Educational Research for COVID-19 Response


These resources highlight research syntheses and recommendations about the effectiveness of potential actions for responding to the COVID-19 pandemic. The focus is on sources which review the broader literature rather than individual studies; however, in the section characterizing current conditions, there are some working papers which have not yet been peer-reviewed. Note that this does not summarize what various educational agencies are currently doing, since that is both fast-changing and vast in scope, and since it is not always clear what the research basis may be that underpins those decisions.

Major sources

Characterizing current conditions

Enrollment, attendance, and engagement

  • Missing in the Margins 2021: Revisiting the COVID-19 Attendance Crisis. (Bellwether; 2021 Oct)
    • In public PK-12 schools, enrollment decreased by 2.7% (>1.3m students) from 2018-19 to 2020-21, with most of the drops in elementary grades, especially kindergarten. Although reporting on attendance data from 2020-21 is sparse, available data point to overall decreases in attendance, with more absenteeism among English learners, low-income students, Black or Hispanic students, students with disabilities, and younger students. Engagement data suggest decreases in participation in online coursework and time spent on learning activities.
  • The Revealed Preferences for School Reopening: Evidence from Public-School Disenrollment. (CEPA; 2021 Aug)
    • While U.S. K-12 public school enrollment typically increases by a fraction of a percentage point each year, it fell by 2% in Fall 2020 (a loss of ~1.1 mil students). Compared to in-person instruction, offering remote-only instruction as of Aug / Sep 2020 reduced enrollment by 1.1% overall and by 3-4% in kindergarten. These results suggest long-term negative fiscal consequences if disenrolled students do not return.
  • Student Attendance and Enrollment Loss in 2020–21. (AIR; 2021 June)
    • According to a survey completed by a representative sample of 565 districts in April 2021, enrollment and attendance rates dropped in 2020-2021, with both rates being higher in districts providing primarily in-person instruction. Enrollment was greater in higher-achieving districts, while attendance was higher in low-poverty districts and districts serving mostly White students. Districts providing remote instruction monitored attendance using teacher reports (94%), learning management systems (89%), and/or homework completion (84%).

