The Pandemic’s Effects on Children’s Education

School closures and switches to hybrid/virtual learning due to the pandemic adversely affected student achievement through several channels, including a decline in skill accumulation and a disruption of peer effects and peer-group formation.

Preliminary evidence suggests that losses took place early in the pandemic and that there has not been an apparent recovery. Also, the impact on students has been far from uniform, as economic losses tend to fall more deeply on younger students and students from disadvantaged backgrounds.

Simply returning schools and instructional practices to where they were prior to 2019 will not avoid such losses. A wide range of remediation policies has been suggested, and evidence suggests that instruction practices — such as tutoring and individualized/small group instruction — appear to be effective.

The pandemic certainly shifted ways of life drastically, and schooling was no exception. As many schools moved to online learning and hybrid schedules, there was bound to be an impact on educational outcomes. In this article, I examine research discussing these impacts and their potential lingering effects.

Declines in Learning

Learning progress slowed substantially in the U.S. during the pandemic. In the three decades before the pandemic, the mean achievement of U.S. eighth graders in math — measured by the National Assessment of Educational Progress (NAEP) — rose by more than 20 points or 0.5 standard deviations (SD). Between 2019 and 2022, math scores declined by 8 points, which means that students lost 40 percent of the previous rise, as shown in Figure 1.

In the U.S. and around the world, evidence shows that in terms of students' scores in standardized tests:

• Declines were observed across all grade levels.
• Learning deficits tended to be larger in math than in reading.
• Learning losses varied widely across states.
• Learning deficits originated early in the pandemic and still persist.

What does a decline of 8 points in the NAEP math test mean? In math, one SD of individual student scores is about 40 points and is roughly equivalent, as a rule of thumb, to three years of schooling. The national average loss of 8 points is equivalent to 0.2 SD, which implies 0.6 years of schooling lost.

The 2023 paper "A Systematic Review and Meta-Analysis of the Evidence on Learning During the COVID-19 Pandemic" looked at more than 40 studies around the world and reached similar conclusions. In the study, the overall size effect is a loss of 0.14 SD, or about 0.4 years of schooling lost.

Factors Explaining These Learning Losses

Several factors contributed to explaining learning losses since the beginning of the COVID-19 pandemic. Among the potential factors were changes to enrollment, instructional time and schooling mode.

A Decline in Enrollment

K-12 public school enrollment declined by about 1.2 million students from fall 2019 to fall 2021.¹ Part of this decline was due to a shift to home-school and private-school enrollment, but a large fraction of the loss is still unexplained.

The 2023 report "Where the Kids Went: Nonpublic Schooling and Demographic Change During the Pandemic Exodus From Public Schools" uses data from 21 states plus D.C. (covering the 2020-21 and 2021-22 school years) to explain what happened with students leaving public schools. The report shows that an increase in home-school enrollment and a decrease in school-age population each explain 26 percent of the decline in public-school enrollment, and 14 percent was explained by an increase in private-school enrollment.

About one-third of the decline, however, remains unexplained.² Tentative explanations for this additional share include an increase in absenteeism, growth in the prevalence of unregistered home schooling and a larger number of young students skipping kindergarten.

Instructional Time

The amount of effective instructional time declined during the COVID pandemic period due to school closures and the sudden switch to remote learning. My 2020 article "The Long-Term Effects of Educational Disruptions" — co-authored with John Bailey Jones — reviewed causal evidence from studies that examine how temporary and unexpected changes in instruction time impact student test scores. These studies look at school closures (involving factors ranging from weather closures to strikes and wars) as well as expected and permanent changes in instruction driven by education policy (including learning losses during summer breaks). The consensus is that increased instruction time has a positive impact on student test scores.

