How Did the COVID-19 Pandemic Reshape Teacher Attrition and Mobility?

This study investigates how teacher attrition and mobility in Michigan shifted before and during the COVID-19 pandemic. Using administrative data on nearly 125,000 public and charter school teachers from 2015 to 2023, the researchers applied an interrupted time series framework. They found that teacher attrition increased by an average of 1.1 percentage points annually after the pandemic began, while within-district moves remained stable and between-district moves declined. Teachers in districts offering fully in-person instruction during 2020–21 were more likely to leave. Contrary to concerns, teachers in charter schools and those serving higher shares of disadvantaged students were less likely to attrit.

Full Citation and Link to Article

Rogers, S., Powell, T., Strunk, K. O., & Hopkins, B. G. (2024). Viral Change: Trends in Michigan Teacher Attrition and Mobility before and during the COVID‑19 Pandemic. Education Finance and Policy. https://doi.org/10.1162/edfp.a.18 

Extended Summary

Central Research Question

This study examines how the COVID-19 pandemic affected teacher attrition and mobility in Michigan’s public schools. The researchers pose three main questions: (1) How did teacher exits from the profession, district transfers, and within-district school changes shift during and after the pandemic? (2) Were these trends different for teachers in districts or communities disproportionately affected by COVID-19, particularly in relation to instructional modality (in-person, hybrid, remote)? (3) Did these patterns vary across teachers with different demographic and credential characteristics, school settings, and district types?

Previous Literature

Teacher attrition and mobility have long been challenges in the U.S., particularly in hard-to-staff schools and subjects such as special education, math, and science. Prior to the pandemic, Michigan experienced rising attrition rates and declining teacher preparation program enrollment. Nationally, teacher turnover rates ranged between 5.1% and 8.4%, with higher turnover observed among novice, non-White, and urban/rural teachers.

The pandemic introduced new factors that could reshape these trends. Surveys in Michigan during 2020 revealed that large percentages of teachers considered leaving the profession due to health and safety concerns. At the same time, instructional modality—whether districts offered remote, hybrid, or in-person instruction—emerged as a key factor that could influence teacher decisions. While early research from states such as Washington, North Carolina, Massachusetts, and Arkansas showed varying degrees of pandemic-related attrition, the existing literature had limited insight into longer-term effects and subgroup-specific trends.

This study contributes to that gap by examining multiple post-pandemic years (through 2022–23), disaggregating different types of mobility, and analyzing variation across demographic, instructional, and organizational dimensions.

Data

The authors use administrative, employee-level data on Michigan public and charter school teachers from fall 2015 through fall 2023. The dataset includes demographics (gender, race/ethnicity, experience), credentials (certification type, subject area), and assignment information (school, district, grade level). In total, the sample covers 123,787 teachers across 3,574 schools.

The researchers also incorporate contextual data:

• County-level COVID-19 case rates (from the Michigan Department of Health and Human Services)

• District-level instructional modality for each month of the 2020–21 school year (via the Education Policy Innovation Collaborative)

• School-level student characteristics, such as race/ethnicity, economic disadvantage, English learner, and special education status

• District characteristics including urbanicity and charter status

The outcome variables track three types of teacher movement: exiting the Michigan public education system, switching school districts, and switching schools within a district.

Methods

The study employs an Interrupted Time Series (ITS) design to estimate changes in teacher attrition and mobility before and after the onset of the pandemic. Linear probability models are used to compare outcomes across years, accounting for teacher, school, and district covariates.

Three primary models are estimated:

1. Baseline ITS examining overall trends in attrition and mobility from 2015–16 through 2022–23.

2. Models interacting post-pandemic years with COVID-19 community rates and district instructional modality.

3. Models exploring heterogeneity in trends by teacher demographics, subject area, school population characteristics, district urbanicity, and charter status.

Instructional modality is a key explanatory variable, categorized as “always in-person,” “sometimes in-person,” or “never in-person” during the 2020–21 school year. COVID-19 rates are divided into terciles (low, medium, high) based on county-level data.