Measuring unfinished learning

  • A systematic review and meta-analysis of the evidence on learning during the COVID-19 pandemic (Nature Human Behavior, 2023 Jan)
    • A meta-analysis of 42 studies published (including preprints) up through August 2022 across 15 countries found that students lost out on about 35% of a school year’s worth of learning. Analyses suggest that learning deficits opened up early in the pandemic and have neither closed nor substantially widened, as of two and a half years into the pandemic. Socioeconomic inequality in education increased, and learning deficits are larger in math than in reading. There was no evidence of variation by grade level, but this may be confounded with longer school closures for older students in some countries.
  • Local Achievement Impacts of the Pandemic (Education Recovery Scorecard, 2022 Oct)
    • These analyses compared state assessment results from grades 4 and 8 across public school districts in 29 states during Spring 2019 and Spring 2022 by converting them to a common scale based on the National Assessment of Educational Progress (NAEP). The median school district experienced achievement declines corresponding to 0.52 grade equivalents in math and 0.23 grade equivalents in reading. For high-poverty districts (>69% FRL), these declines were 0.66 grade levels in math and 0.31 grade levels in reading; low-poverty districts (<39% FRL) saw declines of 0.45 grade levels in math and 0.25 grade levels in reading.
    • While declines were greater in districts which spent more time in remote instruction in 2020-2021, this does not appear to be the primary factor driving achievement declines. Some districts and states (such as California) had longer school closures but smaller achievement losses, while other districts without any remote instruction experienced substantial declines.
    • In many districts, the share of the budget corresponding to lost achievement exceeds the share received through federal ARP ESSER funding..
  • National Assessment of Educational Progress (NAEP) results (administered Jan-Mar 2022; National Center for Education Statistics, results released 2022 Oct)
    • Nationwide proficiency rates on the 2022 NAEP Reading Assessment showed a significant decline from 2019 to 2022 by three points in both grades 4 and 8. California’s decline was not statistically significant.
    • Nationwide proficiency rates on the 2022 NAEP Mathematics Assessment declined from 2019 to 2022 by five points in grade 4 and by eight points in grade 8. California’s proficiency rates declined by four points in grade 4 and by six points in grade 8.
    • Nationwide decreases in 8th-grade reading and math were similar across achievement levels, whereas gaps widened more (by ~5 pts. between the 10th and 90th percentiles) in 4th grade reading and math.
  • The Consequences of Remote and Hybrid Instruction During the Pandemic (CEPR, 2022)
    • Analyzing test data from 2.1m students in ~10,000 schools revealed that districts offering primarily remote or hybrid instruction in 2020-2021 experienced widening socioeconomic and racial/ethnic gaps in learning during the pandemic. In high-poverty schools (>75% FRL), math growth declined by 0.46 SD from 2017-2019 to 2019-2021, compared to 0.30 SD in low-poverty schools (<25% FRL). While white students’ growth declined by 0.208 SD, Black students’ growth declined by an additional 0.119 SD and Hispanic students’ by an additional 0.092 SD. However, accounting for school-level factors reduced these gaps to 0.036 SD for Black students and to 0.032 SD for Hispanic students. This implies that districts should target the hardest-hit schools for additional support, rather than student subgroups.
    • Because these assessments (NWEA MAP) are administered in fall, winter, and spring, declines in achievement could be converted to weeks of instruction or percent of the academic year. Applying this percentage to district budgets revealed that the average district’s planned expenditures from the American Rescue Plan act on academic recovery (28% of total budget) is likely not enough to make up for the amount of unfinished learning (close to 40% of academic year) for students in high-poverty districts which stayed mostly remote in 2020-2021.
  • Learning during COVID-19: An update on student achievement and growth at the start of the 2021–22 school year. (NWEA, 2021 Dec)
    • Compared to a typical year, math performance in Fall 2021 declined by 9 to 11 percentile points, while reading performance decreased by 3 to 7 percentile points. Overall, math declines were similar across grades 3-8, while reading declines were greater in younger grades. While math decreased by 8-10 points across racial/ethnic groups in grades 7-8, gaps widened in younger grades (e.g., for 3rd grade: Asian: -3, White: -7, Latinx: -13, Black: -14). High-poverty schools showed greater declines in younger grades for math (-14 to -8) and reading (-11 to -4), while low-poverty schools showed similar declines across grade levels in reading (between -2 and -4) but larger declines in math for middle grades (ranging from -4 in 3rd grade to -11 in 8th grade). Students with lower pre-pandemic achievement also showed lower gains, particularly in middle-grade reading and in elementary-grade math.
  • Pandemic Schooling Mode and Student Test Scores: Evidence from US States. (NBER; 2021 Nov)
    • State standardized test score data from Spring 2021 across 12 states show overall declines in pass rates compared to prior years (by 14.2% in math and by 6.3% in English). The effect of shifting from 0% to 100% in-person learning over the school year would reduce those drops by 10.1% in math and by 3.7% in English.
  • Changing Patterns of Growth in Oral Reading Fluency During the COVID-19 Pandemic. (PACE; 2021 Mar)
    • Students’ oral reading fluency (gr2-3) is approximately 30 percent behind expectations. Students at lower-achieving schools are falling farther behind, with 10% of students not assessed this fall.
  • Learning during COVID-19: Initial findings on students' reading and math achievement and growth  (NWEA; 2020 Nov)
    • Students (gr3-8) performed similarly in reading but lower in math. Decreases were greater for younger grades and for Hispanic and Black students. Overall, testing rates were 10% lower than in past years. 
  • Fall 2019 to Fall 2020 MAP Growth attrition analysis (NWEA; 2020 Nov)
    • Students who were assessed in 2019 but not in 2020 were more likely to be racial minorities (Black, Hispanic, Asian); among the lowest 30% of scorers in 2019; or attending a district with higher percentages of students who were racial minorities, socioeconomically disadvantaged, classified as English learners, or receiving special education services. The impact of the pandemic is thus likely to be underestimated in analyses based on these assessments.


  • School District Staffing Challenges in a Rapidly Recovering Economy. (CEDR; 2021 fall).
    • ​​​​​​​An analysis of school district job postings across Washington state in October 2021 revealed the most openings for paraeducators, about twice as many as for teachers. Postings per student are especially high for paraeducators in high-poverty districts. Among teachers, the most openings were for substitutes, then special education teachers. Vacancy rates, which adjust for student enrollment and the number of teacher candidates credentialed by the state in a given category, reveal far higher needs in high-poverty districts, especially for bilingual and special education teachers.
  • Stress Topped the Reasons Why Public School Teachers Quit, Even Before COVID-19. (RAND; 2021 Feb).
    • According to a nationwide survey of former public school teachers in December 2020, nearly half of those who left after March 2020 and before retirement cited doing so because of the pandemic. Compared to pre-pandemic, Black teachers were overrepresented among teacher leavers during the pandemic, while teachers with less experience (≤3 years) were underrepresented. Stress was the most common reason for leaving early, almost twice as common as insufficient pay. More flexibility was the most common attribute attracting teacher-leavers to their new job.

District operations

  • Urban and Rural Districts Showed a Strong Divide During the COVID-19 Pandemic: Results from the Second American School District Panel Survey. (RAND; 2021)
    • A survey completed by 434 district leaders revealed that in February 2021, only 17% of urban districts were offering fully in-person instruction, compared to 42% of rural districts. Most districts offering any remote instruction defined attendance as a once-per-day check-in. Half of districts shortened school time, especially those offering remote instruction. The majority of districts offered tutoring and increased socioemotional supports. Districts spent the same or more in 2020-2021 than in 2019-2020, but the majority cut certain expenses (such as building, operations, maintenance, transportation, and supplies) to cover pandemic-related costs.