The impact of instructional time differs across age groups. Research has consistently found that instructional time has larger effects on younger students. The 2008 study "Performance Trajectories and Performance Gaps as Achievement Effect-Size Benchmarks for Educational Interventions," for example, shows that average annual gains in effect size from national standardized tests tend to be higher in lower elementary school grades and decline in the high school years. The 2011 "What a Difference a Day Makes: Estimating Daily Learning Gains During Kindergarten and First Grade Using a Natural Experiment" attempts to identify the average effect of formal schooling on achievement gains. The study finds that kindergarten reading scores increase by 1.6 SD during a standard 250-day school year.

It follows from this research that the COVID-19 pandemic shock and the associated schooling disruptions would have a heterogeneous negative impact on students of different learning ages, with younger children affected substantially more. However, these younger students would have more time to make up for some of the losses before completing school, as opposed to students in high school.

Schooling Mode

The schooling model — in-person, hybrid or virtual — also affects student achievement. The 2023 study "Pandemic Schooling Mode and Student Test Scores: Evidence From U.S. School Districts" estimates the effect of the schooling mode adopted by districts in 11 states on pass rates for students from third grade to eighth grade during the 2020-2021 school year. From 2019 to 2021, pass rates declined at all locations, with an average decline in math of 12.8 percentage points and 6.8 percentage points in English language arts (ELA).

However, districts with less in-person schooling experienced the largest declines. Exploiting within-state variation in schooling mode, the paper shows that districts that offered the most schooling in person saw substantially smaller drops in student test scores compared to the districts that offered the least. Students in districts in the lowest quartile of in-person schooling saw their average test scores in math and ELA fall 18.8 percent and 8.6 percent, respectively. On the other hand, students from districts in the quartile offering the most in-person schooling saw their average test scores in math and ELA fall 5.0 percent and 3.4 percent, respectively.

Offering a hybrid model did not have the same effect as the fully in-person alternative, but it reduced the losses in math and in ELA compared to the full virtual approach. The paper also shows that the positive effects of in-person learning were larger for students in lower grades.

In a different study, the working paper "The Consequences of Remote and Hybrid Instruction During the Pandemic" shows that relative to the pre-pandemic period (2017-2019) student achievement declined as the percentage of the 2020-21 school year spent in remote instruction increased.

Long-Term Outcomes

School disruptions and learning losses have a wide range of economic and social effects. Lower levels of learning due to fewer years of effective schooling translates into deficient development of cognitive skills (measured by scores in standardized tests). In turn, lower cognitive skills will likely reduce the future earnings potential and labor-market opportunities of the students affected. All of this could eventually translate into lower economic productivity for the nation as a whole.

Existing research can offer an approximate idea of the long-term implications of the educational disruptions. For instance, the 2023 paper "The Economic Cost of the Pandemic: State by State" assesses some of the possible economic costs by relying on estimates from studies on human capital and labor markets. In the U.S., the income return per SD of individual test scores is 0.274. This means that an 8-point loss in math scores (or, equivalently, a 0.2 SD decline in achievement) implies that the average student would have 5.6 percent lower lifetime earnings.

An alternative approach to measuring the potential negative long-term effects is based on the relationship between years of schooling and labor market income, known as the "Mincer equation."³ From the 2015 paper "Returns to Skills Around the World: Evidence From PIAAC," it follows that an additional year of schooling increases income by about 11 percent on average in the U.S. If it is assumed that the pandemic and the associated changes in schooling generated a loss of one-third of a school year, then this would translate into a loss of income for the affected students of about 3.5 percent over their entire working life. This would still hold even if schools return to their 2019 levels of performance.

Researchers also claim that fewer years of schooling have other negative implications, such as lower high school graduation rates and higher unemployment as adults. These additional adverse effects are not incorporated into the previous estimates.

The empirical literature referred to above relies mostly on reduced-form estimations and, so far, has been able to quantify only some of the potential channels through which the pandemic shock has affected educational outcomes. Recent work that relies on structural modeling — disciplined by current and pre-pandemic data — assesses the combined impact of different channels and potential long-run repercussions by accounting for a wide range of other effects.

