Describing Teachers’ Well-Being Prior to and 18 Months After COVID-19 School Closures, with a Focus on Early-Career Teachers and Teachers of Color

The COVID-19 pandemic has introduced several unprecedented challenges to the U.S. public education system that have negatively impacted the nation’s teaching workforce. This study describes fourth grade teachers’ depressive and anxious symptoms and burnout prior to and 18 months after prolonged COVID-19 school closures in the 2020/2021 academic year. Within this, we provide further insight into the well-being of early-career teachers and teachers of color in order to highlight groups of teachers that may have been differentially impacted by pandemic-related stressors. Independent samples t-tests and analysis of variance (ANOVA) indicated that teachers who reported on their well-being after schools resumed markedly higher depressive symptoms, anxious symptoms, and burnout than teachers who reported prior to school closures, and that symptoms were highest among early-career teachers and teachers of color in the latter group, with teachers of color in the latter group reporting the highest symptomatology levels across all outcomes. Results can inform efforts to support teachers as we continue to navigate the long-term effects of the pandemic on teachers and students, and provide evidence that early-career teachers and teachers of color should be especially prioritized for targeted support. Plain Language Summary We described elementary teachers’ self-reported mental health symptomatology among two groups of teachers: one group that reported before the onset of the COVID-19 pandemic and one group that reported approximately 18 months later, upon returns to in-person schooling. We focused on early-career teachers and teachers of color as groups that may have experienced particularly heightened mental health challenges. All teachers in the post-closures group reported higher mental health symptoms, and early-career teachers and teachers of color reported the highest levels of symptoms.

The COVID-19 pandemic has introduced unprecedented challenges to the U.S. public education system that have negatively impacted the nation's students and teachers, magnifying already-existing (i.e., pre-pandemic) racial, socioeconomic, and other disparities (U.S.Department of Education et al., 2022).K-12 teachers, in particular, have faced drastic stressors that have exacerbated longstanding (Johnson et al., 2005;Travers, 2001) patterns of chronic stress, reduced well-being, and attrition observed in the education profession.Specifically, the pandemic and related school closures led to marked increases in stress and burnout among teachers as they navigated sudden and unexpected challenges associated with moves to remote instruction (Cardoza, 2021;Kraft et al., 2020;Pressley, 2021;Singer, 2021).Teacher workloads, a notorious career-related stressor even in non-pandemic times (Ladd, 2011), saw an immediate spike in the spring of 2020 upon initial shifts to remote instruction and remained at these higher levels through school closures (Kaufman & Diliberti, 2021).As a likely result, reduced morale and increased burnout among teachers also saw initial spikes followed by steady increases: in May of 2020, 2 months after initial widespread school closures, 25% of teachers reported significant burnout, but instead of this figure decreasing over the months following (indicating adjustment to new conditions), by October of 2020 it more than doubled (Kaufman & Diliberti, 2021).By April 2021, 1 year into the COVID-19 pandemic, 91% of teachers reported severe pandemic-related increases in job stress (EdWeek Research Center Survey, 2021).Considering these patterns, it is not surprising that U.S. K-12 teachers began reporting increased intentions to leave the field.Among key findings from a large survey of teachers conducted by the RAND Corporation, 25% of teachers reported they were likely to leave their jobs by the end of the 2020/21 academic year, and the vast majority of teachers who have quit during the pandemic have cited stress and reduced well-being as direct contributors to their exit (Diliberti et al., 2021).
We anticipate that teachers from underserved groups likely experienced greater struggles through the COVID-19 pandemic and beyond.General population research shows that underserved communities have been disproportionately impacted by the pandemic (Romano et al., 2021;Miller et al., 2021).Along these lines, emerging trends suggest that certain groups of teachers may have experienced more stressors, and thus larger reductions to their well-being, than others.For example, teachers serving low-income communities may have experienced increased challenges in supporting students in transitions to online learning due to lack of resources (technology, etc.), and teachers serving students of color (and teachers of color themselves) may have experienced increased hardships as the nation navigated the dual pandemics of COVID-19 and widespread race-based social injustice taking place simultaneously.It is no surprise, then, that teachers of color initially expressed increased intentions to leave the career (Steiner & Woo, 2021), a pattern that if substantiated via actual attrition could have pointedly negative implications for students when considering the importance of a diverse teaching workforce to the positive development of all students, and especially students of color (Clotfelter et al., 2008;Lindsay & Hart, 2017;Redding, 2019).
Another group of teachers that has not yet been given specific attention in studies of COVID-19 are early-career teachers, or teachers who are in their first 5 years of the profession.Past research examining the well-being of early-career teachers has suggested that this group may be especially vulnerable to declines in mental health and career commitment: McLean et al. (2017) illustrated that beginning teachers' symptoms of depression and anxiety increased upon career entry, and later illustrated that beginning teachers' mental health and career optimism were impacted by a host of classroom-and school-based factors (McLean et al., 2020).Importantly, both of these works illustrated the crucial role of positive relationships with colleagues and school administrators (i.e., school climate) in protecting against declines in mental health among early-career teachers.However, teachers entering the field and navigating the first years of the profession during COVID-19 likely had reduced opportunities to form such relationships and benefit from the broader climate of their school.Adjustment to the career is also a particular priority for early-career teachers, who are applying newly-acquired knowledge and skills for the first time and solidifying their beliefs of their own teaching efficacy and their place within the career.More specifically, early-career teachers spend the initial years of the career putting into practice the skills and knowledge they recently acquired in teacher preparation including establishing classroom routines, applying instruction in various ways, managing student behavior, and building positive relationships with students.These core teaching practices were likely markedly challenging for early-career teachers to practice and adjust in the context of remote instruction.Aligning with this, research among teacher samples has found that they reported feelings of inadequacy and unpreparedness in the switch to remote instruction (Kaufman & Diliberti, 2021).It would stand to reason that these feelings and resulting declines in well-being would be especially profound among early-career teachers, who were still establishing their baseline classroom skills and routines.
In the present study, we elaborate on the above knowledge of teachers' personal experiences and outcomes during the COVID-19 pandemic by describing and comparing the self-reported well-being (depressive, anxious, and burnout symptomatology) of two groups of teachers: 31 who reported their experiences with these symptoms prior to the onset of COVID-19, and 33 who reported 18 months after the initial onset of the pandemic upon returns to in-person schooling, with a focus on early-career teachers and teachers of color.The present study builds upon existing literature describing the negative outcomes of COVID-19 and related school closures on teachers in two important ways.First, we extend the timeline of investigation of the impact of COVID-19 on teachers by describing teachers' well-being before and after school closure (roughly 18 months after initial school closures and having returned to in-person instruction).Second, to empirically expand upon some known trends suggesting the well-being of early-career teachers and teachers of color may be particularly impacted, we also examined differences in well-being among these groups pre-and post-school closure.

