All children deserve to attend welcoming and well-funded schools where they can learn and grow, regardless of race, disability, or income. But funding for public schools, where nearly 90% of all U.S. students learn, is at a near crisis point. The Trump administration’s goals, which are taken right out of Project 2025, seem to be to defund public education to the point that it doesn’t work, then offer private school vouchers as a solution to a manufactured problem. In this post, we highlight five ways public education is on fire in the United States and the damage this will do to students’ abilities to learn and thrive. Instead of cutting funds, lawmakers should invest in public schools, one of the best tools we still have to build a prosperous, equitable country.

Alarm level 1: COVID-19 relief funding for public schools is winding down. In some cases, the administration is ending it prematurely

This academic year (2024–2025) marks the end of the financial support schools were receiving to address the impacts of the COVID-19 crisis, the Elementary and Secondary Schools Emergency Relief III funds (ESSER III). The COVID-19 pandemic, and the changing learning environments that ensued, meant that schools needed funds to address the significant academic, social, emotional, physical, and mental health needs of their students. This funding was distributed in recent years with the last distribution, ESSER III, worth a total of $122 billion allocated to districts around the country. Many students, especially those living in poverty, have not recovered from pandemic-related learning loss. The end of this funding means that districts will now have fewer resources to help students get back on track. Rigorous research has demonstrated that this federal aid to public schools was highly successful, with measurable improvements to student outcomes in states and districts where more aid was spent. Taking the educational challenges imposed by the pandemic seriously would mean recognizing the high value this aid has provided.

However, in late March, the Trump administration canceled extensions that had been granted to states to spend remaining ESSER funds. Effectively, districts are losing out on the funding allocated to them in the form of COVID-19 relief funds. Canceled extensions represent almost $3 billion in lost funding that had already been committed to tutoring services, reading interventions, building improvements, and more. Clawing back these funds jeopardizes improved academic outcomes for many students and their ability to learn in healthy and safe environments. The administration’s refusal to reimburse school districts for funding that has already been spent could force them to cut teaching and other staff positions to make up the cost, ultimately harming students.

Alarm level 2: The administration is lawlessly dismantling the Department of Education and attacking inclusive schools

The winding down and clawing back of ESSER funding are simultaneously occurring at a time when President Trump signed executive orders to (1) dismantle the Department of Education and “return the funding to the states” and (2) regulate curriculum taught in the more than 13,000 public schools in the country.

One order directed Secretary of Education Linda McMahon to shut down several functions of the Department of Education (ED) and send them back to the states. Prior to this order, the White House had directed the ED to lay off 1,300 employees, a directive that is currently in litigation. The other order resulted in a “Dear Colleague letter” from Secretary McMahon demanding that states certify that they will not engage in “illegal DEI practices” as a condition of receiving the federal funds (This order is also currently in litigation.). As it stands, much of the Department of Education funding goes directly to state and local school systems. The ED provides targeted funding to public schools for special education through the Individuals with Disabilities Education Act and supports high-poverty districts through Title I grants. These grants make up for shortfalls in funding that high-poverty districts experience when they get funding from local sources.

To be clear, closing the Department of Education, and reappropriating major funding programs requires an act of Congress, and it is local school districts who have control over what is taught in schools—not federal regulators. Thus, while these executive orders have the potential to inflict a lot of damage, it’s unclear whether these orders can proceed without running afoul of federal laws. If these orders result in delays in funding distributions or outright cuts, students could experience declines in academic achievement, exacerbating existing racial and income disparities and limiting students’ long-term opportunities. If President Trump acts outside of his authority to slash the agencies’ work, the guardrails will essentially be pulled off this funding, which is extremely effective at redistributing funds based on district need. President Trump says he’ll return money to states for them to distribute it, potentially creating a situation where states have to compete for funds. This would create a patchwork in public funding for public schools, one in which some districts risk falling even further behind.

Alarm level 3: Lawmakers are pushing a mounting wave of voucher programs, an increasingly large cost to state-funded education

While many school districts struggle to maintain basic education funding, school privatization efforts are continuing throughout the country, and states like Arizona, Florida, and Ohio are notorious for the budget-breaking cost of universal voucher programs.

Figure A shows the current cost of voucher programs as a share of K–12 education funding in states where over 5% of the budget is currently going to school voucher programs. In the current school year (fiscal year 2025), voucher costs make up anywhere from 5% for states with early voucher programs to upwards of 25% of the entire public education budget for states with mature programs.

Figure A

Because statewide private school voucher programs are funded with state dollars, voucher spending is shown as a proportion of state education funding rather than state and local funding. On average, about 46% of funding for K–12 schools comes from state revenue sources. In states with voucher programs, private schools divert state dollars that could otherwise be available to public schools. For now, local funding for public schools is protected from diversion to voucher programs, although some states with voucher programs are also threatening this source of public school funding by cutting or eliminating property taxes.

Vouchers degrade the quality of education for students who use them

Time and time again research has shown that vouchers harm academic outcomes. Causal studies across three states and Washington, D.C., demonstrate negative effects on test scores for students who use a voucher to switch from public to private school. These test score declines can persist over two years or more and are comparable or worse than declines due to COVID-19 and Hurricane Katrina. Meanwhile, students who leave private schools and return to public schools have experienced increased academic achievement. While some may argue that test scores from the National Assessment of Educational Progress indicate that private school students fare better academically than their public school peers, this is more a reflection of the parents’ socioeconomic status and education level than the impact of private schooling on students. Research also suggests vouchers do not reliably improve high school graduation and college attendance rates. Because of these reasons, lawmakers looking to improve student outcomes should not pursue vouchers.

School vouchers have costs for students who remain in public school

In addition to the direct costs that the state incurs for school vouchers, school districts experience an additional cost when they lose students to private school: the cost of providing the same level of education for fewer students in public education. This cost is entirely borne by the students who remain in public education, even though they affirmatively did not make the choice to take up vouchers. When students leave public schools with a voucher, the school districts must still pay the same amount for costs that can’t immediately adjust to declines in enrollment, such as cooling/heating and utilities. These required payments for a district’s fixed costs mean that districts will have even less to spend on the costs that can adjust due to changes in enrollment. What this means is that public school students who remain in public school will have less funding allocated to them for adjustable costs like teaching, curriculum development, and pupil support services due to other students taking up voucher programs. (To calculate this cost for your district, see EPI’s fiscal externality calculator).

Alarm level 4: National voucher proposals threaten public schools throughout the country

Beyond state voucher programs, Congress is considering national voucher proposals. This would enlarge the scope of vouchers beyond Republican-controlled states. The Educational Choice for Children Act, or ECCA, (H.R. 817, S.292) is a proposal to create a national voucher program. The program would divert over $10 billion per year in tax dollars to private schools and families who homeschool. The bill would do this by establishing a new dollar-for-dollar tax credit for individuals and corporations that make charitable contributions to organizations that give scholarships— or vouchers—for students to attend private schools. Donors who give corporate stocks would receive more back in tax cuts than the after-tax value of the stocks if they had sold them. 

