Abstract
Background:
The COVID-19 pandemic forced districts to rapidly adjust their policies in ways that altered teachers’ working conditions. Teachers’ perceptions of how conditions changed could impact their well-being, job satisfaction, and organizational commitment.
Purpose:
We examined the relationships between Virginia teachers’ perceptions of how their working conditions changed, district pandemic policies, and teacher outcomes (job satisfaction and organizational commitment), thereby highlighting a potential pathway through which the pandemic may have influenced teacher outcomes.
Research Design and Findings:
This descriptive quantitative study leveraged an early spring 2021 statewide survey (n = 56,229) and unique data on districts’ COVID-related policies. Some 43.7% felt working conditions had stayed about the same since 2019, 33.4% said they had improved, and 22.9% said they had worsened. Teachers were more likely to blame COVID for worsened working conditions than to credit COVID for improvements. District policies predicted teachers’ perceptions, which were significantly related to their outcomes. Notably, more in-person learning predicted perceptions of worsened conditions, and a positive view of how districts communicated policy changes and facilitated teachers’ safe return to school predicted perceptions of improvements.
Conclusions:
Although no district policy will be viewed positively by all teachers, how districts make and implement decisions will play a role in how teachers respond. Effective communication and ensuring teachers feel safe and supported in schools can help district leaders address divergent views, buffer teachers from stress, and bolster organizational commitment.
As school districts prepared for the fall 2020 start of school, the COVID-19 pandemic forced them to review all aspects of their operations and identify changes necessary to provide staff and students with a safe and supportive teaching and learning environment. Districts had just months to assemble a team to complete the review, decide on changes, communicate decisions to staff and students’ families, and implement those changes (Virginia Department of Education, 2020). The process continued throughout the 2020–21 school year as the pandemic’s effect on local communities evolved. Some districts made numerous significant policy changes, whereas others made fewer, less substantive ones.
The ways that districts responded to the pandemic altered teachers’ working conditions. Districts’ most prominent policies concerned which students were expected to attend school remotely rather than in person (Darling-Aduana et al., 2022). Other critical policy decisions included which health protocols to adopt, how to accommodate teachers who were uncomfortable teaching in person, how to support students’ internet access, and how to aid teachers with remote instruction (Evidence Project, 2021). All these decisions were consequential for teachers and their students.
Although a dominant narrative during the first full school year after the pandemic’s onset (2020–21) was that teachers’ working conditions had deteriorated, this was not universally the case. Some teachers may have viewed the shift to remote instruction as improving their working conditions. They did not have to believe that remote instruction was superior to in-person teaching but instead that it was appropriate given the ongoing health emergency. Then again, teachers may have held the opposite view, especially if they were less concerned about health risks or more focused on their ability to connect with and effectively educate their students. Additionally, teachers’ opinions on the shift to remote instruction likely varied depending on the supports their districts provided to bolster remote instruction.
Teachers’ perceptions that their working conditions were deteriorating during the pandemic is concerning for multiple reasons. Prepandemic research found that teachers often reported high levels of stress (Diliberti et al., 2021), a trend likely exacerbated by the pandemic’s introduction of new challenges and amplification of existing stressors. When teachers lack sufficient resources to cope with stress, it can lead to burnout (Bakker & Demerouti, 2007, 2017; Demerouti et al., 2001). Supportive working conditions are a crucial resource that helps buffer teachers from the negative impacts of stress. Conditions also contribute to a professional culture where teachers feel empowered and capable, as opposed to unsupported, disrespected, and ineffective (Johnson, 2019; Johnson & Birkeland, 2003). Consequently, it was likely that teachers who perceived COVID had deteriorated their working conditions were more susceptible to burnout and declining well-being and were less likely to either intend to or actually remain in the classroom (Simon & Johnson, 2015).
Much about the pandemic was outside district officials’ direct influence; however, they were able to influence how it changed teachers’ working conditions and their perceptions of those changes. Working conditions are malleable. Districts can intentionally act to improve teacher perceptions, thereby enhancing teacher well-being, retention, and effectiveness. The pandemic offered an opportunity to study the linkages between districts’ policies and teacher perceptions of how their working conditions have changed.
Our descriptive study combines responses to a statewide Virginia teacher survey (n = 56,229) from early 2021 with unique granular data on district COVID-related decisions. We answer these research questions:
How did teachers perceive changes to their working conditions between 2019 and 2021, and did teachers attribute perceived changes to COVID?
How did job satisfaction and retention intentions vary with teachers’ perceptions of how working conditions changed?
How did districts’ COVID policies predict how teachers perceived that their working conditions had changed? How do the predictions vary with teacher and school characteristics?
Our answers to these questions provide guidance to districts on supporting teacher well-being as they move beyond the pandemic.
Background
School districts have considerable capacity to shape teachers’ working conditions via the impact the decisions they make and the policies they adopt have on daily school operations. District leaders’ approach to developing and implementing policies can influence teachers’ commitment to these policies, the fidelity of their implementation, and their ultimate effectiveness (Brazer et al., 2010; Finster & Milanowski, 2018). Involving key stakeholders, such as teachers, in policymaking can be beneficial. It has been positively correlated with teacher satisfaction, organizational commitment, and retention (Bogler, 2005; Ingersoll & May, 2012; Jackson, 2012; Kelly, 2004; Park et al., 2023; Shen et al., 2012; Somech, 2010). Conversely, an overreliance on centralized power during policy implementation, for instance merely passing down new decisions with associated rule changes, can compromise the underlying rationale and commitment to the policy (Brazer et al., 2010).
Teachers connect changes in district policies to salient aspects of their day-to-day experiences and working conditions. In doing so, they form personal views of the policies. Their reactions and responses to policy changes can vary with their individual characteristics, roles, and school culture (Datnow & Castellano, 2000). Teachers’ perceptions of how district-level decisions impact their working conditions contribute to the amount of stress they feel.