The 2022 study "The Long-Term Distributional and Welfare Effects of COVID-19 School Closures," for instance, considers a human capital production function that includes both public investments by schools and time and monetary investments by parents. The study finds that a six-month school closure for children between 4 and 14 years old would increase the share without a high school degree by 7 percent and reduce the share of children with a college degree by 3.2 percent, with younger children experiencing the largest negative effects. Moreover, average lifetime earnings would decline by about 1 percent.

The 2021 working paper "Aggregate and Intergenerational Implications of School Closures: A Quantitative Assessment" also adopts a structural approach, which accounts for general-equilibrium effects and different degrees of substitutability between schools' and parents' investments in education. The paper shows that school closures that last for one year tend to reduce the lifetime earnings of the affected cohorts by about 1 percent. The model also reveals that general equilibrium effects may play an important role. The decline in the supply of human capital increases the return to education, which induces parents to increase educational investment on their children to make up for some of the learning losses.

Distributional Effects

The COVID-19 pandemic shock did not affect students equally in the U.S. The shock is expected to widen the gap in academic achievement between students from high-income and those from low-income families, typically referred to as the gap in educational achievement by socioeconomic status (SES) or simply as educational inequality.

This is relevant to the extent that educational inequality is associated with low social mobility across generations. Even before the COVID-19 pandemic, inequality in educational inputs — access to quality educational institutions (including child care schools) and parental inputs — was increasing. This trend seemed to have intensified during the pandemic: More-educated parents increased the amount of time spent on child care and monetary investments in education considerably more than less-educated parents.

Evidence indicates that the SES-achievement gap was relatively high in the U.S. The 2022 paper "Long-Run Trends in the U.S. SES-Achievement Gap," for instance, states that the SES-achievement gap between the top and bottom SES quartiles for the 1961 cohort is roughly 0.9 SD, or just under three years of learning. The gap has been declining but at a very slow pace: The gaps in math, reading and science achievement are closing by 0.05 SD per decade. At this pace, the gap would not be eliminated until the second half of the 22nd century.

Many recent studies present preliminary evidence indicating that the COVID-19 pandemic shock has increased educational inequality in outcomes. This was observed in both math and reading test scores, for both primary and secondary students, at every stage of the pandemic and regardless of how socioeconomic background is measured. Several factors may have contributed to the widening of educational inequality during this period, including:

• The direct impact of the switch from in-person to virtual schooling on student performance
• The decline in positive peer spillovers from school, which particularly affected children from low-income families
• The fact that parents in low-income families face tighter constraints in supporting their children's learning and compensating for the learning losses

School Closures and the Student Achievement Gap

When considering the impact of the instructional modes, it is important to consider that school closures did not take place uniformly across districts (with the incidence of school closures higher in high-poverty school districts) and that instructional modes affect students' achievement differently depending on their socioeconomic status.

The 2021 study "Large Socio-Economic, Geographic and Demographic Disparities Exist in Exposure to School Closures" shows that school closures occurring from September 2020 to December 2020 were more common in schools with lower third-grade math scores and with higher shares of students who are racial and ethnic minorities, who experience homelessness, who have limited English proficiency and who are eligible for free or reduced-price school lunches.

According to the 2023 paper "The Unequal Responses to Pandemic-Induced Schooling Shocks," students from disadvantaged sociodemographic groups were more likely to face school cancellations at the onset of the pandemic. They also had more limited access to remote learning resources and relied more heavily on schools to obtain access to these resources.

Finally, the aforementioned 2022 working paper "The Consequences of Remote and Hybrid Instruction During the Pandemic" decomposes the impact on the achievement gap from the differential incidence of school closures and the differential impact of the mode of instruction. It shows that 30 percent of the achievement gap between high and low-poverty schools is due to school closures and 50 percent is explained by the differential effect of remote/hybrid learning on poor and rich schools.

Instructional Time and Achievement

Adding instructional time can increase student achievement. The benefits, however, depend on how time is increased and which students receive the additional instruction.