Theoretical Framework
We ground the current study in two theories that outline how the COVID-19 pandemic is likely operating to impact the well-being of the nation's educators by impacting both their personal and professional lives.First, the Bioecological Model (Bronfenbrenner & Cici, 1994;Bronfenbrenner & Morris, 2006), while typically used to inform children's development, well illustrates the varying levels of influence one can experience in life and how these might operate to impact well-being.From life elements most proximal to an individual such as the home environment (the microsystem), to elements of the contexts individuals interact with regularly such as work and community (the exosystem), to elements of the larger social and economic context in which that individual operates (the macrosystem), to patterns of environmental change over time (the chronosystem), the Bioecological model provides a holistic picture of the many sources of influence on how individuals develop and progress through life.The COVID-19 pandemic is a rare example of an event that can have profound impact on every level of individuals' life experiences, from the most proximal, daily routines (shifts to remote work, social distancing) to severe interruptions to social and family support (inability to travel, to see family and friends, sickness and death of loved ones, loss of reliable childcare), to larger societal implications the pandemic has had (large-scale economic implications in communities, supply chain issues lessening availability of important resources).Thus, the pandemic represents an unprecedented influence that operates at every significant level of an individuals' lived experience.
In terms of individuals' professional experiences, the Job Demands-Resources Model (Bakker & Demerouti, 2007), provides a framework for how one's well-being can be impacted by their career-specific experiences.Central to this model is the division of individuals' experiences into demands and resources.Job demands deplete the energy of the individual, while job resources include the knowledge, skills, and outside supports that enable individuals to thrive in their work (Xanthopoulou et al., 2007).When job demands outweigh job resources, and/or when available resources available aren't sufficient for an individual to cope with the demands, and individual's burnout increases and their well-being declines.Studies applying this model to teachers have shown that teachers' job demands including contentious parents and stressful students/classrooms, and job resources including self-efficacy, professional development, and collegial support, are indeed related to their effectiveness and well-being (Curbow et al., 2003;Raskin et al., 2015;Roberts et al., 2019), and that additional, non-job-related resources including social supports and adaptive personal characteristics (resilience, motivation, etc.) can serve as protective factors for teachers' well-being and career commitment (McLean et al., 2017;Bakker et al., 2007).It is clear from existing research that the COVID-19 pandemic drastically increased teachers' job demands.For example, teachers had to make quick shifts to remote instructional modalities despite limited experience and expertise in these areas, with varying degrees of success in adopting these remote modalities depending on teachers' own characteristics (Guoyan et al., 2021).Such shifts likely strained teachers' existing personal and professional resources (time, energy, self-efficacy).At the same time, little to no additional resources were offered to educators on a large scale to offset these increased job demands.This combination of factors has high potential to decrease the well-being of teachers in marked ways.
Considering the Bioecological Model and the Job Demands Resources model in tandem, it is clear that important life demands and resources-both professional and personal-exist at every level of one's life system, all with great potential to be influenced by the COVID-19 pandemic.In addition to the professional demands and resources of teachers described above, non-work-related resources present in teachers' daily lives (microsystems) such as social and family support were likely drastically limited among teachers due to necessary social distancing measures, while non-work-related demands such as caregiving responsibilities likely increased.Macrosystem factors such as supply chain interruptions and interruptions to childcare infrastructure, and larger exosystem factors such as exacerbated gaps in equity (across multiple facets) and pervasive political unrest fueled largely by racism and injustice also had great potential to surface in teachers' (and the general population's) daily lives through increased demands and depleted resources.When considering these factors along with the increased job-related demands and depleted of resources of teachers specifically, it becomes clear the many ways the COVID-19 pandemic had the potential to surface in every level of teachers' personal and professional lives, and ultimately to impact their well-being.