Beyond vouchers harming student educational outcomes, the program itself would be extremely expensive. The bill proposes that Congress allocate $10 billion in tax credits for the voucher programs. But that doesn’t even account for the cost of voucher programs to public schools. The sponsors of the bill estimate that ECCA would provide vouchers for 2 million students. Given that at least two-thirds of students who take up vouchers previously attended private school, we can estimate that 666,667 voucher recipients will come from public school, which is about 1.4% of total public school students. Using our fiscal externality calculator, we estimate that students who remain in public schools would lose an average of $151 per pupil, and public school systems would lose a total of $6.225 billion dollars due to a national voucher scheme.

Alarm level 5: Tax cuts reduce available revenue for public schools

Many states are following a recent trend of reducing revenue available for schools through sweeping tax cuts. Corporate and personal income tax revenue represents about half of state tax revenue, which, in turn, funds about half of K–12 education budgets. From 2021 through 2024, 28 states passed personal or corporate income tax cuts, which will result in hundreds of billions of dollars in lost revenue by 2028, and more states are considering or have passed income tax cuts in 2025. At the same time, some states are cutting and attempting to eliminate property taxes, which account for over a third of revenue for K–12 education on average. States that want to invest in opportunity and long-term economic prosperity and to help their students continue recovering from pandemic-related learning loss should reverse this harmful trend.

Conclusion: What would happen if we boosted public school funding instead?

Given the real and damaging threats to public school funding, we conclude by asking what students actually need to succeed. Growing evidence over the last decade shows that public schooling in the United States simply needs more resources to deliver even better student achievement—not some radical disruption in how it is delivered and by what institutions.

For example, research has shown that school finance reforms between 1972 and 2010 led to a 10% increase in school spending for 12 years, which increased high school graduation rates, wages and family incomes in adulthood for children from districts with the spending increase. Others have similarly found that a $1,000 increase in per-pupil spending for low-income districts would reduce the test score gap between low- and high-income school districts within a state by nearly 40% of the baseline gap.

Increasing funding, rather than withholding federal aid or using public dollars to pay for private schooling, is the path forward for public schools. Public schools have fallen short in many communities because of lawmakers’ choices to underfund them. But the only education system that can fulfill the promise of equal opportunity for all children, regardless of race, disability, or income, is a fully funded system of public schools. Lawmakers interested in building prosperous communities should invest in public schools rather than defunding and privatizing them.

What this report is about

Voucher programs for schools are rapidly expanding across the country. Under these programs, public budgets provide funding to parents to either send their children to private school or homeschool them.

These programs’ growing popularity raises the question of whether letting public money leave the public school system and subsidize private forms of schooling is a way to improve children’s access to an excellent education. EPI’s analysis shows that vouchers harm public schools.

To illustrate the damage, EPI has developed a tool that estimates fiscal externalities—the dollar costs to school districts from students leaving public schools with a voucher. An externality produces an outcome for those who aren’t responsible for the decision at hand. In this case, the fiscal externality is the negative effect that voucher programs will have on public school systems: Voucher programs redirect money away from traditional public schools.

Users of the tool can try out different scenarios to see how much money students will lose out on. The fiscal externality helps to put a number to the reality that children who don’t participate in voucher programs will still bear the cost for educational choices that others make. The fiscal externality does not quantify the entire cost of voucher programs. It represents a piece of those costs, but an important, and often hidden, cost.

Some factors affecting the impact of voucher programs on public schools

• How many children will go to private schools or be homeschooled in a given year?
• How quickly will enrollment numbers in public schools fall?
• How many of the school district’s costs are fixed and can’t be changed in response to lower enrollment numbers? (For example, heating and cooling costs for school buildings will remain the same regardless of enrollment.)
• How many of the school district’s costs are variable and can be changed in response to the drop in enrollment numbers?

Download the fiscal externality calculator

Summary

Universal voucher programs for schools are rapidly expanding across the country. Under these programs, states give parents stipends to either homeschool their children or send them to private school.

The growing popularity of vouchers raises a host of crucial questions and concerns. Key to informing the debate are questions of public finance and education quality. Is allowing public money to leave the public school system and follow kids to private schools the most effective or equitable way to make sure every child has access to an excellent education? Our view is that it’s not. Public dollars allocated to education should go to boosting spending in public systems, not subsidizing private education.

Proponents of vouchers try to claim that expanding them would not harm public resources for education. Their argument hinges on the fact that public school spending is generally determined by governments setting a per-pupil allocation and then multiplying this allocation by projected enrollment. This funding model allows voucher proponents to claim that if vouchers pull children out of public schools, it still leaves per-pupil spending untouched, even though vouchers might reduce overall spending. In effect, proponents are arguing that vouchers would not degrade public schools’ ability to provide educational services.

But there are many flaws in this argument. Most obviously, introducing any new demand on public expenditures will lead to some pressure throughout public budgets—including school budgets—even if it does not immediately come directly from public school allocations. Many vouchers end up going to students who have never been in public schools, so funding these students’ private school expenses introduces new pressures on public budgets.

This report focuses on another often underrecognized, but extremely important, source of stress that voucher programs can impose on the ability to provide an excellent public education for all. This is the fiscal externality imposed on school districts, students, and their families when declines in student enrollment numbers intersect with the fixed nature of many school costs. An externality produces an outcome for those who, otherwise, aren’t responsible for the decision at hand. In this case, the fiscal externality is the negative effect that voucher programs will have on public school systems that will now receive less money from the government.

The key contribution of this report is the introduction of a calculator for policymakers, researchers, and advocates to assess the quantitative impact of this fiscal externality for every school district in the country. In what follows, we walk through the general nature of the fiscal externality and then explain how it is operationalized in our tool.

We offer an example in Ohio, a state with one of the oldest active voucher programs in the country and where vouchers have grown substantially. We calculate the fiscal externality for a 5% decline in enrollment for the Cleveland Metropolitan School District, which has its own voucher program that works in conjunction with several state voucher programs.1 The results show that Cleveland public school students stand to lose between $364–$927 per pupil in education spending, which adds up to $12–$31 million in a total fiscal externality. 

Background on vouchers

As a policy tool, school vouchers have a long and questionable history. Following the Brown v. Board of Education ruling in 1954, several Southern states used vouchers as a tool to undermine school integration efforts, with states offering voucher programs to enable parents of white children to afford segregated private schools (Ford, Johnson, and Partelow 2017). Starting in the 1990s, states have enacted “modern” voucher programs with the claim of supporting students with special needs or students in low-income districts, offering a small number of these students pathways to private school. The number of students using vouchers stood at just 11,000 in 2000 but has increased to over 600,000 in 2021 (Welner, Orfield, and Huerta 2023). 

More recently, legislation has broadened the applicant pool for vouchers by creating universal programs in which any student can use public funds to pay for private education. Eight states have enacted universal voucher legislation in the past three years, bringing the total number of states with universal voucher programs to 11 (EdChoice 2024).

Figure A shows the 31 states with voucher programs and highlights the 11 states with universal programs. A program is considered universal when nearly all students are eligible for the benefit. Voucher programs are dominant in the Southeastern and Midwestern parts of the country, places that are also less likely to have robust public education budgets in the first place.