Like all occupations, teaching has specific risk factors associated with stress and reduced well-being. According to the Job Demands-Resources theory, teachers experience job-related stress when they lack sufficient resources to cope with the job’s demands (Bakker & Demerouti, 2007, 2017; Demerouti et al., 2001). Working conditions, features of the school’s environment that affect teaching (Merrill, 2021), are multidimensional and comprise both job demands and resources. Supportive working conditions can contribute to a professional culture that empowers teachers to feel and be successful. Conversely, unsupportive conditions can leave teachers feeling disrespected, unsupported, and ineffective (Johnson, 2019; Johnson & Birkeland, 2003). Teachers in schools with supportive working conditions are more satisfied with their jobs (Grissom, 2011; Johnson et al., 2012), more committed to their jobs (Rosenholtz & Simpson, 1990), more likely to plan to remain teaching (Kraft et al., 2021; Ladd, 2011), and more likely to remain teaching at the same school (Ingersoll, 2001; Loeb et al., 2005).
Job demands require sustained effort, resulting in physiological or psychological costs to the worker (Bakker & Demerouti, 2007, 2017; Demerouti et al., 2001). Examples of working conditions that are job demands for teachers include emotional labor, workloads, accountability mandates, and time pressures (Demerouti et al., 2001; Lee & Ashforth, 1996). Although demands are not inherently harmful, they can become stressors and cause strain when teachers struggle to recover and bounce back from the energy expended addressing them (Hakanen et al., 2006). Consequently, job demands are an antecedent of exhaustion, which predicts reduced teacher performance, overall well-being, and student achievement (Bakker et al., 2004, 2005; Granziera et al., 2023; Hakanen et al., 2006).
On the other hand, job resources enable teachers to respond to demands and perform their jobs effectively (Bakker & Demerouti, 2007, 2017; Demerouti et al., 2001). Working conditions that are resources for teachers include collegiality, autonomy, professional learning opportunities, administrative support, and feedback (Admiraal & Kittelsen Røberg, 2023; Collie et al., 2018; Demerouti et al., 2001; Johnson, 1990; Kraft et al., 2021; Lee & Ashforth, 1996; Pressley et al., 2022). Teachers with sufficient job resources tend to have higher work engagement, which predicts stronger organizational commitment (often measured by job satisfaction and intention to remain teaching) and job performance (Bakker et al., 2004; Bakker & Demerouti, 2007; Hakanen et al., 2006; Schaufeli & Bakker, 2004).
Teacher Well-Being During the Pandemic
The COVID-19 pandemic introduced unprecedented challenges. Districts’ responses to those challenges altered how teachers perceived their job demands and accessed resources. As the virus spread and evolved, districts implemented sudden policy changes to prioritize the safety of their students and staff. Most teachers were unaccustomed to remote instruction and had to swiftly acclimate to new, unfamiliar teaching methods while accommodating students whose well-being was at risk because of the pandemic. Aware of these new job demands, districts provided additional resources such as technical support to teachers and students, computers to students, and internet connectivity. These adjustments occurred while teachers simultaneously coped with personal pandemic concerns and the ongoing struggle to maintain a work–life balance. Researchers promptly sought to investigate whether and how these combined challenges might have impacted teachers’ well-being.
There were concerning indicators in the months immediately after the spring 2020 shutdown of nearly all public schools in the United States. Many teachers felt disconnected or left without resources to cope with online teaching (Baker et al., 2021). However, not all teachers responded in this manner. Teachers participating in a different prepandemic study rated their jobs as significantly less stressful, their coping better, and their satisfaction higher in May 2020 than in fall 2019 (Herman et al., 2021).
As disruptions to normal operations grew throughout the 2020–21 school year, multiple surveys highlighted the significant struggles teachers were experiencing. One large-scale online survey collected responses from over 2.6 million professionals across the United States between September 2020 and March 2021, including 135,000 teachers (Kush et al., 2022). Alarmingly, teachers were significantly more likely than all other workers to report symptoms of anxiety and significantly more likely than healthcare workers to report depressive symptoms (Kush et al., 2022). One nationally representative online survey from early 2021 revealed that over three-quarters (78%) of teacher respondents reported experiencing frequent job-related stress, while 27% reported symptoms of depression (Steiner & Woo, 2021). Another online survey of the same sample found that 51% of teachers felt that burnout was a major concern (Zamarro et al., 2022). These findings suggest that teachers’ rapidly changing, highly demanding working conditions during the pandemic may have contributed to their declining well-being.
COVID-Related Policy Decisions, Teachers’ Working Conditions, and Teacher Well-Being
Prior research has highlighted two notable district-level policy changes that impacted teachers’ working conditions and well-being: health and safety protocols and decisions about teaching modalities.
Health and Safety
Concerns over COVID exposure and inadequate mitigation strategies emerged as major sources of stress and anxiety for teachers (Pressley et al., 2023; Steiner & Woo, 2021; Zamarro et al., 2022). Although teachers might have been used to dealing with students’ illnesses prepandemic, COVID introduced a heightened level of risk as fears mounted over potentially acquiring a severe or fatal case or experiencing unquantified long-term consequences. District policies around health and safety, such as masking, social distancing, and ventilation, played a crucial role in alleviating or exacerbating teachers’ concerns.
For some teachers, dissatisfaction with COVID-19 health policies and unenforced mitigation strategies were among the top reasons for leaving or considering leaving the profession (CDC Foundation, 2021; Diliberti et al., 2021). Nearly 20% of teacher respondents who quit between March and December 2020 stated that their salaries were insufficient to compensate for the stress and safety risks (Diliberti et al., 2021). Furthermore, COVID-related anxiety predicted teacher burnout during this period (Pressley, 2021; Zamarro et al., 2022).
Teaching Modality
Policies that shifted instruction away from in-person instruction—changes to teaching modality—placed unprecedented demands on teachers during the 2020–21 school year. Districts’ decisions about instructional modality ranged between fully remote and fully in person, with many adopting a hybrid approach where students learned in person on some days and remotely on others, but where teachers were often required to teach in person most days. These decisions often varied across the grades and throughout the year. While some teachers continued primarily teaching in person, the majority faced learning how to instruct remotely, managed simultaneously teaching students in person and remotely via livestream, and/or adapted to transitions between modalities as their districts changed policies throughout the year (Goldhaber et al., 2022; Halloran et al., 2021; Sachs et al., 2022).