Foundational Knowledge of Teachers' Well-Being
Outside of the COVID-19 pandemic, teaching has repeatedly been identified as among the most stressful careers (Gallup, 2013;Johnson et al., 2005;Travers, 2001), with teachers simultaneously navigating a myriad of job-related challenges including heavy workloads, struggles with classroom management, isolation from colleagues, limited resources, low wages, and low professional prestige (Curbow et al., 2003;Day & Qing, 2009;Montgomery & Rupp, 2005, Raskin et al., 2015).We examine three indicators of teachers' well-being in the present study: depressive symptoms (indicating clinical depression), anxious symptoms (indicating generalized anxiety disorder), and burnout.Depression is considered a dampening of positive affect with symptoms including fatigue and feelings of worthlessness.Alternately, anxiety is characterized by excessive worry or fear (American Psychiatric Association, 2013).Burnout is considered the endpoint of an individual's unsuccessful coping with longterm stress (Jennett et al., 2003) and includes more immediate emotional exhaustion as well as later-onset depersonalization (cynical attitudes toward students or colleagues; Maslach et al., 1996;Maslach & Jackson, 1981).
It has been well-established prior to COVID-19 that teachers experience depressive symptoms, anxiety, and burnout at higher rates than individuals in the general population and among other professions.Whitaker et al. (2013) found that teachers reported higher rates of negative mental health symptomatology than the general population, and Johnson et al. (2005) illustrated that teachers experience levels of stress comparable to emergency medical responders, police officers, and prison guards.Further, results from the 2017 Education Quality of Life Survey reported that 61% of U.S. teachers surveyed found their job ''highly stressful'' and a similar percentage characterize their present mental health as ''not good'' (American Federation of Teachers, 2017).
The negative implications of reduced well-being (i.e., burnout) among teachers for both teachers and their students has been well illustrated before the pandemic.More burnout among teachers is related to dampened motivation and professional performance (Hakanen al., 2006;Llorens et al., 2007), as well as to lower job satisfaction (Brackett et al., 2010;Klassen & Chiu, 2010;Skaalvik & Skaalvik, 2009, 2010).Further, teachers' mental health symptoms have been found to adversely impact their job satisfaction and performance (Ferguson et al., 2012;Fernet et al., 2012;Kyriacou, 2001), and burnout has been identified as a primary contributor to teachers' intentions to leave the field (Martin et al., 2012).
In addition to impacting their own outcomes, teachers' well-being also has implications for students across a wide range of outcomes and likely operates through key classroom processes.Elementary teachers' depressive symptoms have been found to negatively relate to students' math achievement (McLean & Connor, 2015) and social/emotional skills (Roberts et al., 2016).Regarding classroom processes that involve both teachers and students, teachers with more depressive symptoms have been found to provide less frequent positive feedback to their students and to underutilize types of instruction that require more effort (McLean et al., 2017;McLean & Connor, 2018), and more depressed teachers have been shown to struggle more with student behavior management (Aloe et al., 2014;Li Grining et al., 2010).Studies of how teachers' anxiety impacts students are comparatively sparse; however, it has recently been illustrated that a teachers' anxiety can impact students through processes of emotional transmission (Frenzel et al., 2021) whereby increased teacher anxiety leads to increased student anxiety (McLean et al., 2023).Teachers' burnout relates to their abilities to support students (Shen et al., 2015) and maintain student engagement (Covell et al., 2009).Additionally, burnout has been found to impact teachers' appraisals of student behavior (Egyed & Short, 2006;Kokkinos et al., 2005).Given the importance of teacher well-being for both teachers themselves and for their students, it is critical to illustrate how school closures during the COVID-19 pandemic may have changed teachers' levels of well-being, likely for the worse.This approach allows for an understanding of mean-level patterns across teachers both pre-and post-school closure and who also are characterized by additional individual factors (i.e., race/ethnicity, teaching experience) that may result in differing experiences and well-being.