Figure A

While additional costs to provide quality education are not a problem in itself, (and we at EPI are huge advocates of increasing funding for public schools), study after study has found that voucher programs do not improve student achievement and, hence, are not a cost-effective way to spend any additional dollars that states or localities are willing to commit to K–12 education. In three states that enacted voucher programs —Louisiana, Indiana, and Ohio, researchers assessed student test scores in periods following program enactment and found that academic performance worsened.2

Students in Louisiana’s Scholarship voucher program experienced declining achievement in both language arts and mathematics during their first two years in the program (Mills and Wolf 2023). Similarly, Waddington and Berends (2018) found that voucher students in the Indiana Choice Scholarship voucher program experienced declining achievement in mathematics one year after attending private school. Under the Ed Choice Program in Ohio, voucher students who previously attended high-performing public schools performed worse than they would have had they remained in public school (Figlio and Karbownick 2016).

Instead of boosting student achievement in equitable and cost-effective ways, voucher programs generally end up putting large new demands on state and local budgets and increase the cost of educating students who remain in public school. In Arizona for example, where 75% of voucher program users are already in private school (SOS 2024), legislators grossly underestimated the cost of voucher programs—by tenfold. The voucher program was initially projected to cost $65 million in 2023 but actually cost upwards of $708 million. Because of these overruns, Arizona’s voucher policy is now leading to a state budget crisis (Hager 2024).

The intuition behind the fiscal externality

Public schools—like nearly every other economic entity in the modern economy—require a mix of inputs, some of which are variable and some of which are fixed. Variable inputs are those that are needed to produce any extra increment of a good or service. Fixed inputs are those that are needed to produce even a single increment, but whose costs don’t rise in lockstep with how much is produced. Take the example of a gas station. The fixed costs of the gas station are the pumps and the land. The owners must have secured these before they can sell a single gallon of gas to customers. And once the owners have bought the pumps, owners won’t need to buy them again as more gas is pumped. The variable costs of the gas station include the gas that is dispensed. Owners need to purchase each new gallon from suppliers before selling the gas to customers.

The revenue brought in by each new provision of the good or service must be sufficient to cover both variable and fixed costs. Otherwise, problems will arise. Say that a gas station rents its pumps and land for $10,000 per month and pays $2.00 per gallon to suppliers for each gallon of gas it sells. With the gas it sells each month, the gas station must earn not just enough to pay suppliers for the variable costs, but also the $10,000 to cover fixed costs. If the gas station doesn’t earn this much, it will eventually go bankrupt.

Unlike gas stations, schools don’t go bankrupt, but if they lose revenue and have a significant share of total costs accounted for by inputs that are not adjustable in the short run, then deep problems can result. Take the case of revenue loss driven by uptake of vouchers. Total revenue for public school systems is generally based on enrollment, with a per-pupil allocation of total spending being multiplied by the school district enrollment to determine total funding. At first, it would seem that reducing enrollment would reduce both total revenue and the number of students needing educational services proportionately, which should leave the schools’ ability to provide education unaffected.

But if any significant share of school costs is fixed, this is not true. Fixed costs, such as building electricity or utilities, do not automatically fall when student enrollment declines. As a result, when total revenue declines, districts are stuck paying more per pupil for costs they can’t adjust. All the downward adjustment that occurs when enrollment is reduced must be absorbed by variable costs, which fall even on a per-pupil basis. The fiscal externality is the per-pupil funds each district would require to maintain the same level of variable cost spending for remaining public school students due to voucher programs. This cost is entirely borne by state and local education budgets and leaves districts unable to deliver the same level of instruction to the remaining public school pupils.

What exactly would it mean for schools if variable costs per pupil fell? There would be fewer resources available on a daily basis to educate kids—fewer teachers and other staff members and fewer curriculum and education supplies. Education quality would suffer. To the degree that daily services provided to students by schools require all inputs—variable and fixed—it is fair to say that this reduction in variable inputs per pupil translates 1-for-1 into reduced services for students.

The key pieces of information for assessing how much stress this fiscal externality would place on public schools’ ability to provide high-quality education are these: how large and how quickly an enrollment decline could be triggered by the introduction of vouchers and what share of public schools’ costs are fixed.

Figure B

Figure C

To illustrate the fiscal externality, let’s take the example of a hypothetical school district of 30,000 students. The colored bars in Figures B and C show a stylized breakdown of aggregate expenditures (Figure B) and per-pupil expenditures (Figure C) before and after a decline in student enrollment for this school district. The navy bar on the far left in Figures B and C shows the aggregate spending and per-pupil spending before a decline in enrollment. This hypothetical district’s total expenditures are $500 million and, when divided by 30,000 students, yield a per-pupil spending of $16,667. The light blue bar illustrates the change to aggregate and per-pupil spending after a decline in enrollment by 10%. When enrollment declines by 10%, total spending reduces by 10% to $450 million, but per-pupil spending stays the same.

If schools were able to reduce all expenditures in direct proportion with declines in enrollment, then the fiscal externality of voucher programs would be zero. For example, schools might be able to reduce instruction costs by not renewing teacher contracts. These expenditures would be considered variable, in that they can adjust to enrollment changes.

In the short run, however, schools will still have to continue heating buildings and conducting basic maintenance and paying interest on accumulated debt regardless of enrollment changes. These fixed costs cannot be reduced commensurate with enrollment. The middle bars in Figure B and C show the changes in total and per-pupil fixed costs when enrollment is reduced by 10%. Because schools cannot reduce their fixed costs regardless of a change in enrollment, their total fixed costs stay the same, even if they receive less money to pay their costs.

That leaves schools no other option but to reduce their spending on variable costs more than proportionally to the decline in enrollment. The far-right bars in Figures B and C show the change in variable costs due to a 10% enrollment decline. In Figure B, the school district has to reduce total variable spending more than in direct proportion with enrollment declines because they cannot change their fixed costs when enrollment is reduced. As a result, when enrollment declines by 10%, total fixed costs remain at $250 million, while total variable costs drop by $50 million to $200 million.

Figure C shows per-pupil spending on fixed and variable costs if both spending categories could adjust to changes in enrollment in the middle and far-right bars. Like overall spending, if fixed and variable costs were both adjustable, per-pupil spending would stay the same at $8,333 for both categories. However, because total fixed costs cannot be reduced in the short run, per-pupil fixed costs increase, and variables costs decline by more than in direct proportion with enrollment change, from $8,333 to $7,407.

We can calculate the fiscal externality by taking the difference between the per-pupil variable costs before and after the enrollment decline. It doesn’t take a substantial drop in public school enrollment for these costs to be significant. In this hypothetical example, the externality is $926 per pupil or $25 million overall.

Considerations in calculating the fiscal externality

The stylized example in Figures B and C illustrates the potential impact of a decline in enrollment due to a voucher program that draws students away from public schools toward private schools. Turning these illustrations into policy-relevant magnitudes for specific school districts requires some careful considerations around the change in enrollment rate and their unique combination of fixed and variable costs. In the following sections, we describe some considerations users should take when using our accompanying calculator to project the fiscal externality of voucher programs for their own district.

Changes in public school enrollment

The fiscal externality will be proportional to the fraction of public school students that transition to voucher programs in a given year. Using the 2021–2022 National Center for Education Statistics (NCES) data, we provide enrollment for every district, and the user can choose the projected rate of enrollment loss they believe might occur in a district.