Teaching in new modalities emerged as a highly ranked source of job stress in early 2021 (Steiner & Woo, 2021). Compared to teachers who predominantly taught in person in the 2020–21 school year, those who mainly taught in the hybrid modality had a higher probability of worrying about burnout and contemplating leaving due to COVID (Zamarro et al., 2022). Remote teachers also reported higher rates of depressive symptoms during late 2020 and early 2021 (Kush et al., 2022).
Changing modality was a crucial factor in teachers’ well-being. Teachers who switched modalities at least once were more likely to consider quitting or changing schools, even when adjusting for burnout concerns (Zamarro et al., 2022). In a sample of nearly 8,000 teachers across nine states, all reported struggles transitioning to remote instruction (Kraft et al., 2021). Moreover, teaching in new modalities required extensive use of and reliance on technology (Padhi & Miller, 2022; Pinckney & Miller, 2022). An analysis of teachers who left by December 2020 due to COVID found they reported more technology-related challenges than those who left for other reasons (Diliberti et al., 2021).
Study Contribution
In this study, we explore the relationships between districts’ responses to the pandemic and teachers’ perceptions of how and why their working conditions have changed since 2019. Based on the literature, we infer that teachers who perceived that their working conditions had worsened also perceived their jobs as more demanding and/or as providing insufficient resources to help them cope with those demands compared to before the pandemic. In turn, these teachers were more likely to experience declining well-being. We also predict the opposite: teachers who perceived their working conditions had improved viewed their jobs as providing sufficient resources to handle their fluctuating job demands, thereby supporting their well-being.
Our study makes three substantial contributions. First, we constructed and incorporated a rich set of district COVID-related policy measures not explored in other studies. Second, by investigating the role of districts’ COVID-related policies on teachers’ perceptions of changes in their working conditions, we have highlighted one potential pathway through which the pandemic may have impacted teacher well-being, job satisfaction, and retention, which other studies have examined (Camp et al., 2023; Hirschberg & Davidson, 2023; Redding & Nguyen, 2024). Third, we examined how the relationship between district policies and teachers’ perceptions of changes to their working conditions varied among key groups of teachers. Lastly, we conducted our study in Virginia, a state that has not been the primary focus of much research but where COVID’s impact on teacher retention has been particularly pronounced (Barnum, 2024; Katz & Miller, 2023). Our findings speak directly to districts’ decisions during the pandemic and provide important guidance for district policymakers moving forward as they seek to keep and bolster their teacher workforce.
Methods
Data
Our primary data source was teachers’ responses to the 2021 Virginia School Survey of Climate and Working Conditions (the VSS). Virginia administered the survey from January through March 2021 to teachers in all regular public schools. 1 All but eight eligible schools participated, resulting in 67,110 teacher responses (71% response rate) at 1,847 schools (over 99%) across all school districts in the Commonwealth. The survey’s anonymous nature meant we did not have information on nonrespondents. Although we cannot guarantee that the respondents represented all Virginia teachers, the survey captured teachers at nearly every school. The sample’s gender and ethnic-racial composition—81% female, 79% White, 10% Black, and 4% Hispanic—was very similar to all teachers statewide—79% female, 80% White, 13% Black, and 3% Hispanic. We supplemented the survey data with publicly available data on schools and their communities from numerous district, state, and federal sources. Sources include the Virginia Department of Education (VDOE), the Virginia Department of Health, the Virginia Board of Elections, the National Center of Education Statistics’ Common Core of Data, Broadband Now, the U.S. Census Bureau, districts’ reopening plans, and districts’ online platforms (websites, Facebook pages, and Twitter feeds).
Sample
The final analytic sample had 56,229 teacher respondents who expressed an opinion about how their working conditions had changed since 2019. Although the sample excluded the 16% of respondents who did not have an opinion on a key survey question, all 1,847 schools remained in the sample. There were only a handful of differences between the included and excluded teachers, as demonstrated by Cohen’s d and variance ratio tests in Table 1. For one, excluded teachers were more likely to be Hispanic and less likely to be White than included teachers. Also, the experience profiles differed between the groups, with excluded teachers having less experience than included. We expected this difference because we reasoned that teachers not yet teaching in 2019, relative to those who were, would be less likely to have an opinion on changes. Schools likely hired these new teachers to fill vacancies from turnover in prior years. Turnover rates increased with the schools’ concentration of minoritized students, which may explain why the schools in which less experienced excluded teachers taught had higher concentrations than included teachers’ schools (Katz & Miller, 2023). To support inferences about all teacher respondents, we applied inverse propensity weights to make our analytic sample resemble the respondent population, as shown in the last two columns of Table 1.
Unweighted Sample Compared to Weighted Sample with Balance Statistics.
Notes: LN = natural log, SR = square root, DCR = district’s COVID response; aCohen’s d imbalanced if <0.1 or >0.1; bVariance ratio imbalanced if <0.08 or >1.25. Robust standard errors in parentheses.
Measures
Using multiple data sources, we constructed measures of how teachers perceived their working conditions changed, job satisfaction, retention intentions, districts’ COVID-related policy decisions, COVID-relevant district characteristics, and teacher and school characteristics.
How COVID Changed Working Conditions
Our single outcome variable identified teachers’ perceptions of both how their working conditions changed and COVID’s role in that change. The measure combined teacher responses to two survey questions. The prompt “Over the last two years, the working conditions for teachers in this school have. . .” gauged how teachers perceived their working conditions had changed, if at all. We divided teachers into three groups: those who felt their working conditions had worsened (responses of “become much worse” or “become somewhat worse”), “stayed about the same,” or had improved (“become somewhat better” or “become much better”). As mentioned earlier, we excluded teachers who responded, “I have no opinion.” Teachers reporting changes were asked a follow-up: “To what extent do you think COVID-19 contributed to this change in working conditions for teachers?” We divided teachers into those who attributed their changed working conditions primarily to COVID (responses of “very much” and “it is the only reason”) versus primarily to something else (“not at all,” “very little,’’ or “some”). Our outcome measure, therefore, identified five groups of teachers: those who felt their working conditions had worsened because of COVID, had worsened because of other reasons, had stayed about the same, had improved because of other reasons, or had improved because of COVID. 2 Although the pandemic was likely the largest and most consequential change from 2019 to 2021 for most teachers, it was not the only change. Our outcome variable allowed us to assess teachers’ perceptions of the pandemic’s role in driving change in their working conditions.