Although the impact of instructional time on academic achievement has been well established, there are significant discrepancies at different points in the achievement distribution, both between students and across schools. Evidence seems to indicate that a uniform (non-targeted) rise in instruction time benefits high achievers more than low achievers, consistent with the idea that "skills beget skills."⁴

The Role of Social Interactions and Parental Resources

In terms of the peer environment, schools operate as a "social equalizer," in that they provide children from disadvantaged families the opportunity to socialize with children from families with higher educational attainment. The quantitative model developed by the 2022 paper "When the Great Equalizer Shuts Down: Schools, Peers and Parents in Pandemic Times" accounts for these effects. School closures, in the paper's framework, affect educational inequality through three channels:

• Learning technology: Schools are substituted for other, less conventional and newer learning technologies.
• Peer environment: Students interact less with other students from different socioeconomic backgrounds and more with peers residing in their neighborhoods.
• Parent demands: Remote learning places greater demands on parents in terms of both time and financial resources.

The paper finds that these channels jointly cause a skill loss of 0.4 SD for children from poor neighborhoods, with peer effects playing the largest role. In terms of grades, the skill loss entails an average loss for disadvantaged children comparable to a change from a B to a C in about half of school subjects.

After schools closed, the pattern of online searches for learning resources differed widely across U.S. households. The 2021 paper "Inequality in Household Adaptation to Schooling Shocks: COVID-Induced Online Learning Engagement in Real Time" shows that search intensity for both school- and parent-centered online learning resources nationwide had doubled by April 2020 relative to pre-pandemic levels. Larger increases in search intensity, however, occurred in areas of the country with higher income, better internet access and fewer rural schools.

The same pattern of online searches was observed worldwide. For instance, the 2020 paper "Inequalities in Children's Experiences of Home Learning During the COVID-19 Lockdown in England" shows that primary school students in the 10th percentile of family income distribution devoted about 35 minutes less learning per day than those from median-income families and 70 minutes less than children from families in the 90th percentile.

Other research also accounts for the role of parents in compensating for lower instructional quality. The aforementioned 2022 paper "The Long-Term Distributional and Welfare Effects of COVID-19 School Closures" and the 2023 paper "The Fiscal and Welfare Effects of Policy Responses to the COVID-19 School Closures" show that parental actions reduce the negative effects of school closures but do not fully offset them. Children with poorer parents tend to suffer more, mostly because wealthier parents can increase investment in their children in response to school closures. Also, elementary and private schools were closed, on average, for fewer days than secondary and public schools. As a result, younger children and children from richer families experienced lower welfare losses. Finally, earnings and welfare losses were largest for children who started public secondary schools at the onset of the crisis.

The aforementioned 2021 working paper "Aggregate and Intergenerational Implications of School Closures: A Quantitative Assessment" shows that, after accounting for general equilibrium effects, unexpected school closure shocks have moderate long-lasting adverse effects on macroeconomic aggregates. They do, however, generate lower intergenerational mobility. Substitutability between public and parental investments is key in explaining the outcomes. For instance, when substitutability is lower, the negative impact on the aggregate economy and overall incomes of the affected children would be higher, but the negative impact on intergenerational mobility will be smaller.

Teacher Effectiveness and Student Outcomes

Before the pandemic, studies found that teacher effectiveness had a critical impact on student achievement. The 1992 study "The Trade-Off Between Child Quantity and Quality" states that the best teachers provided 1.5 years of academic growth for students each school year, while the least effective teachers provided only 0.5 year's learning. The 2014 paper "Measuring the Impacts of Teachers II: Teacher Value-Added and Student Outcomes in Adulthood" examines the effectiveness of grade 3-8 teachers and their students' future income. The study shows that replacing a bottom 5 percent teacher with a teacher of average effectiveness would increase the present value of lifetime income for students by approximately $250,000. Moreover, the lasting impact of a good (or bad) teacher is later seen in college attendance and labor market earnings.⁵

Instructional Practices and Student Achievement

Teachers adopt a wide range of approaches in classroom instruction, including:

• Traditional direct instruction, such as lecturing and use of textbooks
• More class time devoted to students working with classmates
• Individual practice and assessment

Some of these approaches are more effective in promoting student learning than others, independent of the teacher's skills/effort.