At present, the share of public school students that use voucher programs to transition to private schools is small. A large share of students that use voucher funding is already enrolled in private school. For example, in Wisconsin, over 80,000 students are enrolled in voucher programs, and research shows only 25% of these students are coming from public schools, suggesting the total loss in public school enrollment to vouchers is only 2% of public school enrollment in the state.3 In the accompanying calculator, users need to input the decline in public student enrollment to calculate the fiscal externality, not the share of total students enrolled in voucher programs, which is likely significantly larger than the decline in enrollment from public school.

In our calculation of the fiscal externality, we assume that the composition of public school students leaving the district to take up voucher programs is the same as the composition of public school students remaining in the district. This is more likely to be true when the voucher program is universal in nature, allowing all students to participate. However, if the voucher program targets a specific group—for example, students below a certain income threshold or those with disabilities—the composition of students leaving and those staying will be different, and results should be interpreted with caution.

For example, if a voucher program exclusively targets students with disabilities and these students take up the program, the resulting composition of public schools will have fewer students with disabilities, and the composition of private schools will have more of these students. This changing composition would alter the types of resources required for public schools, and districts would need to plan accordingly. Similarly, we don’t make assumptions about the grade level a student might be coming from in a transition away from public school to private school, which might have implications for the types of cost-cutting strategies districts can take. In reality, the choices a school has for cost reduction would be different if all the students came from a single grade, compared with if they were spread out across several grades.

How do schools adjust to less funding? An explainer on fixed and variable costs

The lost revenue of students transitioning away from public school should be blunted if public schools are able to commensurately reduce expenditures in proportion to the reduction in enrollment. However, as described in our hypothetical example in the previous section, only certain costs can likely be adjusted in the short run. Many educational costs, such as ongoing construction contracts, heating and utilities, curriculum development and principals’ offices, cannot be changed in the first several years of an enrollment decline. These costs are fixed.

Using district-level data from the National Center for Education Statistics, we categorize education costs into three major categories: instruction, service, and capital. In this section, we describe these categories in detail and provide some context for determining what share of each cost category might be considered fixed or variable. Ultimately, users can rely on their context-specific knowledge to identify the share of each category that is fixed or variable.

Instruction costs: The first major cost category for districts is instruction, which includes teacher salaries, benefits, and non-personnel costs (supplies and materials) for regular, special, and vocational programs. Prior literature has categorized the vast majority of instruction cost as variable: When enrollment declines, districts can reduce the number of teachers much faster than they can halt a building being constructed (Ladd and Singleton 2020; Bifulco and Reback 2014).

Yet, there may be several instances when not all instruction costs can be reduced commensurately with enrollment even over a year or two. First, not all teachers are subject to annual contract renewals, so even if a district decided to lay off some teachers, it might not be possible to do so in the time frame being estimated. Second, many teachers are spread out across grade level, subjects, and even schools, which makes reductions in teachers hard to justify. If, for example, fourth graders were exclusively targeted by voucher programs, school districts could reduce the number of fourth grade teachers in response, but often the decline in enrollment is much more diffuse, making the choice to let go of any one teacher difficult. Third, non-personnel costs like materials and supplies may be bought in bulk and can’t be changed in the short run in response to enrollment declines. Fourth, districts of different sizes may have different abilities to adjust instruction costs (Lapp et al. 2017). Rural districts, which are smaller in population, may make different choices in adapting to enrollment reductions than large urban districts, which may have greater ease in moving resources around.

Service costs: Service costs include the costs a district incurs to support instruction, student development, and achievement. These costs include pupil services (expenditures for attendance recordkeeping; social work; student accounting; counseling; student appraisal; record maintenance and placement services; medical, dental, nursing, psychological, and speech services), instructional staff services (supervision and instruction service improvements; curriculum development; instructional staff training; and instructional support services, such as libraries, multimedia centers, and computer stations for students outside the classroom), costs associated with general and school administration, and food services. Service costs also include any nonelementary and secondary services provided by the district such as adult education and English-language learning,

Prior literature has categorized at least some service costs as variable, while categorizing other service costs as fixed (Bifulco and Reback 2014). While some districts may view all services as nonadjustable in the very short run, other districts may be able to cut certain types of services, such as nursing or mental health services, in an effort to protect resources that directly support student learning in the short run. Users looking to identify a longer time horizon to estimate a projected change in district expenditures may decide that a larger share of service costs is variable, while users looking to identify a fiscal externality in the short run may decide to categorize a majority of service costs as fixed.

Capital costs: The final category of district costs is the cost of capital and debt service: school buildings, operations and maintenance, and outstanding payments that a district owes. Prior literature has often categorized these costs as rigidly fixed over the short and even medium run because the time horizon for schools to make such large cost-cutting changes (closing a building, for example) is much longer than the time horizon needed for districts to make changes to instruction or other services. However, if a researcher is looking to project the medium- to long-run effects of voucher programs, capital costs may very well be adjustable, as schools decide to take increasingly drastic measures to respond to enrollment reductions.

Table 1 shows the average cost of instruction, service, and capital costs across all public school districts in the U.S in the 2021–2022 school year. Overall instruction costs average to $9,237 per pupil or about 51% of total costs. Service costs made up 27% of a district’s overall spending, costing $4,845 per pupil. Total pupil services ($1,012 per pupil), total instructional staff services ($756 per pupil), and school administration ($836 per pupil) comprised the majority of service costs. Capital costs make up the remaining 22% of the budget, and construction costs and operation and maintenance costs were the largest costs in this category. 

Table 1

Categorizing costs as fixed and variable through adjustment rates

The accompanying calculator lets users determine the share of each cost category that is variable and the share that is fixed. These choices are formally defined as picking an adjustment rate.

The adjustment rate identifies the rate at which costs can be reduced in relation to enrollment changes, and ranges from 0 to 1. When costs can be reduced exactly proportionally to enrollment changes, they are considered variable. When costs cannot be reduced at all, they are considered fixed, and the adjustment rate is 0, meaning that enrollment declines do not reduce costs at all. When costs are fully adjustable, the adjustment rate is 1, meaning that each percentage decline in enrollment reduces costs by 1%.

For example, if the adjustment rate for instruction costs is equal to 1, then 100% of instruction costs can decline commensurate with changing enrollment. If the user chooses an adjustment rate of less than 1, say 0.5 for example, spending is thought to be “stickier,” or less adjustable. In the case of instruction, given that not all teacher contracts can be halted when there is a change in enrollment, it’s likely that instruction costs are not fully adjustable in the short run. In such an instance, the adjustment rate might be closer to 0.8. In effect, 80% of their costs might be adjustable in the short run, but 20% are fixed. 

Service costs are likely to have a lower adjustment rate than instruction since many functions of education in this category are required to run and maintain a school district regardless of how many students are enrolled. For example, school districts will always need nurses, guidance counselors and social workers, food prep workers and staff professionals to develop curriculum materials for students. These needs imply that service costs are likely a bit stickier than instruction costs. In the short run, service costs might have an adjustment rate of 0–0.5, meaning that each percentage point decline in enrollment can only reduce service costs by 0–0.5 percent—the rest remain fixed. In the longer run, these costs will likely become more adjustable, districts will have time to respond to declines in enrollment, and the adjustment rate should increase commensurately.