The Present Study
Guided by foundational literature illustrating the importance of well-being to teachers' effectiveness and career longevity, as well as more recent work demonstrating the stark negative implications of COVID-19 and related school closures for teachers' well-being, we sought to address the following research questions: First, how do depressive, anxious, and burnout symptoms differ between teachers who reported before (''pre-closures'') and teachers who reported 18-months after (''post-closures'') COVID-19 school closures?We anticipated more depressive, anxious, and burnout symptoms among the post-school-closures compared to the pre-closures group.We also anticipated that the proportions of participants within each group who fell above the clinical cutoffs for depression and anxiety would be substantially higher among post-closures teachers.Second, are there differences in the depressive, anxious, and burnout symptoms among early-career teachers and teachers of color compared to other teachers?We anticipated that early-career teachers and teachers of color in the post-closures group would report the highest depressive, anxious, and burnout symptoms.

Procedures
Data were collected as part of a federally funded study of elementary teachers' emotions and emotion-related experiences.This study included multiple cohorts of fourth-grade teachers recruited from public elementary schools in a single Southwestern U.S. state.Data from three cohorts are included in the present study: pooled data from the first two cohorts of participants (2018/ 2019 and 2019/2020 academic years, the ''pre-closures'' group) are compared to data from the third cohort of participants (2021/22, the ''post-closures'' group) to address the research questions.Each cohort included unique teachers who underwent the same data collection procedures during their participating year.The preclosures group completed all data collection prior to March 2020 school closures, and the post-closures group completed data collection after schools resumed full-time in-person instruction in August 2021.No data were collected in the 2020/21 academic year.
Teachers were invited to participate in the study in the summer prior to their participating year.Participating teachers were asked to report on basic demographic information at the time of enrollment via a short electronic survey.During their participating year, all teachers completed electronic surveys in the fall (August) and winter (February) in which they reported on their emotions and emotion-related experiences including wellbeing.All teachers were provided monetary compensation for the completing study activities.
Teachers' fall survey data for all three cohorts are used in the present study, with the exception of Cohort 1's burnout data.The burnout measure was not added to the data collection plan until the winter of 2019; thus Cohort 1 only reported on burnout at that time point.Therefore, Cohort 1's winter measurement, and Cohort 2 and 3's fall measurements of burnout are used in the present study.

Participants
This study was conducted in partnership with 23 public elementary schools in six school districts in a single Southwestern U.S. state.All districts were located within a 50-mile area but spanned rural, urban, and suburban settings.A wide range of school socioeconomic status (SES) as indicated by percentage of students in a school enrolled in a Free and Reduced Meal (FARM) program was observed; schools ranged from 6% (considered a high-SES area) to 94% (considered a low-SES area) of students enrolled in FARM.
Fifteen teachers from 5 schools participated in Cohort 1 (2018/2019), 18 teachers from 9 schools participated in Cohort 2 (2019/2020), and 31 teachers from 9 schools participated in Cohort 3 (2021/22) for a total combined sample in the present study of 64 teachers.This combined sample was split into pre-closures (n = 33; Cohorts 1 and 2) and a post-closures (n = 31; Cohort 3) groups.In the pre-closures group, the majority of teachers were female (88%) and White (67%; 24% Latino/a, 6% African American, and 3% multiracial).Pre-closures teachers ranged in years of experience from 0 to 38 years (M = 10.33 years, SD = 8.68 years), and 54% were earlycareer teachers with five or fewer years of experience.In the post-closures subsample, the majority of teachers were female (90%) and White (84%; 13% Latino/a, 3% African American).Post-closures teachers ranged in years of experience from 0 to 27 years (M = 10.68 years, SD = 8.14 years), and 38% were early-career teachers with five or fewer years of experience.

