two types of self-disclosure may be for several reasons. Receiving academic self-disclosures from students may be viewed as ‘part of the job’ and can be dealt with objectively, requiring a less emotionally-taxing response from educators. They may even have standard responses to specific disclosures that occur regularly and require little emotional labour (e.g. sharing learning strategies and directing students to relevant academic resources). Additionally, educators likely have the resources to meet student academic needs and therefore feel better equipped to respond to those disclosures. This system, in turn, is likely to lessen the impact on their work-related stress and personal wellbeing. In the context of COVID-19, there may have also been greater institutional support and resourcing available for educators to pass on to students, particularly in relation to the transition to online learning.

The increased burden of receiving personal self-disclosures from students aligns with the literature we surveyed earlier (Becker et al. 2002; Laws and Fiedler 2012; McAllister et al. 2014; Stanley and Manthorpe 2001). To minimise the impact of student disclosure on educator wellbeing, it is critical for institutions to proactively and reactively support the management of educators’ emotions. Institutions could provide psychological first aid training or similar professional development opportunities (Margrove, Gustowska, and Grove 2014). However, in doing so, institutions would have to acknowledge that staff do engage in substantive amounts of emotional labour as a part of pastoral care. This is essential, as this reality has historically been underestimated and undervalued in tertiary education (McAllister et al. 2014). Furthermore, the association between personal disclosures and reduced wellbeing may indicate compassion fatigue. This phenomenon occurs when employees in helping professions experience stress brought on by working conditions associated with learning about the details of another person’s stressful circumstances (Radey and Figley 2007). We agree with Cordaro (2020) that tertiary educators – especially in the context of COVID-19 – may be vulnerable to compassion fatigue by receiving personal disclosures from highly distressed students.

To support students and staff, institutions should ensure the student support pipeline is clear, effective and reliable. Feeling overwhelmed by student concerns but not trusting the system will likely create further problems for all parties involved. While we suspect that COVID-19 may have heightened the impact of personal student disclosures on educators, the existing literature strongly suggests it exists under normal circumstances and is not sufficiently addressed by tertiary institutions (Huyton 2009; McAllister et al. 2014). Institution-wide procedures which are consistent and coherent across courses and programmes could reduce the demands that seem to be placed on individual educators. Based on the current findings, such practices could be particularly beneficial to educators responsible for very large courses, as they appear to experience markedly more student self-disclosure.

This study also highlights that educators’ perceptions of workplace support – from their supervisors, colleagues and a general sense of work empowerment – may lessen the adverse effects of student self-disclosure challenges on work-related stress. Sharing the emotional burden of hearing student disclosures with fellow educators may alleviate stress by providing a space for them to embrace and articulate their emotional response to disclosure. This may also present opportunities for educators to receive advice from colleagues or supervisors. It is heartening that collegial relationships can ameliorate the impact of receiving disclosure from distressed students, in line with the extant literature on social support for academics (Sabagh, Hall, and Saroyan 2018). However, this may be complicated by limitations that arise as a consequence of a public health crisis. Workplace support in education, for instance, may feel uncomfortable when teachers are restricted to online interactions. We argue that the potential influences of social support on tertiary educators should be further explored in research that considers both in-person and online workplace settings.

Institutions must take meaningful steps towards supporting staff in their capacity to manage and respond to student self-disclosure related to physical, emotional, and mental wellbeing. Although social support from fellow staff may help educators cope with student disclosure, we argue it would be insufficient for institutions to rely on that collegiality alone. Resting on the shoulders of others may help offset individual stress in the short term; however, such reliance may cause group-level stress in the long term. To ensure that students and educators are adequately supported, institutions must effectively reduce (or, ideally, prevent) the associated educator stress and presenteeism. To foster and enhance social support among collegial networks, institutions could consider mentoring approaches, buttressed by quality training for mentors.

We also note that since our data collection period, the COVID-19 landscape has endured and shifted. Tertiary education systems and the people within them – students and staff alike – have adapted to new variants, public health measures, and institutional policies regarding teaching and learning. What appears to be consistent across this period is that COVID-related effects on social, emotional, mental, and financial wellbeing persist. It is unclear when these will cease. It seems likely then, that impacts on student engagement and performance at university will continue, with the consequent demand placed on staff via student self-disclosure potentially having cumulative effects on educator stress and burnout. Educators have always been a target for student self-disclosure, but as individuals try to cope with challenges in the wake of COVID-19, the impacts of disclosure may be compounded.

To address this, the aforementioned recommendations for professional development and robust student support services should be underpinned by accurate accounting of the time and energy many educators put into pastoral care (McAllister et al. 2014). Without institutional acknowledgement and systems put in place that recognise increased staff responsibilities regarding pastoral care – especially with regard to student self-disclosure – educators may continue to struggle, experiencing frustration, presenteeism, and burnout. Institutions should work with educators to devise support systems that reflect current realities. This requires institutions to (a) acknowledge that students are in need of support, (b) understand that educators do unaccounted for pastoral work that calls for training and support, and (c) ensure that institutional programmes involving students, educators, and support staff are clear, accessible, and work effectively.

From a theoretical perspective, our findings support a social cognitive model that reflects how student disclosure contributes to educator wellbeing. Teaching is inherently social: regular interpersonal exchanges in the classroom facilitates good teacher-student relations and on this basis, students may disclose to educators. Upon receiving messages of distress from students, educators may struggle to cope with the cognitive and emotional demands that disclosure provokes. Consequently, educators can experience greater stress, decreased wellbeing, and increased negative workplace behaviours like presenteeism. However, inabilities to cope with student disclosure may be offset by information and comfort-seeking coping strategies within the workplace environment. In our structural model, social support and work empowerment seemed to be essential buffers against personal and professional difficulties among tertiary educators. This further reinforces the value of social relationships for educator practice and wellbeing. Overall, further research that considers how initial coping mechanisms and institutional systems could be improved are urgently needed.

Our study was situated in the context of COVID-19 and we do not have data to suggest how the trends here may compare to trends in student disclosure to educators under pre-pandemic circumstances. Pre-COVID evidence suggests there was already a high prevalence of student disclosure to educators (Lindecker and Cramer 2021). The collectivity of the experience of the pandemic and widespread messaging about mental health may have encouraged students to speak to educators about their issues. Additionally, educators themselves may have been more vulnerable to the negative impacts of receiving disclosure from students as they had already been subject to the stresses associated with teaching amid a global pandemic. We note further that because we surveyed educators across multiple higher education institutions, we are limited in how much we can speak to issues at an institutional level (e.g. accessibility and availability of counsellors) since different institutions have varying structures and services available. There is also a potential sampling bias as participants responding to the survey invitation may be especially concerned with their students and their profession.

This study offers a perspective on educator experiences of stress and student disclosure. Much of the limited literature on this topic entails qualitative studies of small samples of educators. While the paucity of research makes exploratory research valuable, our findings offer much needed quantitative insights into the association between student self-disclosure and educator stress, wellbeing and presenteeism. In addition to quantifying the prevalence of student self-disclosure, our findings clearly illustrate the emotional toll disclosure can have on educators. Our view is that tertiary institutions need to better support educators who engage in the pastoral care of students. This begins with acknowledging and valuing the emotional labour that pastoral care requires and providing meaningful pathways for managing the wellbeing impacts educators may experience as a consequence of student self-disclosure.

Disclosure statement

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

Notes

1. We used a dichotomous, cisgender (male–female) categorisation of gender as we had an insufficient subsample size to adequately represent experiences by gender‐diverse participants and those preferring not to report their gender. Our sample for this analysis was, therefore, n = 293.