Capital costs likely have the lowest rate of adjustment for districts in the short run. This is because districts must continue to pay for construction, operation and building maintenance costs, and to make payments on debt, regardless of small- or even medium-sized enrollment changes. In the longer term, it’s possible these costs will be more adjustable. Prior literature generally categorizes 100% of capital costs as fixed, or as having an adjustment rate of 0 in the short run. However, users looking to make longer-term estimates can decide whether to raise the adjustment rate of capital costs.

When using the calculator, users considering an analysis in their district should think carefully about which costs are adjustable in the time horizon they choose and alter the adjustment rates accordingly.

Outside the scope of this report and calculator is how each voucher program is financed or how it interacts, in an accounting sense, with K–12 budgets. We assume that public K–12 schools are essentially funded on a per-pupil basis, so a 10% reduction in enrollment will lead directly to a 10% reduction in total state/local funding. We don’t describe how vouchers are financed or how they might compete with other educational resources.

Case study: Calculating fiscal externality for Cleveland’s public school district

Among states offering school voucher programs, Ohio currently has five distinct programs. Ohio currently ranks sixth in the country for voucher enrollment (NCES 2024) with 4.1% of its students receiving voucher payments. Three of the programs have expanded eligibility over the past 10 years and are now universal. One such program, the Cleveland Scholarship, is Ohio’s oldest voucher initiative and the second oldest in the nation. Established in the 1995–1996 school year, the Cleveland Scholarship provides between $6,000 and $9,000 in tuition assistance. All students within the Cleveland Metropolitan School District are eligible, with 20% of spots reserved for families earning less than 200% of the federal poverty level. As of 2022, 7,889 students were enrolled in the Cleveland Scholarship, a significant increase from 5,525 in 2010 and 1,994 in 1997 (Fleeter 2023).

Table 2 provides a breakdown of Cleveland Metropolitan School District’s costs in the 2021–2022 school year. In 2022, 35,319 students were enrolled in Cleveland Metropolitan School District, and 51.2% of the district’s costs went to instruction. Salaries comprise 32% of the total budget and 63% of instruction costs. Benefits comprise 13% of the total costs and 26% of instruction costs. Service costs make up 32% of the overall budget and were comprised of business/central costs (payroll and accounting services, and development and evaluation services), school administration costs, and pupil service costs. The remaining 16% of the budget are made up of capital costs, a majority of which were operation and maintenance costs for buildings and facilities (9.4% of the overall budget and 58% of capital costs).

Table 2

Calculating the fiscal externality for an enrollment change in the Cleveland Metropolitan School District requires users to identify the adjustment rate for each of the three major cost categories: instruction, service, and capital expenditures. Table 3 shows how different choices of adjustment rates for each category would translate into fixed and variable costs for the Cleveland Metropolitan School District.

Table 3

The far-left columns provide the per-pupil cost of instruction, service, and capital costs. The subsequent columns move from a completely variable adjustment rate (1) on the left side of the table to a fixed adjustment rate (0) on the right side of the table. When the adjustment rate is 1, 100% of the costs can be reduced proportionately with changes in enrollment—costs are effectively characterized as variable costs. On the far right, when the adjustment rate is 0, 0% of the costs can be reduced proportionately with changes in enrollment—they are all set to be fixed costs. The middle columns show how adjustment rates between 0 and 1 would lead to different shares of each cost category being categorized as fixed or variable. Adjustment rates closer to 1 lead to a larger share of the cost category varying with changes in enrollment. Adjustment rates closer to 0 will have a smaller share of the cost category varying with changes in enrollment.

Once the adjustment rates are determined for the cost categories, the calculator provides an estimate for the fiscal externality. Table 4 shows a range of estimates for a hypothetical fiscal externality in the Cleveland Metropolitan School District, under a scenario of a 5% enrollment drop due to voucher programs. The table shows a range of adjustment rates for instruction, service costs, and capital costs. The top panel provides estimates when the adjustment rate for capital costs is set to equal 0, and the bottom panel provides estimates when the adjustment rate is set to equal 0.2. We chose these scenarios to reflect the fact that, in the short term, capital costs will likely not change in response to changes in enrollment and so the vast majority of this category is considered fixed.

Table 4

Under each capital adjustment rate scenario, we provide a range of adjustment rates for instruction and service costs, reflecting different possibilities that school districts like Cleveland Metropolitan may face in the wake of an enrollment drop. The columns indicate different adjustment rates for instruction. The left-most column assumes all instruction is fully adjustable to changes in enrollment, and the right-most column assumes an adjustment rate of 0.5, which means that instruction costs can only be cut by 5% for every 10% reduction in enrollment. The rows indicate different adjustment rates for services. We pick adjustment rates that are smaller to reflect the fact that many cost functions categorized as a service cost are functions required to maintain basic education regardless of enrollment size.

In the top left of the table, when the adjustment rate of the instruction cost is 1 and the adjustment rate of service and capital costs is 0, the fiscal externality is $606 per pupil. In the short run, it’s likely that not all instruction costs are variable. In the middle column in which the adjustment rate for instruction is 0.8, the fiscal externality ranges from $533–$734. In the last column, instruction costs are much less adjustable (the adjustment rate is 0.5), and the fiscal externality ranges from $726–$927. A plausible scenario for the school district might be one where the adjustment rate for instructions costs is 0.8 (80% of instruction costs are variable) and an adjustment rate for service costs is 0.2 (20% of service costs are variable). In this case, voucher policies that reduce enrollment by 5% mean that the Cleveland Metropolitan School District must reduce services by $654 for each public school student. This decline translates to a total fiscal externality of $22 million a year (in 2022 dollars).

The results show considerable fiscal externalities associated with declines in enrollment from voucher programs. A 5% decline in enrollment is just an example, but given that 4.1% of Cleveland’s students are already enrolled in voucher programs and these overall program costs often grow quickly (Abrams and Koutsavlis 2023), it’s not impossible to see a scenario where large voucher programs could lead to sizable fiscal externalities.

These externalities are not just a problem for public budgets. Students stand to lose out on their potential educational achievement when funding to schools is cut (LaFortune, Rothstein, and Schanzenbach 2018). When that funding is reduced, students, particularly in high-poverty neighborhoods, are likely to have worse outcomes than they would have had if their schools had retained the previous level of education funding.

Download the fiscal externality calculator

Conclusion

Vouchers undermine efforts to make excellent public education available to all children by pressuring fiscal resources available to public schools along many different dimensions. In this report we provide a tool that can help interested parties put estimated dollar amounts on one particular channel of this pressure that is often underrecognized: the fiscal externality of voucher programs on public education. While this report focuses on reductions in enrollment that can occur from voucher programs, the fiscal externality can also be calculated for any program that puts downward pressure on public school enrollment, such as charter schools.

Building on approaches that examine the fiscal impacts of charter schools (Bifulco and Reback 2014; Ladd and Singleton 2020), we provide a method and accompanying calculator for stakeholders all over the country to estimate the fiscal externality of vouchers in their home districts. We use the prior literature as a guide to categorize spending into categories that are likely adjustable in response to declines in enrollment (variable costs) and costs that are likely nonadjustable to changes in enrollment (fixed costs). However, we provide the opportunity for users to define what share of each of these costs is actually adjustable when they do their own calculations. Users should use their discretion in determining the share of variable and fixed costs that are adjustable in their given district.