Measures
Teacher Demographics and Grouping Variables.Teachers reported their race/ethnicity and years of teaching experience upon enrollment into the study.We used these variables to create grouping variables for career stage and race/ethnicity.For the career stage variable, teachers with five or fewer years of experience were coded as ''1'' representing early-career teachers and all other teachers were coded as ''0''.For the race/ethnicity variable, teachers of color were coded as ''1'' and all other teachers were coded as ''0''.In addition, each participant was coded in the data as either belonging to the pre-closures group or the post-closures group.
Depressive Symptoms.Participants reported on their depressive symptoms in the fall of their participating year.Depressive symptoms were measured on a 4-point scale (rarely or none of the time (less than 1 day) to most or all of the time (5-7 days)) in the fall using the 10-item version of the Center for Epidemiologic Studies Depression Scale (CESD-10; Radloff, 1977), which asks users to rate how often over the two weeks they experienced each of ten symptoms associated with depression.Example items include, ''I was bothered by things that usually don't bother me'' and ''I could not get going.''This scale is frequently used for studies of depressive symptomatology and has garnered robust evidence of reliability and validity (Orme et al., 1986;Roberts, 1980).This measure displayed adequate internal consistency in the present study (alpha = .73).
Total scores on this measure range from a minimum of 0 to a maximum of 30 points, with a score above 10 indicating potentially diagnosable clinical depression.Participants' total scores were summed across items and these sum scores were used in analyses.A dummy variable was created to indicate whether participants were above or below the clinical cutoff of 10, with participants scoring 0-9 coded as 0 (below clinical cutoff) and participants scoring 10 or above coded as 1 (above clinical cutoff).
Anxious Symptoms.Participants reported on their anxious symptoms in the fall of their participating year.Anxious symptoms were measured with the Generalized Anxiety Disorder Scale (GAD), a 7-item scale that asks users to report how often in the past 2 weeks they have been bothered by seven symptoms of anxiety.Examples include, ''Not being able to stop or control worrying'' and ''worrying too much about different things.''Teachers responded on a 4-point scale (0 = not at all, 4 = nearly every day).The GAD scale is established as valid and efficient tool for measuring anxiety (Spitzer et al., 2006), and displayed high internal consistency in the present study (alpha = .91).
Total scores on this measure range from a minimum of 0 to a maximum of 28 points, and a score of above 10 indicates potentially diagnosable clinical anxiety.Participants' scores were summed across items and these sums were used in analyses.A dummy variable was also created to indicate whether participants were above or below the clinical cutoff of 10, with participants scoring 0-9 coded as 0 (below clinical cutoff) and participants scoring 10 or above coded as 1 (above clinical cutoff).
Burnout.Participants in Cohorts 2 and 3 reported on their burnout symptoms in the fall of their participating years, and Cohort 1 reported on their burnout symptoms in the winter of their participating year.Burnout symptoms were measured using the ''frequency'' portions of the Maslach Burnout Inventory (MBI; Maslach et al., 1996), Emotional Exhaustion and Depersonalization subscales.Users rate on a 7-point scale (0 = never, 7 = every day) the frequency with which they have recently experienced each of 14 burnout symptoms.Emotional Exhaustion is captured by nine items including ''I feel emotionally drained from my work'' and ''Working with people all day is really a strain for me,'' and Depersonalization in captured by five items including ''I've become more callous toward people since I took this job'' and ''I worry that this job is hardening me emotionally.Previous studies have shown high reliability and good internal consistency of this scale among teachers (Gold, 1984;Iwanicki & Schwab, 1981), and high reliability was observed for the overall scale in the present study (alpha = .93),as well as for each of the burnout components utilized (alpha = .94for Emotional Exhaustion, .70 for Depersonalization).Higher scores on the Emotional Exhaustion and Depersonalization dimensions indicate more burnout.

Analytic Approach
To address the first research question assessing the differences in depressive, anxious, and burnout symptoms between the pre-closures and post-closures groups, we examined descriptive statistics for all outcomes for each group.Next, we conducted independent samples t-tests to determine if group means on study variables were significantly different.Next, we used frequency estimations of participants' total scores on the depressive and anxious symptoms measures to determine the proportions of each group at or above the clinical cutoff for depression and anxiety.We then used chi-square tests to determine if observed differences in these proportions between groups were significant.To address the second research question assessing differences in depressive, anxious, and burnout symptoms between teachers in the pre-closures and post-closures groups depending on career stage and race, we conducted analysis of variance (ANOVAs) to determine if teachers in each group differed based on teaching experience and identifying as a teacher of color.Tukey's HSD post-hoc tests were run to probe significant omnibus F-tests.

Comparing Depressive, Anxious, and Burnout Symptoms Pre-Versus Post-Closures
Descriptive statistics for all study variables by school closures group are presented in Table 1.We conducted independent t-tests with school closures group as the grouping variable to determine if there were significant mean-level differences in depressive, anxious, and burnout symptoms.Teachers in the post-closures group had significantly higher levels of both depressive and anxious symptoms than teachers in the pre-closures group, t (49.37) = -3.93,p \ .01,t (42.71) = -3.48,p \ .01respectively (see Table 1 for means).On average, depressive symptoms were 163% higher and anxious symptoms were 140% higher among the teachers in the postclosures group.There were marginally significant differences in mean-level emotional exhaustion between groups (t (53.13) = u`1.68,p = .10);on average, emotional exhaustion symptoms were 41% higher among teachers in the post-closures group.There were no significant differences between groups in depersonalization, t (57.94) = 1.47, p = .15.
We used the dichotomized clinical cut-off variables for depression and anxiety to assess the proportion of participants within each group who scored at or above the cut-off for each measure (see Figure 1).Chi-squared tests of association were conducted between group and clinical cut-offs, separately for depressive and anxiety symptoms.For depressive symptoms, there was no significant association between group and clinical cut-off group, x 2 (1) = .64,p = .42.For anxious symptoms, there was a significant association between group and clinical cut-off group, x 2 (1) = 6.92, p \ .01,such that significantly more teachers in the post-closures group were above the cutoff (;27%).