Job Satisfaction and Career Plans
The survey included questions about teachers’ job satisfaction and retention intentions. Teachers’ job satisfaction was captured on a six-point Likert from strongly disagree (1) to strongly agree (6) to the statement, “Overall, my school is a good place to work and learn.” We operationalized job satisfaction using z-scores in addition to indicators for being unsatisfied (i.e., disagreeing) or most satisfied (i.e., “strongly agree”). Teachers also picked from eight responses to indicate their immediate career plans. We categorized their retention intentions as staying teaching at the same school, transferring to another school in Virginia, and leaving teaching in Virginia.
Covariates
To understand the variation in teachers’ perceptions of how and why their working conditions changed, we constructed measures of COVID-relevant factors, teacher characteristics, and school characteristics. We selected covariates for our models based on research and news stories about how schools, teachers, and students responded to the pandemic, as well as relevant literature on teachers’ well-being.
District COVID-Related Policies
Districts implemented many decisions that could have influenced teachers’ perceptions of how their working conditions changed and, by extension, their well-being. One of the most visible consequences of the pandemic was the shift away from traditional in-person instruction. Studies on pandemic teaching modalities found that Virginia relied more heavily on remote learning and less on fully in-person schooling than other states (Goldhaber et al., 2022; Halloran et al., 2021). For context, 80.9% of Virginia’s students were fully remote in September 2020 and only 2.7% were fully in person; however, by the end of the survey window, just 4.3% were fully remote, 84.3% were hybrid, and 11.4% were fully in person (Sachs et al., 2022). Access to in-person learning varied across grade levels. Kindergarteners, on average, could attend school in person for 42.2% of the school year compared to 35.4% for 5th graders and 28.6% for 12th graders (Sachs et al., 2022). We developed measures to reflect this shift from students’ and teachers’ perspectives.
As part of a larger project examining the pandemic’s impact in Virginia, we assembled a daily tracker of each district’s grade-specific decisions regarding whether and which students could attend school in person each day (Sachs et al., 2022). Utilizing this tracker, we calculated the percentage of the academic year (up to the teacher’s survey completion date) that students in the teacher’s school had the opportunity to learn in person—18.5% on average. Some districts, however, changed their in-person learning policy throughout the year, causing disruptions to students’ and teachers’ routines. To account for the potential strain related to these changes, we created an indicator of whether each district had altered the in-person learning offered at the teacher’s school as of their survey-taking date.
Decisions regarding students’ access to in-person learning directly impacted teachers, because they needed to provide instruction wherever the students were. To accommodate teachers’ potential preferences or medical concerns, some districts adopted approaches to allow for flexibility. The survey captured this aspect by asking teachers how they had taught students since the beginning of the school year, with the options to select all that applied from in person, remotely from a school building, or remotely from elsewhere outside the school. We assigned each teacher to one of four groups: (1) those who taught only in person, (2) taught in person and remotely, (3) ever taught remotely from inside a school, and (4) only taught remotely from outside the school. The third group of teachers had not instructed students in person but may have also taught remotely from outside the school building. Comparisons among these four teacher groups allowed us to test for differences between dual-modality (those who taught both in person and remotely) and single-modality teachers (group 2 versus the other three), between teachers who taught students in person versus those who did not (groups 1 and 2 versus groups 3 and 4), and between teachers who taught from a school building versus those who did not (groups 1, 2, and 3 versus group 4).
The shift to remote instruction emphasized using technology as a vital tool for teaching and learning. To assess the extent of support districts provided to facilitate this transition, we created five indicator variables from the qualitative coding of Virginia districts’ 2020–21 school year reopening plans (Pinckney & Miller, 2022). These indicators captured whether the district (1) supplied some or all students with technology devices, (2) offered technical support to students or (3) staff, and (4) provided students with connectivity in homes or (5) at designated community locations like parking lots, libraries, or community centers. We hypothesized that the more technology-related support the district provided, the better equipped teachers would feel to meet their students’ needs.
The above variables represented just some of the districts’ responses to the pandemic. To capture how positively each teacher viewed their district’s COVID response, we averaged their responses to four survey questions, each with a six-point Likert scale from strongly disagree (1) to strongly agree (6). 3 The four items were: (1) “My division’s response to the pandemic was well communicated with staff”; (2) “My division’s overall response to the pandemic was appropriate given the circumstances”; (3) “I felt comfortable about returning to school this year after the COVID-19 closures”; and (4) “I have sufficient resources to help my students make up the learning lost as a result of COVID-19.” 4
COVID-Related District Characteristics
How districts responded to the pandemic correlates with various community characteristics (Schueler et al., 2023). We included four measures to account for districts’ local contexts. First, we captured each community’s political leanings with the mean Republican party vote share across the 2016 presidential, 2017 gubernatorial, 2018 U.S. House of Representatives, and 2019 Virginia House of Delegates elections as recorded by the Virginia Department of Elections. Next, we calculated the COVID case and death rates (per 10,000 residents) for each district using the Virginia Department of Health data and local population data from the U.S. Census Bureau. We assigned each teacher the values for the day before they took the survey. Finally, using data from Broadband Now, we estimated the percentage of residences covered by high-speed broadband (100 mbps) in June 2020.
Teacher and School Characteristics
We used standard measures of teacher and school characteristics. Teacher covariates from the 2021 VSS were gender, race, experience, primary subject taught, and whether the teacher grew up in the same or a neighboring district. The VDOE provided data on school type (prekindergarten, elementary, middle, high, or combined), total school enrollment, and student populations for the 2019–20 school year. 5 The student population characteristics included each school’s percentage of non-White (minoritized) students, students with disabilities, students that were English learners, and economically disadvantaged students. Last, we identified each school’s locale (rural, town, suburban, urban) from the National Center of Education Statistics’ Common Core of Data.