In the example we gave with Cleveland, we find that the fiscal externality of a 5% decline in public school enrollment from a voucher policy could be very large, ranging from $12 million to as high as $31 million when a large share of a district’s cost functions is considered to be fixed. We stress that these calculations are mere projections based on a possible decline in public school enrollment. Users will have to provide their best guess as to how much enrollment will actually decline in order to calculate the fiscal externality in their home districts.

Finally, we stress that these are just some of the costs of voucher programs, not the only costs. There are many other costs of subsidizing private education with public funds. The fiscal externality is a key part of measuring the cost of vouchers and of deep importance if we care about the quality of public schooling in our country. Good public education requires strong funding and stable enrollment numbers. Voucher policies threaten to destabilize public education funding and student learning, in part, by introducing an element of uncertainty. If school administrators don’t know what their future enrollment numbers will look like, this will prevent them from being able to make plans. The one certainty is that student achievement will suffer as a result.

Notes

1. Cleveland Metropolitan School District was previously named Cleveland Municipal School District. In the National Center for Education Statistics, the district is named Cleveland Municipal School District.

2. Critically, these studies were conducted prior to the COVID-19 pandemic and, thus, do not reflect changes in education achievement due to the pandemic.

3. The commitment of public dollars to students who have never been in public schools is an additional source of fiscal stress for states; however, it is not the fiscal stress documented in this report and tool.

References

Abrams, Samuel E., and Steven J. Koutsalvis. 2023. The Fiscal Consequences of Private School Vouchers. Southern Poverty Law Center and Education Law Center, March 2023.

Bifulco, Robert, and Randall Reback. 2014. “Fiscal Impacts of Charter Schools: Lessons from New York.” Education Finance and Policy 9, no. 1: 86–107. https://www.jstor.org/stable/educfinapoli.9.1.86

EdChoice. 2024. “School Choice in America.” School Choice in America [data dashboard], accessed October 23, 2024. 

Figlio, David, and Krzysztof Karbownik. 2016. Evaluation of Ohio’s EdChoice Scholarship Program: Selection, Competition, and Performance Effects. Thomas B. Fordham Institute, July 2016.

Fleeter, Howard B. 2023. Ohio School Voucher Overview. Ohio Education Policy Institute, June 2023.

Ford, Chris, Stephenie Johnson, and Lisette Partelow. 2017. The Racist Origins of Private School Vouchers. Center for American Progress, July 2017.

Hager, Eli. 2024. “School Vouchers Were Supposed to Save Taxpayer Money. Instead They Blew a Massive Hole in Arizona’s Budget.” ProPublica, July 16, 2024.

Ladd, Helen F., and John D. Singleton. 2020. “The Fiscal Externalities of Charter Schools: Evidence from North Carolina.” Education Finance and Policy 15, no. 1: 191–208. https://doi.org/10.1162/edfp_a_00272

Lafortune, Julien, Jesse Rothstein, and Diane Whitmore Schanzenbach. 2018. “School Finance Reform and the Distribution of Student Achievement.” American Economic Journal: Applied Economics 10, no. 2: 1–26.

Lapp, David, Joshua Lin, Erik Dolson, and Delia Moran. 2017. The Fiscal Impact of Charter School Expansion: Calculations in Six Pennsylvania School Districts. Research for Action, September 2017. 

LeFebvre, Joanna. 2024. “EARNCon 2024 Voucher Presentation” (slide presentation). Data retrieved from Education Commission of the States [database], September 12, 2024.

Mills, Jonathan N., and Patrick J. Wolf. 2023. “Vouchers in the Bayou: The Effects of the Louisiana Scholarship Program on Student Achievement After 2 Years.” Educational Evaluation and Policy Analysis 39, no. 3: 464–484. https://doi.org/10.3102/0162373717693108

National Center for Education Statistics (NCES). 2024. “Private School Universe Survey (PSS)” [database], 2021–2022.

Save our Schools Arizona Network (SOS). 2024. “Universal Vouchers: The Verdict Is In.” Save Our Schools Arizona Network, January 17, 2024.

Waddington, R. Joseph, and Mark Berends. 2018. “Impact of the Indiana Choice Scholarship Program: Achievement Effects for Students in Upper Elementary and Middle School.” Journal of Policy Analysis and Management 37, no. 4: 783–808. https://doi.org/10.1002/pam.22086

Welner, Kevin, Gary Orfield, and Luis A. Huerta, eds. 2023. The School Voucher Illusion: Exposing the Pretense of Equity.  New York: Teachers College Press at Columbia University.

The teacher shortage in the nation’s public schools—particularly in our high-poverty schools—is a crisis for the teaching profession and a serious problem for the entire education system. It harms students and teachers and contributes to the opportunity and achievement gaps between students in high-poverty schools and their more affluent peers. Policy choices as well as policy inaction have contributed to the factors that have eroded the appeal of the teaching profession and it will take a comprehensive policy agenda to address those factors.

EPI’s teacher shortage policy agenda plots a course to return teaching to a profession in which teachers are compensated on par with their college-educated peers, operate in environments where they can teach effectively, get the training they need early in their careers and the professional development they need throughout their work lives, have their professional judgment incorporated, and have the opportunity to use the expertise they attain to help shape what goes on in their classrooms and their schools. This policy agenda has two components: a set of four foundational recommendations for how to understand the context and approach the problem in a way that will actually solve it, followed by specific policies that, if implemented together, could go a long way toward solving the teacher shortage crisis. This policy agenda builds on an in-depth analysis of the size, scope, and drivers of the teacher shortage, as detailed in the first five reports in the EPI’s “Perfect Storm in the Teacher Labor Market” series. The agenda addresses the factors we identified in the series as well as other specific or underlying factors that our past research and understanding of the issue tell us are important, such as lack of a properly resourced system, the lack of a diverse teaching workforce, student loan debt payments burdening teachers, and the uneven effectiveness of initiatives for teachers-in-training. We review the analysis, present the principles underlying the agenda, and then present the specific agenda items, which fall into two categories:

Overarching principles for how to approach the teacher shortage problem

• Understand that the teacher shortage is caused by multiple factors and thus can only be tackled with a comprehensive set of long-term solutions.
• Understand that the complexity of the challenge calls for coordinated efforts of multiple stakeholders.
• Increase public investments in education.
• Treat teachers as professionals and teaching as a profession.

Specific proposals in the policy agenda to address the teacher shortage

• Raise teacher pay to attract new teachers and keep teachers in their schools and the profession.
• Elevate teacher voice, and nurture stronger learning communities to increase teachers’ influence and sense of belonging.
• Lower the barriers to teaching that affect teachers’ ability to do their jobs and their morale.
• Design professional supports that strengthen teachers’ sense of purpose, career development, and effectiveness.