Recommended practices and strategies

General guidance


  • How Schools Can Help Children Recover from COVID School Closures: A Letter from Education Researchers. (2020 Aug)
    • Provide substantial additional resources to prevent looming school budget cuts.  Implement universal internet and computer access. Target resources to those most in need. Provide the most personalized and engaging instruction possible under the circumstances, even when it is necessary to be online.  Address the learning losses created by the crisis by expanding instructional time in ways that challenge, support, and engage students. Offer tailored, integrated support to each child in order to address social-emotional, physical health, and family well-being. Make decisions about teachers that support pedagogical quality and equity.
  • Reducing district budgets responsibly. (ERR; 2020 Jul)
    • School resources are particularly important for supporting lower-income students and addressing between-school inequities. Examine tradeoffs if reducing extracurricular activities, support services, or after-school programs, which may affect student engagement, academic performance, and access to peer and adult support. 
    • Since layoffs have negative consequences for students, consider delaying pay raises or furloughing non-working staff to reduce layoffs. Avoid “last-in-first-out” layoffs, which disproportionately affect schools with underserved students, focusing instead on effectiveness. 
    • Be strategic about what to adopt (e.g., by diversifying and strengthening the workforce) and what to abandon (e.g., by dropping redundant or ineffective programs).

Systems and operations

Addressing unfinished learning

Assessment and differentiation



Student engagement, socioemotional learning, and school climate

  • Supporting Child and Student Social, Emotional, Behavioral, and Mental Health Needs. (US DOE; 2021 Oct)
    • Prioritize wellness. Enhance mental health literacy and reduce barriers to access. Implement continuum of evidence-based prevention practices. Establish integrated framework of support for all. Leverage policy and funding. Use data for decision-making to promote equitable implementation.
  • Evidence-Based Practices for Assessing Students' Social and Emotional Well-Being. (ERR; 2021 Feb)
    • Consider: Comprehensive system with validated instruments for monitoring student well-being, supplemented with student surveys for understanding mindsets and other unobservable characteristics, as part of a larger student support strategy. Screeners and monitoring strategies can help identify and refer students for targeted support by specialists or connect them to resources.
    • Avoid: Survey questions which may re-traumatize, stigmatize, or marginalize students. Poor survey design (e.g., double-barreled questions, double negatives, questions that are difficult for younger students to interpret).
  • Research-Based Strategies for Effective Remote Learning: Student Engagement. (REL Appalachia; 2020 Dec)
  • Ways to Promote Children's Resilience to the COVID-19 Pandemic. (ChildTrends; 2020 Apr)
    • Protective factors: Sensitive, responsive caregiving. Meeting basic needs. Emotional support for children. Support for caregiver well-being. Social connectedness.  
  • Bringing evidence-based decision-making to school safety. (ERR; 2020 Sep)
    • Consider: School-wide mental health trauma programs built around tiered interventions. SEL lessons incorporated into academic classes; small-group counseling; individual check-ins with caring adults. Mechanisms for student input on school rules and classroom processes. Comprehensive, disaggregated, validated school climate data. Train staff to recognize and respond to cues without stereotyping students. Connect with community mental health professionals.

Focal populations

Digital learning

Family engagement / At-home supports

Staff supports

  • Digital Professional Learning for K-12 Teachers: Literature Review and Analysis (WestEd; 2020 Dec)
    • Well-designed virtual communities of practice support implementation. Customization (software features, personalized goal-setting, modes for content delivery and assessment) can encourage more active engagement. Learning may be extended through asynchronous access to archived resources, collaboration with colleagues, continuous improvement cycles, and monitoring student progress. Strong and seamless facilitation helps; digital access opens up possibilities for virtual coaches and outside experts, although local facilitators can offer immediate feedback and contextual knowledge. Layer multiple modalities, with job-embedded learning opportunities and opportunities for social interaction.
  • Remote Professional Development: Rapid Evidence Assessment. (2020 Sep)
    • 1-pg summary. Remote coaching, mentoring, and expert support can be effective alone or to complement PD programs. Video is particularly effective for enabling teachers to reflect on teaching practice, if paired with other resources (e.g., viewing guides, coaching conversations). More interactive content increases time on task and completion rates. Collegial collaboration may improve outcomes through reflective practice and collective problem-solving. Supportive school conditions include leader support, protected time, and effective technology platforms and training.
  • District systems to support principal leadership. (ERR; 2020 Sep)
    • Principal Supervisors: Dedicate time to instructional leadership growth. Coach with focus on teaching and learning, even amidst operational demands.
    • Teaching and Learning: Support use of common, research-based definition of high-quality, culturally responsive teaching. Help foster teacher learning communities with necessary autonomy and resources.
    • Human Resources: Recruit and select teachers based mainly on performance. Emphasize recruiting and retaining teachers of color. Partner with principals to match teachers to roles and teams.
    • Data Systems: Provide principals with ready access to information about their students and staff, with strengths-based, antiracist approach.
    • Operational Staff: Partner with principals to provide high-quality services that ensure that facilities, transportation, and food services support high-quality instruction.
  • Sustaining teacher training in a shifting environment. (ERR; 2020 Jul)

This page was last updated on September 12, 2023