Analytic Strategy
We employed simple descriptive statistics and cross-tabulations to answer our first and second research questions. To assess the role of districts’ COVID-related decisions on teachers’ perceptions of how their working conditions changed, our third research question, we estimated a multinomial logistic regression model (Equation 1).
The model predicted how teacher i in school s in district d perceived COVID to have changed their working conditions since 2019 (Yisd as a function of district COVID-related policies (Disd), district COVID-relevant factors (Cd), teacher characteristics (Tisd), and school characteristics (Ssd), all discussed above. We clustered the standard errors at the district level to reflect how teachers are nested within districts. To improve the normality of covariates, we used the square root of three variables (percentage of English learners, percentage of students with disabilities, and COVID death rate) and the natural log of two others (school enrollment and COVID case rate). To examine how these predictions varied across teachers and schools, we altered Equation 1 by adding interactions between district policies (Disd) and one teacher or school characteristic at a time. Finally, we assigned each teacher the weight previously discussed.
Instead of discussing the relationships between districts’ COVID-related policies and teachers’ perceptions in terms of changes to odds ratios, we used the fitted coefficients to calculate the probability a teacher held a specific perception. Specifically, we calculated the average teacher’s probability of holding each of the five perceptions captured by our outcome under different district COVID-related policies. To calculate the probability of perceiving working conditions had worsened or improved, we combined the probabilities of attributing the change to COVID and to other reasons. 6 We then tested whether the change in each probability associated with a change in a single district policy significantly differed from zero. We tested all pairwise comparisons for the binary and categorical measures of district policies (e.g., provided community connectivity versus did not). For the continuous policy measures, we compared the probabilities when the policy was set at the 25th and 75th percentiles (e.g., the amount of in-person learning the district offered students was at the 75th versus the 25th percentile). A strength of this approach was that it simultaneously modeled the relationships between district policies and a teacher’s full probability profile. In doing so, it did not force a policy’s relationship with one end of the perception spectrum (worsened/blame) to be the inverse of the relationship to the other end (improved/credit) (i.e., more likely on one and less likely on the other).
Results
Perceived Changes to Working Conditions
When teachers completed the survey, between five and eight months into the 2020–21 school year, 43.7% of teachers felt their working conditions had stayed about the same since 2019. The remaining teachers were more likely to say their working conditions had improved (33.4%) than had worsened (22.9%). Although this finding had less than a quarter of teachers aligned with the dominant public narrative that working conditions were deteriorating, the survey’s question wording and context are important. Teachers who selected the “stayed about the same” response option may have, in fact, felt that conditions had changed, but that the change was not sizable enough to qualify as “somewhat” of a change. Another reason that would contribute to fewer teachers reporting worsened conditions would be that they may have been hesitant to say negative things about their school on a state-administered survey. Conversely, it was also possible that teachers, concerned for their health and safety, viewed instructing students remotely as an improvement to their working conditions. The survey would not have captured many teachers’ shift back to in-person learning, a change many teachers found disruptive, as most teachers responded before the governor required districts to offer in-person instruction to at least some students. Furthermore, the survey administration occurred months before hostile public messaging replaced the public’s vocal support for teachers that was present at the start of the school year. We found evidence of a shift in teachers’ perspectives during the survey administration period. Teachers who responded in March, after the governor’s mandate went into effect, were 2.1 percentage points (9.5%, p < 0.001) more likely than those who responded in January to perceive that their working conditions had worsened and 1.9 percentage points (5.5%, p < 0.05) less likely to believe their working conditions had improved.
When we layered on teachers’ views of COVID’s role in the changes, we found that slightly more teachers credited COVID for the improved working conditions than blamed COVID for worse working conditions (15.4% versus 14.3%). However, a different pattern emerged when we conditioned on how they perceived their working conditions had changed. Of teachers who said working conditions had worsened, 62.6% identified COVID as the main reason. In contrast, just 46.1% of those who said conditions had improved credited COVID for that improvement.
Association with Job Satisfaction and Career Plans
Perceptions that working conditions changed were positively correlated with job satisfaction and retention intentions (Table 2). Job satisfaction among teachers reporting improvement was 83% of a standard deviation higher than the satisfaction of those reporting deterioration. Teachers who perceived improvement were also 74% less likely to be unsatisfied (4.8% versus 18.6%) and more than 2.5 times more likely to be most satisfied (55.5% versus 15.6%). Teachers reporting improved working conditions were more likely to intend to remain teaching at their current school (89.3% versus 68.0%) and less likely to intend to stop teaching in Virginia’s public schools (6.4% versus 15.3%).
Job Satisfaction and Career Plans by Perception of Changes in Working Conditions.
Notes: n = 56,229. Unsatisfied teachers are those who selected “Strongly Disagree,” “Disagree,” or “Somewhat Disagree.” Most satisfied teachers selected “Strongly Agree.” All statistics are weighted.
The direction of differences in job satisfaction and career plans were the same among teachers blaming and crediting COVID as they were for teachers who indicated overall worse or better. Surprisingly, disparities in job satisfaction and career plans were not as stark among those attributing changes to COVID as we anticipated from the overall findings, suggesting that the variations we observed were primarily driven by larger differences among teachers who attributed changes to non-COVID reasons. Perhaps this implies that teachers, at that point, viewed the pandemic’s effect as temporary and expected that the changes would be reversed once the pandemic ended. Conversely, other factors driving change may have been more entrenched and enduring.
District COVID Policies and Perceived Changes to Working Conditions
We have presented how district policy responses relate to the predicted probability of how working conditions changed and COVID’s role in that in Table 3. In the online supplemental materials, we have provided the fitted regression coefficients in appendix Table A1.
Predicted Probabilities of the Perceptions of How Working Condition Had Changed Due to COVID Policies.
Notes: n = 56,229. +p < 0.1, *p < 0.05, **p < 0.01, ***p < 0.001. DCR = district’s COVID response. See appendix Table A1 for the fitted regression coefficients and appendix Figures A1–A4.