The models disaggregate attrition (leaving the public teaching workforce) from between-district moves and within-district school switches, recognizing that different stakeholders (e.g., principals, superintendents, state policymakers) may be concerned with different types of turnover.

Findings/Size Effects

The pandemic had a significant but nuanced effect on teacher labor market dynamics in Michigan.

Overall Attrition and Mobility Trends

• Teacher attrition declined slightly after 2019–20 (–0.49 percentage points) but increased notably after 2020–21 (+1.01 points) and 2021–22 (+1.81 points), representing 13.8% and 23.5% increases over pre-pandemic attrition, respectively.

• By 2022–23, attrition remained 1.3 percentage points above trend but had started to decline, suggesting possible normalization.

• Across-district moves dropped sharply after 2019–20 (–2.43 points; a 46% decline from the previous year) and remained below pre-pandemic trends.

• Within-district school switches were steadily declining pre-pandemic and did not change significantly during the pandemic.

Instructional Modality

Instructional modality was a stronger predictor of attrition and mobility than community COVID-19 rates:

• Teachers in “always in-person” districts during 2020–21 were more likely to leave teaching or switch districts.

• Teachers in “never in-person” districts were significantly less likely to leave or move, particularly in 2019–20 and 2020–21.

• These findings suggest that policy decisions about reopening schools played a more prominent role in shaping teacher behavior than local COVID case rates themselves.

Demographic Variation

• Black and Latino teachers were less likely than White teachers to leave or switch schools/districts after 2019–20 and 2020–21.

• Asian and Black teachers also had lower attrition rates in some years, reducing concerns that the pandemic worsened teacher workforce diversity.

• Novice teachers (0–3 years) were less likely to leave in 2019–20 but more likely to attrit in later years. They were significantly less likely to switch districts throughout the pandemic.

School and District Characteristics

• Teachers in schools with higher proportions of economically disadvantaged and non-White students were less likely to leave the profession after 2019–20 and 2020–21.

• These teachers were also less likely to switch districts.

• Teachers in charter schools and urban areas exhibited different patterns, but school setting variables such as instructional modality and student demographics were more predictive of attrition than teacher demographics themselves.

Subject Area Trends

• The study also explored subject-specific attrition (e.g., special education, ESL, STEM), though detailed results were more limited. Some variation by subject area emerged, particularly among shortage areas, but general trends mirrored the broader findings.

Comparison with Other States

Compared to Arkansas, Massachusetts, North Carolina, and Washington, Michigan experienced a larger decline in district switching immediately after the pandemic but a more modest rebound. Other states saw spikes in turnover in 2021–22, whereas Michigan’s rebound was slower and less pronounced. This may reflect differences in policy context, school reopening timelines, or workforce demographics.

Conclusion

The COVID-19 pandemic had substantial but heterogeneous effects on teacher attrition and mobility in Michigan. While attrition increased—resulting in an estimated average annual loss of 1.1 percentage points more teachers—within-district moves remained stable and between-district moves declined. Teachers in districts offering fully in-person instruction in 2020–21 were more likely to exit or move, suggesting that instructional modality had a more significant impact on teacher behavior than the actual COVID rates in a community.

Importantly, teachers in higher-need schools (serving more non-White and economically disadvantaged students) and those in charter settings were generally less likely to leave, alleviating some concerns about widening inequities. In fact, the school setting—particularly instructional modality and student composition—was a stronger predictor of attrition than teacher demographics.

The authors conclude that decisions made by districts during the pandemic, especially those related to instructional modality, had lasting impacts on teacher stability. While some effects may be reverting toward pre-pandemic norms, the pandemic created both challenges and opportunities for understanding and addressing teacher workforce dynamics.

These findings have important implications for policymakers. Tailored retention strategies may be warranted for specific school settings or instructional contexts, especially in future crises. Moreover, recognizing the trade-offs associated with decisions about school operations could help mitigate the negative impacts of policy responses on the educator workforce.