Outcomes for Early-Career Teachers and Teachers of Color
Early-Career Teachers.We created four groups to classify participants by school-closure group and teaching experience (pre-closure early-career teachers, pre-closure mid/late-career teachers, post-closure early-career teachers, and post-closure mid/late-career teachers).This grouping variable was used as the independent variable in ANOVAs to examine mean-level differences among groups for each of the outcomes.
There was a significant difference by group for depressive symptoms, F (3, 58) = 8.37, p \ .01.Tukey's HSD post-hoc analyses showed that post-closure early-career teachers had significantly higher depressive symptoms than both pre-closure early-career teachers and preclosure mid/late-career teachers.Post-closure mid/latecareer teachers also had significantly higher depressive symptoms than pre-closure mid-career/veteran teachers (see Figure 2).
There was also a significant difference by group for anxious symptoms, F (3, 58) = 5.17, p \ .01.Tukey's HSD post-hoc analyses showed that post-closure earlycareer teachers had significantly higher anxiety levels than both pre-closure early-career teachers and preclosure mid/late-career teachers (see Figure 2).

Teachers of Color.
Using the same approach as with career stage, we created four groups to classify teachers by their school-closure group and self-reported race/ethnicity (pre-closure teachers of color, pre-closure white teachers, post-closure teachers of color, and post-closure white teachers).This grouping variable was used as the independent variable for the ANOVAs to examine meanlevel differences among groups for each of the outcomes.
There was a significant difference by group for depressive symptoms, F (3, 58) = 9.50, p \ .01.Post-closure teachers of color had significantly higher levels of depressive symptoms than all other groups (see Figure 3).
There was also a significant difference by group for anxious symptoms, F (3, 58) = 7.15, p \ .01.Post-closure teachers of color had significantly higher levels of anxiety symptoms than all other groups.
There was a significant difference by group for emotional exhaustion, F (3, 58) = 2.95, p = .04.Teachers of color in the post-closure group had significantly higher levels of emotional exhaustion than both teachers of color and white teachers in the pre-closure groups.There was no significant difference by group for depersonalization, F (3, 56) = 2.03, p = .12.

Discussion
The present study builds on existing literature describing the impacts of COVID-19 and related school closures on the lives of U.S. teachers.We anticipated that the farreaching impacts of the COVID-19 pandemic at every level of individuals' ecological systems (Bronfenbrenner & Morris, 2006), as well as the extreme shifts in the presence of job demands and availability of job resources for teachers (Bakker & Demerouti, 2007) would result in our observing much higher rates of depressive, anxious, and burnout symptoms among teachers who reported on these factors in the Fall of 2021.As well, we anticipated that rates of symptoms would be highest among earlycareer teachers and teachers of color in our post-closures group.These hypotheses were largely substantiated in our descriptive analyses.
Recent studies have described the current state of U.S. teachers' stress, well-being, and intentions to exit the field in the context of the pandemic, however there is still much to be understood about how our nation's educators are faring during this unprecedented time, how their outcomes have shifted compared to what was observed prior to the pandemic, and how impacts of the pandemic on well-being vary depending on group belonging.As such, we highlight four contributions of the present study: First, most COVID-19 studies among teachers to date have reported rates of stress, burnout, morale, etc. among teachers mid-pandemic, but have not directly compared these to rates among teachers prior to COVID-19 within the same study.As such, our comparison of pre-closures versus post-closures teachers offers more direct conclusions about the implications of the  pandemic on teachers' well-being.Second, most related studies have examined teachers' stress, workloads, morale, and other broader constructs that do not speak directly to potentially diagnosable conditions.As such, our focus on depressive and anxious symptomatology including clinical cutoffs, and our consideration of both emotional exhaustion and depersonalization within our measurement of burnout, provides a more nuanced view of exactly where the negative impacts of the pandemic are surfacing among teachers.Third, most studies to date have described teachers' experiences immediately after the start of and during prolonged school closures.