Policies Regarding Student In-Person Learning and Changes to Those Policies
District policies that made more in-person learning available to students were associated with teachers having stronger negative opinions of how their working conditions had changed and COVID’s role in those changes. Teachers in districts that offered more, versus less, in-person learning were 4.9 percentage points (p < 0.001) more likely to state their working conditions had worsened but only 3.4 percentage points (p < 0.001) less likely to say conditions had improved. We found a similarly large difference between the negative and positive perspectives for teachers’ COVID attributions. Teachers in districts that offered more in-person learning were 3.7 percentage points (p < 0.001) more likely to blame COVID for worsening conditions but only 1.5 percentage points (p < 0.01) less likely to credit COVID for improving them, compared to those in districts that offered less in-person learning. Any change to this policy was not significantly associated with teachers’ perceptions.
Policies Regarding Where and How Teachers Taught
With the shift toward remote instruction, teachers could instruct students in person, remotely, or both. Some districts required teachers to deliver remote instruction from the school or to teach students in person and remotely simultaneously. Comparisons among teachers who instructed students in multiple modalities and/or from different locations revealed that teachers’ experiences were related to their perceptions of working conditions.
First, our findings suggested that teachers who entered schools to teach students in person or remotely were more likely to believe that their working conditions had worsened. Relative to teachers who only taught remotely from outside the school building, in-person-only teachers were 3.7 percentage points (p < 0.01) more likely to say working conditions had worsened. The same was true for dual-modality and remote-inside teachers who were 5.0 and 2.2 percentage points more likely to say conditions worsened than remote-outside teachers (p < 0.001 and p < 0.01), respectively. All three groups of teachers who taught from inside school buildings were also more likely to blame COVID than teachers who did not (between 1.5 and 4.8 percentage points, p < 0.001). Across the three comparisons, there were inconsistent patterns for reporting working condition improvements and crediting COVID for that.
Second, teachers’ exposure to students, and potentially COVID, might explain why both groups of teachers who taught students in person at any point were more likely than remote-only teachers to view their working conditions negatively. This was especially true for dual-modality teachers who were 2.8 percentage points (p < 0.01) more likely than remote-inside teachers and 5.0 percentage points (p < 0.001) more likely than remote-outside teachers to say that conditions had worsened. Additionally, in-person-only teachers were 3.7 percentage points (p < 0.01) more likely than remote-outside teachers and, although not significant, more likely than remote-inside teachers to say conditions had worsened, consistent with the overall pattern. Again, there were inconsistent findings for perceiving improvements.
On the other hand, the attribution analysis revealed consistent patterns. However, they suggested that teachers who taught in person were more likely to both blame and credit COVID than those who only taught remotely. For example, in-person-only teachers were more likely than remote-outside teachers to both blame and credit COVID (3.1 percentage points, p < 0.001, and 2.0 percentage points, p < 0.05, respectively). The same was true for dual-modality teachers relative to remote-inside teachers (3.3 percentage points, p < 0.001, and 1.2 percentage points, p < 0.01, respectively).
Third, dual-modality teachers held more negative perceptions of their working conditions than single-modality teachers. They were more likely than remote-inside and remote-outside teachers to believe their working conditions had worsened (2.8 percentage points, p < 0.01, and 5.0 percentage points, p < 0.001, respectively). Although not statistically significant, the difference from in-person-only teachers followed the overall pattern. Furthermore, dual-modality teachers were more likely to blame COVID than all three groups of single-modality teachers (between 1.7 and 4.8 percentage points, p < 0.01). The results for the relationships between dual-modality teaching and perceptions of improved working conditions and crediting COVID were not consistent.
Policies Regarding Technology
Only two of the five technology policies we examined had a significant relationship with teachers’ perceptions, and the findings for these policies ran contrary to what we had expected. First, teachers in districts that provided devices to their students, compared to those in districts that did not, were 3.7 percentage points (p < 0.05) more likely to report that their working conditions worsened and 2.6 percentage points (p < 0.01) more likely to blame COVID for the change. Device provision was not significantly related to views of improved working conditions or COVID’s role in that. Second, teachers in districts that provided staff with technical assistance were 2.9 percentage points (p < 0.01) less likely to believe their working conditions had improved and 1.1 percentage points (p < 0.05) less likely to credit COVID than teachers in districts without the support.
The lack of support for what we had expected may be due, in part, to a lack of statistical power. Unlike the other policies we examined, the technology policies were fixed across teachers in all schools in a district. There were only 132 districts in Virginia. Lowering the statistical significance threshold to 10%, there was some evidence that the policies that provided students with technical support were associated with more teachers stating that their working conditions had improved (2.5 percentage points) and crediting COVID for that (1.3 percentage points). Also, providing connectivity within the community was associated with fewer teachers blaming COVID for worsened working conditions (1.8 percentage points).
Teachers’ Assessment of Their District’s Overall COVID Response
Teachers who assessed their district’s overall COVID response more positively were more satisfied with how the district communicated its decisions, felt the district’s responses were appropriate for the situation, felt more comfortable returning to school, and believed they were sufficiently resourced to address students’ learning loss. Compared to their peers with less positive views, these teachers were 20.2 percentage points (p < 0.001) less likely to report that their working conditions worsened and 19.0 percentage points (p < 0.001) more likely to report improved conditions, a difference of 1.2 percentage points. The attribution results revealed a larger difference between the negative and positive perspectives. Teachers with more positive views, relative to those with less, were 12.5 percentage points (p < 0.001) less likely to blame COVID, but only 7.1 percentage points (p < 0.01) more likely to credit COVID. These findings, perhaps more than the others, demonstrate the potential for districts to influence how teachers perceive and feel about their working conditions.
Moderating Role of Teacher and School Characteristics
The main model results revealed significant differences across various teacher and school characteristics in how teachers perceived that COVID had changed their working conditions (see Table 4). For example, female teachers were more likely to say their working conditions had worsened and less likely to say conditions had improved than male teachers. Similarly, White teachers were more critical of perceived changes than their Black or Hispanic peers, as were teachers in schools with smaller, compared to larger, concentrations of economically disadvantaged or minoritized students. Consequently, we explored whether these characteristics moderated the relationships between district policies and teachers’ perceptions (see Table 5). Here we have focused on the two policy variables with especially large relationships in the main model—student in-person learning and view of the districts’ COVID response. Results for other district COVID-related policies are in appendix Table A2.