The need for an agenda: A review of the size and scope of the shortage and the factors behind it

The teacher shortage in the nation’s public schools is an increasingly recognized but still poorly understood crisis. Much attention has focused on the size of the shortage (about 110,000 teachers in the 2017—2018 school year, by one estimate), its monetary costs, and some of its negative effects on students and teachers.1 But the full size of the teacher shortage when accounting for credentials, the multiple causes of the teacher shortage, and the unequal distribution of the shortage across low- and high-poverty schools had received less scrutiny.  In 2019, EPI published the “Perfect Storm in the Teacher Labor Market” series of reports examining the full magnitude of the teacher shortage and the working conditions and other factors that contribute to the shortage. In the series, we looked at public schools overall, and compared trends and conditions in low- and high-poverty schools.2 We found a sizable and in some cases growing share of teachers without the teaching credentials associated with being a highly effective teacher, and even higher shares of teachers without these credentials in high-poverty schools.3 We traced the struggle to adequately staff schools to high turnover and attrition and to a sharp drop in the number of people on track to pursue a career in teaching.4 And we identified a number of factors that are prompting teachers to quit and dissuading people from entering the profession, including low relative pay, safety hazards and other challenging aspects of the school environment, a lack of respect for teachers’ knowledge and judgment, and uneven access to useful types of professional development.5 We found turnover and attrition a bigger problem in high-poverty schools than in low-poverty schools.6 The higher turnover and attrition in high-poverty schools is perhaps not surprising given that shares of teachers reporting difficult working conditions were higher in high-poverty schools than in low-poverty schools.7

As the series warned, the teacher shortage isn’t just a sudden, intense, and increasing decoupling of the number of new teachers needed in our public schools and the number available to be hired each year since 2013. The shortage of teachers is a crisis for the teaching profession, and a serious problem for the entire education system. It harms students, teachers, and the public education system as a whole. It deters student learning, reduces teachers’ effectiveness, consumes economic resources that could be better deployed elsewhere, and makes it more difficult to build a solid reputation for teaching and thus to professionalize it, which further contributes to perpetuating the shortage. Finally, the uneven distribution of highly qualified teachers across low- and high-poverty schools impedes the goal of providing a sound education equitably to all children in the United States. The greater shortage of teachers in high-poverty schools contributes to the opportunity and achievement gaps that plague our system as a result of underfunding, poverty, and inequality. The series, and a summary of its findings, are available on the Economic Policy Institute’s Teacher Shortage page, https://www.epi.org/research/teacher-shortages/.8

This is the sixth and final report in the “Perfect Storm in the Teacher Labor Market” series, and it presents a policy agenda to confront the teacher shortage—including closing the gaps in access to sufficient, highly qualified teachers across schools. It addresses the factors we identified in the series as well as other specific or underlying factors that our past research and understanding of the issue tell us are important, such as lack of a properly resourced system, the lack of a diverse teaching workforce, student loan debt payments burdening teachers, and the uneven effectiveness of initiatives for beginning teachers.

But, admittedly, reaching an equilibrium between the number of teachers available and the number of teaching positions to fill in a given year is a complicated process. None of the factors or conditions that affect supply and demand for teachers operate in isolation, and they are likely to change, at any time, due to myriad reasons. Despite this fact, we do know that some of the factors that create a need for new teachers are only going to increase in the years ahead. For one, the expected enrollment in public and secondary schools is expected to increase by close to a million students between 2019 and 2028.9 Student-to-teacher ratios, another quality indicator, have shown a decline when long-term trends are examined, but are still not below the pre-Great-Recession parameters in the aggregate and in many states, and are substantially larger in public schools than in private schools (NCES 2019c; Hussar and Bailey 2020). And quality arguments would, if anything, recommend smaller class sizes (Chetty et al. 2011; Mishel and Rothstein 2002). Demand for teachers is also spurred by the need to hire specialized teachers who can provide instruction in changing content areas and competencies. Finally, the teaching workforce is aging, which adds to demand by increasing the number of teacher vacancies while shrinking supply (the number of teachers available for rehire) (Ingersoll et al. 2018).10

Unless decisive action is taken, the realities indicate that the teacher shortage crisis will potentially turn into a new perpetual ill of the educational system, leaving students, schools, and the education system as a whole behind where they need to be. This set of policy recommendations would also make teaching a more attractive career and make possible a shift in how we see and treat the teaching profession.

The policy agenda

The policy agenda to confront the teacher shortage while closing the gaps in access to highly credentialed teachers by school poverty has two components: a set of four foundational recommendations for how to approach the problem in a way that will actually solve it, followed by specific policies that, if implemented together, could go a long way toward solving the teacher shortage crisis.

This collective set of policies would improve the working conditions and other factors that have diminished the appeal of teaching as a career. In addition, the policies in this agenda that treat teaching as a true profession could help nudge a societal and political shift reinstating the prestige of the profession, a profession only partially respected: While people say they highly value public schools and have fond memories of the impact teachers have had in their lives, a record low of 46% of those polled said they would like their child to become a teacher (PDK 2018).

Overarching principles for how to approach the teacher shortage problem

The success of the specific policies in the proposed agenda to tackle the teacher shortage hinge on four key overarching system-level recommendations. These overarching recommendations have to do with how to approach the problem from the outset. We make a distinction between the overarching and the specific recommendations for the following two reasons. First, because the teacher shortage problem is profoundly shaped by the broad education context, recommendations that tackle the broad context also tackle the drivers of the shortage. Second, recommendations that tackle the broad context would if fully implemented make the specific recommendations easier to implement, and, in some aspects, even unnecessary. If, as our overarching recommendations suggest, we address the challenges that threaten equity and excellence in our education system, treat teachers as professionals, get all stakeholders involved, and tackle the issues with a set of comprehensive long-term solutions, then factors such as worsened conditions in high-poverty schools or a lack of useful professional development would take care of themselves.11 These overarching recommendations are discussed in detail below.

Understand that the teacher shortage is caused by multiple factors and thus can be tackled only with a comprehensive set of long-term solutions

In public policy and media spheres, single “magic remedies” to U.S. education challenges sometimes gain traction, with their appeal of easy—and fast—solutions. But while well-intentioned short-term remedies may temporarily ameliorate some of the conditions that lead to the teacher shortage crisis, the benefits of these quick fixes will be short-lived and not at the scale of the problem.12 A lasting, sufficient response to the teacher shortage thus requires a systemic policy strategy tackling multiple underlying, interdependent causes simultaneously. Specifically, we are referring to low relative teacher pay and compensation, teachers lack of say in how they do their jobs, difficult working environments (“school climates”), and lagging professional supports.

Note, however, that this list of components in our policy strategy is likely not exhaustive. Not addressed here are other drivers that we lack the data to assess or for which there are not yet evidence-based policy solutions that we can build upon.13 All these explored and unexplored factors are interrelated, and all reflect and respond to the context in which they occur. This is the nature of the problem and the realities states and school districts encounter.14 Thus, although our analysis of the data and learnings from other research provide a substantial understanding of the problem, we acknowledge that our agenda is limited by the absence of complete data or experimental evidence on some factors that play a role in the teacher shortage. We call for continued research on the problem, including controlled experiments to test the effectiveness of policy solutions and the use of a comprehensive look of all the factors at play. We also strongly urge researchers and policymakers scrutinizing teacher labor markets and the drivers of the shortages to use the quality and equity framework used in our series so that we can protect and improve the equity and excellence in our education system.