The 2023 paper "Teachers' Use of Class Time and Student Achievement" provides evidence that students tend to score higher on math exams when their teachers devote more class time to individual practice and assessment. In a math class where the teacher allocates "most of the time" to practice and assessment, scores are 0.08 SD higher than in the typical class. For English exams, students score higher when they spend more time working and talking with classmates, with similar predicted gains in magnitude as those in math.

As mentioned earlier, the 2016 paper "What Differences a Day Can Make: Quantile Regression Estimates of the Distribution of Daily Learning Gains" and the 2017 paper "The More, the Better? The Impact of Instructional Time on Student Performance" show that additional instructional time is more beneficial to higher achievers, contributing to the growth and persistence of the achievement gap. Finally, the 2010 paper "Teacher Effects and the Achievement Gap" states that disadvantaged students are more likely to get the weakest teachers who spend less time on instruction.

It should be acknowledged, however, that the literature described above only covers a sample of a much larger related research that focuses on the improvement of instructional practices. This work includes (among other areas of study) research on the use of phonics versus whole language in reading, competing math curricula, the value of homework, the impact of class size, the use of technology in the classroom and the value of bilingual education.

Other Effects of COVID and School Closures on Students

School closures can also be expected to have consequences for the socioemotional and motivational development of children and adolescents. The psychological strain on families generated by the COVID-19 shock affected students' development in these areas and may adversely affect economic potential. Empirical evidence has linked school closures to several factors, including rising mental health concerns, lower levels of engagement, reports of violence against children, rising obesity, increases in teenage pregnancy, rising levels of chronic absenteeism and dropouts, and overall deficits in the development of socioemotional skills due to social isolation from networks and peers.

Other labor market outcomes are also closely associated with educational attainment. Students who do not read well in elementary school are more likely to not graduate high school on time or to drop out of high school altogether. Fewer years of schooling systematically lead to higher unemployment, as noted in the 2020 working paper "The Economic Impacts of Learning Losses." Conversely, higher educational attainment reduces dropout rates, increases the likelihood of attending college and lowers teen motherhood, incarceration and arrest rates, as found by the 2022 working paper "What Do Changes in State Test Scores Imply for Later Life Outcomes?"

The 2023 working paper "Societal Disruptions and Child Mental Health: Evidence From ADHD Diagnosis During the COVID-19 Pandemic" shows a large decline in cumulative new ADHD diagnoses from March of 2020 until early 2021. These "missed diagnoses" could be extremely costly both at the individual level and to society as a whole.

The 2022 working paper "In-Person Schooling and Youth Suicide: Evidence From School Calendars and Pandemic School Closures" examines the impact of in-person schooling on youth suicide. Historically, suicides among 12-18-year-olds were highest during months of the school year and lowest during summer months. This pattern, however, changed in 2020, as teen suicides declined in March 2020, remained low throughout the summer and increased in the fall of 2020 when many K-12 schools returned to in-person instruction. Returning from online to in-person schooling was associated with a 12-to-18 percent increase in teen suicides. The paper suggests that bullying victimization may have been important factor explaining this phenomenon.

Remediation Policies

What kind of policy interventions could mitigate the instructional and learning losses? Simply returning schools to where they were in 2019 would not allow students to make up for the lost ground. The pace of student learning in most districts would need to accelerate beyond where it was pre-pandemic for students to catch up academically. The work cited in this article indicates that it would be necessary to increase both instruction quality (teacher effectiveness and appropriate use of class time) and quantity (hours of instruction). Moreover, research that has analyzed the impact of unexpected changes in instruction time on educational attainment suggests that the cohort of affected students are permanently scarred.