Predicted Probabilities of the Perceptions of How Working Condition Had Changed by Teacher and School Characteristics.
Notes: n = 56,229. +p < 0.1, *p < 0.05, **p < 0.01, ***p < 0.001.
Differences Across Teacher and School Characteristics in the Predicted Probabilities of Perceptions of How Working Conditions Had Changed by COVID Policy.
Notes: n = 56,229. Results from separate models that interacted one teacher or school characteristic with all policies. Teachers’ race/ethnicity was the only characteristic that had significant difference in the differences for the percentage of the school year in person. See appendix Tables A3 through A6 estimated coefficients from these models. +p < 0.1, *p < 0.05, **p < 0.01, ***p < 0.001.
Views of Districts’ COVID Response
Our results showed a consistent pattern in the association between teachers’ positive views of their district’s COVID response and how they perceived their working conditions changed. The magnitude of those associations, however, differed significantly with the moderators. For instance, the same change in how positively teachers viewed their district’s response was associated with a larger reduction in the probability of reporting worsened conditions among female than male teachers (1.2 percentage points, p < 0.05) and among White than Black teachers (3.8 percentage points, p < 0.001). The trend was reversed on the other end of the response spectrum. The increase in probability of reporting improved conditions was larger among male than female teachers (1.4 percentage points, p < 0.001) and among Black than White teachers (5.4 percentage points, p < 0.001). A similar pattern was evident with respect to the school-level moderators. The reduction in the probability of blaming COVID for worsened conditions was larger among teachers in schools with smaller concentrations of economically disadvantaged or minoritized students (1.3 and 1.6 percentage points, respectively, p < 0.05), while the increase in the probability of crediting COVID for improved conditions was greater among teachers in schools with larger concentrations (1.2 percentage points, p < 0.01, and 1.3 percentage points, p < 0.05, respectively).
In-Person Learning for Students
We found very few differences in how the amount of in-person learning offered to students was related to teachers’ perceptions of their working conditions. There were no significant differences across the teacher or school characteristics in the policy’s relationship with teachers’ beliefs about working condition changes. The only significant differences for COVID attributions was by teachers’ race/ethnicity. For the same change in the amount of in-person learning offered, the increase in the probability of blaming COVID was larger among White than Black teachers (2.5 percentage points, p < 0.001). Also, the reduction in the probability of crediting COVID was larger among Black than Hispanic teachers (6.7 percentage points, p < 0.01).
Discussion
Our primary focus was to understand how and why Virginia teachers perceived their working conditions changed during COVID and how districts’ COVID policies were related to those perceptions. We lacked an objective measure of change in working conditions; however, studying teachers’ perceptions has tremendous value, because teachers’ subjective experiences shape their well-being and related outcomes (Johnson & Birkeland, 2003; Zee & Koomen, 2016). As of early 2021, teachers were most likely to believe that working conditions had stayed about the same and were more likely to believe that working conditions had improved than worsened. Although not what might have been expected, we believe these findings may be partly due to the timing of the survey, as we have previously stated. Had the survey been administered later in the pandemic, the distribution of teachers’ perceptions may have been different. Nonetheless, teachers’ early perceptions were important because they provided a signal of teachers’ well-being and likely influenced how teachers felt and acted later.
We found evidence that districts’ COVID policies were more strongly associated with teachers’ perceptions of worsening working conditions and attributing blame to COVID than their perceptions of improving conditions and crediting COVID. On average, teachers’ probability of reporting that conditions had worsened was 2.2 times more responsive to district policies than their probability of reporting that conditions had improved. Similarly, their probability of blaming COVID was 3.1 times more responsive than their probability of crediting COVID.
For most policies, we did not uncover an inverse relationship between worsening and improving or between blaming and crediting COVID (e.g., a policy was associated with an increase in blaming and a decrease in crediting). Just two policies had an inverse relationship: the percentage of the school year the teacher’s school offered in-person learning and teachers’ positive views of their district’s COVID response. For the other policies, only one of the relationships was significant, neither were, or both were, but in the same direction. The latter only occurred with teaching locations, for which three comparisons exhibited the pattern. Teachers who taught only in person were more likely to both blame and credit COVID than those who taught remotely from (1) inside 7 or (2) outside schools, as well as (3) teachers who taught both in person and remotely than those who taught remotely from inside schools. The fact that, across all three comparisons, teachers who ever taught in person were more likely to blame and credit COVID for changing working conditions suggests a divide among teachers in their perceptions of this prominent district COVID policy.
Although not necessarily indicative of another divide among them, our findings highlighted how various groups of teachers perceived changes in their work stressors and demands differently. For instance, we found that female teachers were more likely to have negative perceptions and less likely to have positive perceptions of changes to their working conditions than their male peers. This was consistent with pandemic-era research that illustrated a gender-based trend in teachers’ reports of COVID. Female teachers were more likely to report high levels of job-related stress, mental health symptoms, and concerns about job-related burnout than male teachers (Hirschberg & Davidson, 2023; Kush et al., 2022; Steiner & Woo, 2021; Zamarro et al., 2022).
Furthermore, our findings suggested that teachers accustomed to challenging work environments or high job demands may have been less sensitive to negative changes or viewed policy-related changes less drastically than their peers used to more stable conditions. Teachers in schools with a higher concentration of economically disadvantaged students and minoritized students were more likely to perceive working conditions improved, more likely to credit COVID, and less likely to blame COVID. Higher-poverty, higher-minority schools tend to have higher rates of bullying (National Center for Education Statistics, 2022), lower student achievement (Willms, 2010), higher rates of teacher turnover (Carver-Thomas & Darling-Hammond, 2017), and less favorable teacher views of working conditions (Boyd et al., 2011; Grissom, 2011; Johnson et al., 2012). Given these disparities, our findings may suggest that teachers from more disadvantaged schools were more accustomed to high job demands and, accordingly, better prepared to handle COVID-related challenges. Initially, our results seemed to contradict Kraft and colleagues (2021), who found that teachers in majority Black and high-poverty schools had difficulty adjusting to COVID-induced challenges. However, our findings are consistent with changes in turnover among Virginia teachers. Katz and Miller (2023) found that turnover among schools with the highest concentrations of economically disadvantaged and minoritized students decreased slightly relative to spring 2019 to the end of the 2020–21 school year. In contrast, turnover rates increased by more than 2 percentage points at schools with the lowest concentrations of economically disadvantaged students and increased by more than 5 percentage points at schools with the lowest concentrations of minoritized students. These turnover patterns may have also been related to districts’ learning modality policies, as White students had access to more in-person learning than did Black or Hispanic students (39%, 28%, and 27%, respectively; Sachs et al., 2022).