Depressive Symptoms
Anxious Symptoms  Our comparison of teachers' well-being before versus 18months after school closures and after returns to inperson instruction can speak to the longer-term impacts of the pandemic on teachers' well-being.And fourth, little work has examined how certain groups of teachers may have been differentially impacted by the pandemic, and we address this by describing the specific experiences of two groups of teachers we anticipated would be at higher risk for adverse well-being outcomes: early-career teachers and teachers of color.We anticipated that depressive, anxious, and burnout symptoms would be significantly higher in the postclosures group, and that the percentage of individuals in the post-closures group at or above the clinical cutoff of depressive and anxious symptoms would be higher.Findings closely aligned with our predictions: the postclosures group reported 163% and 140% more depressive and anxious symptoms, respectively, and a higher proportion of post-closure teachers exceeded the cutoff for anxiety.We also predicted that depressive, anxious and burnout symptoms would be higher among earlycareer teachers and teachers of color in the post-closures group compared to their counterparts in the pre-closures group.This hypothesis was also largely substantiated: post-closure early-career teachers had 239% more depressive and 135% more anxious symptoms than their pre-closure early-career counterparts, and differences were significant between post-closure early-career teachers and all pre-closure teachers.Even more staggering, post-closure teachers of color had 490% more depressive, 342% more anxious, and 137% more emotional exhaustion symptoms than their pre-closure counterparts, and again these differences were significant between postclosure teachers of color and all pre-closure teachers.Notably, post-closure teachers of color were the only group to show a significant increase in emotional exhaustion.Following we offer some immediate conclusions as well as some broader considerations.

Conclusions
Our first set of findings that depressive and anxious symptoms were significantly higher among post-closure teachers (and that emotional exhaustion trended higher) provides broad evidence of the long-term negative impacts of the COVID-19 pandemic on the well-being of all educators, even after nationwide returns to in-person instruction.This is especially salient when examining the proportions of teachers above the clinical cutoffs for diagnosable mental health conditions: in past studies of elementary teacher mental health, teachers reported relatively few symptoms, and proportions of teachers falling above the clinical cutoffs were small (McLean & Connor, 2015, 2018).In the present study, 27% of postclosures teachers fell at or above the cutoff for potential generalized anxiety disorder diagnosis compared to just 3% in the pre-closures group.This is concerning when paired with the knowledge that even mild levels of negative mental health symptomatology can surface in the classroom to impact teachers and students (McLean & Connor, 2015, 2018;McLean et al. 2017).
Our second set of findings that the well-being of early-career teachers and teachers of color fared even worse through the pandemic than their mid-/late-career and white counterparts speaks to the uneven, and clearly inequitable, influence of pandemic-related trauma on teacher populations that were already at risk of declines in mental health and intentions to leave the field (McLean et al., 2017(McLean et al., , 2020;;Steiner & Woo, 2021).This is especially true for teachers of color, among whom we observed unparalleled increases in mental health symptomatology (137%-490% increases) who were the only group to show significantly increased symptoms of emotional exhaustion.This finding aligns with previous reports that teachers of color report intentions to leave the field at a higher rate.These patterns are immensely concerning firstly in regards to the well-being and career longevity of teachers of color themselves, but also given the well-established importance of a diverse teaching workforce to the positive development of students (Clotfelter et al., 2008;Lindsay & Hart, 2017).The disproportionate loss of teachers of color, a group already undersupported in the field of education, will very likely exacerbate the inequities among teacher and student groups that have long been observed in the education system (Lankford et al., 2002) and that have become even more marked in the wake of COVID-19 (National Assessment of Educational Progress, 2022).
Although descriptive in nature, these findings underscore the need for prioritized support for groups of educators who are more likely to experience adverse outcomes, and within this suggest that this support should be especially prioritized for teachers of color.These findings also illustrate the utility of targeted investigations into how pandemic-related stress and trauma might surface differently among different groups of teachers and students, as these uneven impacts on groups have been observed in the general population in almost every study of COVID-19 patterns and implications (CDC, 2021).While not directly tested here given our modest sample size, it is likely that teachers navigating the pandemic who are both in the early-career stage as well as identify as teachers of color may be among those most vulnerable to adverse well-being outcomes, and school administrators, district leaders, and policymakers should give this group careful consideration when prioritizing when and how to support teachers.
Lastly, we did not observe any significant differences between groups on depersonalization, which is interesting when considering foundational evidence in the burnout literature that that emotional exhaustion typically comes first, followed more distally by depersonalization.
Relating this to the present study, while there is evidence that the post-closures group may have been experiencing the emotional exhaustion phase of burnout, the lack of findings for depersonalization may suggest that they had not yet have progressed to this stage of burnout, or alternately that the stressors leading to burnout in the case of COVID-19 do not as readily translate to depersonalization from students and colleagues.