Research can once more offer some guidance on which policies could be effective at addressing the educational challenges. For instance, as mentioned earlier, additional hours of instruction could help make up some of the losses. However, should the extension take place by adding days or by lengthening the school day? The positive effect of expanded learning time on student achievement appears strongest for extending the school year.

Tutoring Programs

The 2022 working paper "The Challenges of Implementing Academic COVID Recovery Interventions: Evidence From the Road to Recovery Project" claims that making up the gap would require approximately 40 to 100 hours of high-quality tutoring for the average student (slightly less for reading than math). This approach, however, faces several implementation challenges, such as:

• Reaching targeted students
• Staffing and scheduling
• Family engagement
• Building central office capacity and internal systems

Tutoring programs yield consistent and substantial positive effects on learning outcomes, according to the 2020 working paper "The Impressive Effects of Tutoring on PreK-12 Learning: A Systematic Review and Meta-Analysis of the Experimental Evidence," with an overall pooled effect-size estimate of 0.37 SD. The effects tend to be strongest among the earlier grades, with similar overall effects for reading and math. Moreover, tutoring programs conducted during school have larger impacts than those conducted after school.

The implementation of effective tutoring programs — especially at a larger scale — could be challenging and costly. It relies on, among other things, the ability of administrators to schedule additional class time and to attract and retain talented tutors who can consistently deliver high-quality instruction.

The 2023 paper "Not Too Late: Improving Academic Outcomes Among Adolescents" examines the impact of individualized, intensive, in-school tutoring on adolescents and finds that this instructional approach increased math test scores by 0.37 SD in Chicago and that the effects are persistent.

Instructional Time

Given that the impact of instructional time varies across the achievement distribution, policies that increase instructional time could either increase or decrease achievement gaps. Since the disruption likely increased dispersion in learning levels within classrooms, it would be necessary to adopt more individualized or small-group instruction — as noted in the aforementioned 2020 working paper "The Economic Impacts of Learning Losses" — to close the gap.

The 2023 paper "The Fiscal and Welfare Effects of Policy Responses to the COVID-19 School Closures" shows that extending schools by three months (or six weeks over the next two summers) generates significant welfare gains for the children and raises future taxes to pay for the cost of this schooling expansion. Expanded schooling generates substantial welfare gains of 0.21 percent in terms of consumption equivalent variation. Also, welfare gains are highest for children from income-poor households.

There is plenty of evidence supporting the positive effects of childhood investments on postsecondary attainment and degree completion. Examples include the 2011 paper "How Does Your Kindergarten Classroom Affect Your Earnings? Evidence From Project Star" and the 2013 paper "Experimental Evidence on the Effect of Childhood Investments on Postsecondary Attainment and Degree Completion."

The evidence presented by the aforementioned 2023 paper "Pandemic Schooling Mode and Student Test Scores: Evidence From U.S. School Districts" highlighted earlier informs, among other things, on which districts and students should be targeted and offers a note of caution when considering school closures in the future.

COVID-related learning losses were undoubtedly large and unequally distributed across households. Parents' ability to compensate for school inputs was crucial during this time, but it may also have contributed to increasing educational inequality. However, there is hope that carefully thought-out, targeted instructional practices might be able to make up for some of these losses.

Santiago Pinto is a senior economist and policy advisor in the Research Department at the Federal Reserve Bank of Richmond.

To cite this Economic Brief, please use the following format: Pinto, Santiago. (August 2023) "The Pandemic's Effects on Children's Education." Federal Reserve Bank of Richmond Economic Brief, No. 23-29.

This article may be photocopied or reprinted in its entirety. Please credit the authors, source, and the Federal Reserve Bank of Richmond and include the italicized statement below.

Views expressed in this article are those of the authors and not necessarily those of the Federal Reserve Bank of Richmond or the Federal Reserve System.