Districts had to weigh numerous considerations with uncertain implications when determining the appropriate amount of in-person learning to offer. Research has since revealed that students who spent more time learning in person during the reopening year performed better (Darling-Aduana et al., 2022; Fahle et al., 2023; Ross et al., 2024). Unfortunately, our results reveal a tension. We found that teachers who spent less time in person reported better perceptions of their working conditions, suggesting that teachers’ well-being and students’ academic performance were at odds. Regardless of how districts balanced these competing interests, it was the responsibility of leadership to communicate those decisions and rationales to their teachers.
How positively teachers assessed their district’s responses to COVID emerged as the factor with the largest associations with teachers’ perceptions of changes to their working conditions. The measure captured how positively teachers viewed their district’s response with a composite of communication effectiveness, decision appropriateness, resource provision, and comfort with returning post-closures. Our analysis revealed a clear pattern. Compared to teachers with less positive views of their district’s response, those with more positive views were consistently less likely to report worsened conditions and blame COVID, while being more likely to report improved conditions and to credit COVID for the improvements. Furthermore, the measure was also unique in that it was the only one where teachers’ responsiveness varied significantly across all the teacher and school characteristics in the moderation analysis. For instance, a teacher’s probability of blaming COVID was more responsive among female, White, and Hispanic teachers as well as those in schools with lower concentrations of economically disadvantaged and minoritized students. These findings underscore the critical importance of districts involving teachers in decision making, striving to meet teachers’ needs, clearly communicating decisions to teachers, and helping teachers become more comfortable with the decisions.
The rich measures of district policies that we used in this descriptive noncausal study enabled us to examine the potential role of said policies in teachers’ perceptions of working conditions and provide us with insight into teacher well-being. These measures, however, were not an exhaustive list of all decisions that could have played a role. For instance, we lacked measures that captured how districts facilitated ongoing communication with families and students, or the specific kinds of new professional growth opportunities districts provided teachers for how to teach during the pandemic. Additionally, because we could not interview district officials, we relied on districts’ written plans for their policy decisions, which may not perfectly match their policy implementation. Finally, we also lacked additional data on the policies we analyzed—namely, how teachers felt about each of these policies. The measure of teachers’ positive views of their district’s COVID response was the only measure that directly reflected teachers’ opinions of district policies. This may be why its relationship with their perceptions of how their working conditions changed was larger than the relationships for the measures that simply indicated the presence of a policy. Despite these limitations, our findings make important contributions by testing a potential pathway through which the pandemic may have influenced teacher outcomes.
Conclusion
District-level decisions shape teachers’ working conditions. Supportive working conditions can help teachers cope with job demands and maintain their well-being. Unsupportive working conditions, however, can reduce teacher effectiveness and increase turnover. Leveraging unique data from Virginia collected in spring 2021, we examined how teachers perceived that their working conditions had changed, including COVID’s role in that change, and then explored how those perceptions were connected to district pandemic policies and teachers’ outcomes. Notably, more in-person learning predicted perceptions of worsened conditions (and attributing this to COVID), and a positive view of how districts communicated policy changes and facilitated teachers’ safely returning to school predicted perceptions of improvements (and attributing this to COVID). We found that teachers who perceived their working conditions to have worsened were less satisfied and less likely to intend to remain at their current school than teachers who perceived their working conditions had improved. White teachers, as well as teachers at schools serving less disadvantaged student populations, were more likely than their peers to blame COVID for worsened working conditions and less likely to credit COVID for improvements. District COVID policies were more strongly related to teachers’ probability of blaming COVID for changes than their probability of crediting COVID. Finally, we found that the relationships between district policies and teacher perceptions varied across teacher and school characteristics.
Although this study occurred during the first full school year of the COVID pandemic, our findings are relevant to districts’ future policy decisions. Policymaking is a complicated process that does not occur in a vacuum. All policy decisions require tradeoffs and include risks of unintended consequences. Districts often must make decisions with incomplete or ambiguous information. They must weigh different factors and search for a solution that prioritizes the best interests of their students and teachers. The COVID pandemic was an instance, albeit extreme, of an external factor to which districts had to respond.
Our findings show that no district decision or policy will be viewed positively by all teachers; however, the processes that districts follow to make and implement decisions will play a role in how teachers respond. Effective communication is crucial to addressing divergent views in such scenarios. Similarly, whatever the decision, whatever the policy, leadership must help teachers feel safe and supported in their schools. In doing so, districts can help teachers cope with job stress and strengthen their commitment to their jobs and the district.
Supplemental Material
sj-docx-1-tcz-10.1177_01614681251322575 – Supplemental material for District Pandemic Policies and Teachers’ Perceptions of Their Working Conditions in Virginia
Supplemental material, sj-docx-1-tcz-10.1177_01614681251322575 for District Pandemic Policies and Teachers’ Perceptions of Their Working Conditions in Virginia by Luke C. Miller and Erica Sachs Langerhans in Teachers College Record
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research reported here was supported by the Institute of Education Sciences, U.S. Department of Education, through Grant R305S210009 to the Virginia Department of Education and Grants R305A200106 and R305B200005 to the University of Virginia. The opinions expressed are those of the authors and do not represent the views of the Institute, the U.S. Department of Education, or the Virginia Department of Education.
Supplemental Material
Supplemental material for this article is available online.
Notes
Author Biographies
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