Limitations
Some characteristics of our sample and study design limit our ability to generalize to all U.S. K-12 teachers.First, this sample only included fourth-grade teachers in a single U.S. state.It could be that the effects of the pandemic differed among teacher s based on grade level taught and U.S. region of residence, and so patterns detected here and conclusions drawn should be interpreted with caution as they may not represent the experiences of all U.S. teachers.Second, our sample size was modest at 64 teachers.As such, the present study may be underpowered and thus at higher risk of Type II error or failing to detect existing effects.It is encouraging, then, that significant associations were detected; however, the findings of the present study should be replicated among a larger and more varied teacher sample.Third, all data in the present study were self-report by teachers and are at increased risk of reporter bias.We attempted to minimize this bias by ensuring that surveys were private, by describing our data privacy and security protocols to teachers in detail, by encouraging teachers to answer honestly, and by affording teachers the option to decline to answer any questions that made them uncomfortable.Last, our multiple-cohort study design (which was determined prior to COVID-19 with the purpose of studying other classroom processes) did not allow for the longitudinal investigation of variables within the same group of teachers.Rather, different groups of teachers were compared; half of whom underwent data collection prior to the pandemic and half of whom underwent data collection 18 months after initial school closures.We recognize that comparing two different groups of individuals is not as rigorous as comparing symptoms at different time points among the same group of individuals.Due to this, we reiterate that results should be interpreted with caution, should not be generalized, and should instead be used to guide future, more rigorous research on these topics.Limitations withstanding we frame our findings as an exploratory first step toward fully understanding what teachers experienced during this time.

Future Directions and Broader Implications
The findings here can inform next steps for future research.First, we again frame these results as highly preliminary insights that other researchers can use to inform next steps with more rigorous design and analytic methods, for example, using state and/or national teacher survey data to examine on a large scale how the COVID-19 pandemic has differentially impacted key outcomes among different groups of U.S. educators.Second, we feel the patterns revealed here could further substantiated using qualitative and/or mixed methods approaches that more closely attend to teachers' personal experiences through the COVID-19 pandemic.Indepth interviews and other more person-centered data collection approaches could capture a more nuanced view of how U.S. teachers responded to, and continue to navigate, the COVID-19 pandemic.Such an approach could have potential to inform more precisely how the pandemic impacted teachers (and individuals more generally) at every level of the ecological system and within this, the demands and resources that were present (or absent) in their occupational settings.
Before the COVID-19 pandemic, teachers routinely faced a wide range of barriers to well-being and career longevity, with barriers especially marked among early career teachers and teachers of color (McLean et al., 2017(McLean et al., , 2020;;Lankford et al., 2002).The added stressors of COVID-19 related long-term school closures, all of which took place simultaneously with extreme equityand race-based political and social unrest in the U.S., likely magnified the negative effects of existing barriers to well-being among teachers, and most especially among teachers belonging to the racial/ethnic and socioeconomic groups that were most negatively impacted by these unprecedented stressors.Additionally, it appears that widespread availability of vaccines and returns to in-person instruction may not have reversed the negative impacts of the pandemic on educators' mental health and burnout as some may have expected they would.A future direction that will be taken by the investigative team is to examine which buffers show the most promise in protecting teachers against these declines in well-being including positive school climate, low classroom-level adversity, and the provisions of professional development, social supports, and resources.
Investigations and resulting supports focusing on teacher well-being in and after the COVID-19 pandemic could provide much-needed direction for school leaders, education policymakers, and implementation scientists regarding which areas to target in their efforts to help teachers during this time.Such supports could see farreaching benefit for both teachers and their students, as past research has illustrated that teachers' well-being is associated with a host of student outcomes including student well-being and social, emotional, and academic development (McLean & Connor, 2015, 2018;McLean et al., 2017;Hamre & Pianta, 2004;Harding et al., 2019;Herman et al., 2018;Maricua`oiu et al., 2023;Roberts et al., 2016;Sandilos et al., 2015).Reforms to policies dictating teacher compensation, in-school supports for teachers' well-being, efforts to increase school staffing including school counselors, school psychologists, and paraprofessionals, representation of training on emotion regulation and mindfulness strategies in teacher preparation and professional development, reductions in teacher workloads and increases in proven job resources such as planning time, resources to support teachers in teaching across multiple modalities and among multiple student populations, and perhaps most importantly the provision of targeted supports to groups teachers who need it most that are responsive to the unique challenges and needs of these groups are all approaches that could positively shift the landscape of the U.S. teaching workforce and enable all teachers, and their students, to thrive.

Figure 1 .
Figure 1.Proportion of groups at or above clinical cutoff for diagnosable conditions.

Figure 2 .
Figure 2. Increases in mental health symptomatology by groups.

Figure 3 .
Figure 3. Increases in emotional exhaustion symptomatology by group.

Table 1 .
Demographic Information and Descriptive Statistics by School Closure Group.