Understand that the complexity of the challenge calls for coordinated efforts of multiple stakeholders

As noted, policy solutions to address the teacher shortage must be comprehensive and long term. Adding to the complexity of the challenge, redressing these longstanding and multifaceted problems that impede children’s success and teachers’ ability to do their jobs effectively will require that stakeholders at all decision-making levels—schools and districts in local communities, states, and even federal entities—coordinate their efforts.

Both research and practice tell us that many institutions beyond schools and school districts—such as states, parent-teacher associations, school boards, and unions—play a role in shaping the overall status of teachers and their working conditions. This underscores the importance of aligning all the institutions that need to work together to address the teacher shortage.15

Increase public investments in education

Insufficient school funding is a well-documented problem that affects everything from the state of school facilities to student performance—and is a particular problem in high-poverty schools.16 With regard to the teacher shortage, the lack of funding is evident in many of the factors that contribute to teacher attrition and the declining attractiveness of teaching, such as low pay relative to other college-educated professionals, challenging working conditions, and inadequate professional development. Thus boosting funding overall is a foundational step: At the very least we must bring funding up to a basic, adequate level to educate children to the levels desired and to meet growing needs in our schools. Furthermore, funding needs to be progressive—greater for high-poverty schools so that resources are equitable (i.e., given the current inequities, the highest-poverty districts need more resources in order to deliver a sound education to all their students).17

In addition, we need specific investments that directly address all the specific causes of the teacher shortage, like low teacher pay and challenging working conditions. In summary, better-funded systems overall would make the profession more attractive and supportive of teaching and learning. Society claims that teachers are incredibly important and yet when it comes to investing in teachers and education our actions don’t match our professed values. Increasing the investment in the system overall will demonstrate that society does value teaching.

Treat teachers as professionals and teaching as a profession

As a society we claim—and policymakers often grant—that teachers are professionals, but the ways we treat them indicate otherwise. The prestige of the profession is questionable (PDK 2018) and teachers’ voices are often ignored. Generally, professionals’ voices are central in key decisions regarding how they do their work, the kinds of supports they need to do it well, or their interactions and relationships with peers and supervisors. As described below, none of these is true with respect to teachers.

Teachers report being frustrated, demoralized, and dissatisfied due to the combination of underfunding, challenging working environments (described in García and Weiss 2019d and summarized below), disrespect for their profession, and generalized distrust between teachers and administrators (Weingarten 2019; Schultz 2019). Little respect for teachers’ professional judgement and consideration for their knowledge is a problem across the board. Teachers don’t feel supported and valued: According to a recent survey of public school teachers, just 10% say they feel they’re valued a great deal, whereas 42% say their community values them a good amount and 48% say they feel valued either “just some” or a little or not at all (PDK 2019). Our data showed that teachers’ voices are systematically missing in decisions around their own working conditions. Specifically, our analysis in the Perfect Storm series looked at what teachers reported about their situations and found that the following are factors in the teacher shortage: A lack of teacher influence and autonomy, poor learning communities (i.e., environments in which teachers have little opportunity to learn from one another or through professional development activities), and low satisfaction with working conditions and appreciation for the profession. We compared teachers who had quit the profession by the time of a follow-up survey with those who had stayed and found that among teachers who quit, larger shares lacked influence over school policy or the activities, routines, and rules that take place in their classrooms, and were not fully satisfied with teaching at the schools—with gaps ranging from 3 to 17 percentage points. We also showed that larger shares of teachers who quit worked in less cooperative environments (without a great deal of cooperation among staff members), and said they did not get strong support or encouragement from the administration (García and Weiss 2019d, 2019e).

Puzzlingly, though these issues of teacher say were a clear factor in the teacher shortage, it is hard to find other research or policy papers on teacher retention or recruiting that have explicitly looked at the importance of giving teachers a voice.18

In short, teachers do not feel treated as the professionals they are, and policies don’t reflect a high value placed on teaching, which further reduces the prestige of teaching and impedes the professionalization of teaching. Overall, then, the recommendation is simply to stop doing what we are doing wrong and put in place the steps needed to treat teachers as true professionals.

To illustrate the point that teachers are not treated as professionals, we can look to physicians as an example. We do not expect a new medical school graduate to march into the operating room and be in charge of a full procedure. This is not because we do not value medical school or doubt the ability of a new doctor, but because we recognize that it takes many years of hands-on training, peer-to-peer and veteran mentoring, and ongoing professional training to be fully prepared. Yet many teachers go right into classrooms, often ones in which students have multiple unmet needs, with little on-the-ground training or support, and are judged based on their performance in those circumstances.

Taking the legal and judicial professions as further examples: Lawyers and judges find that their views and judgments are respected and that their profession is accorded a great deal of prestige. Yet many teachers find their judgment is ignored when it comes to making decisions on the priorities, practices, tools, and other features of daily classroom work.

The status of university professors provides another example: They keep up with the changing needs and nature of their profession by participating in continuous training and accessing new career opportunities. In contrast, K‒12 teachers are told what types of professional supports they will receive, and when and how they receive them (“professional development [is] done to the teachers, not with them” as noted in Ferlazzo 2018), and often face incentive systems and schedule limitations that push them to compete rather than to collaborate.

This overall change in the way we view and treat teachers would make many of the specific policy and practice recommendations outlined in the next section either much easier to implement or, in some cases, unnecessary because the problems they seek to fix would cease to exist.

Specific proposals in the comprehensive policy agenda to address the teacher shortage

In what follows, we focus on specific policies designed to tackle the underlying factors that drive the teacher shortage, as described in the previous installments of the “Perfect Storm in the Teacher Labor Market” series.

As noted earlier, all components—low relative teacher pay and compensation, teachers’ lack of say in how they do their jobs, difficult working environments (“school climates”), and lagging professional supports—are interrelated and all need to be addressed. Given the likely difficulty of addressing all the factors at once, the proposals detailed below will likely need to be implemented incrementally. Policymakers who must make difficult choices should prioritize first steps based on the unique circumstances of their states, districts, and schools. But it is important to remember that the teacher shortage has great costs and thus that not moving forward on this agenda to address the teacher shortage would incur even greater costs for schools, students, communities, and the nation as a whole.19

Raise teacher pay to attract new teachers and keep teachers in their schools and the profession

One concrete reason why teachers are leaving the profession and why fewer people are interested in becoming teachers is low pay, in both absolute and relative terms. In 2018, average teacher wages adjusted for inflation were lower than they were in 1996 and the average teacher earned 13.1% less in wages and benefits than other comparable college graduates, a pay gap that had grown substantially from 1993, when teachers earned just 2.9% less (Allegretto and Mishel 2019).20 Teacher salaries have been described as too low “to support a middle-class existence” (Podolsky et al. 2016), and, depending on family size, may put teacher heads of household below the income threshold to receive public subsidies.21

Another indicator that low teacher pay is a source of financial distress is the growing share of teachers who are taking second jobs, on top of their full-time jobs at school (59.0% took on additional work in the 2015–2016 school year, up from 55.6% in 2011–2012). For these teachers, moonlighting during the school year brings in on average $4,100, making up 7.0% of their combined base salary and moonlighting income (García and Weiss 2019c). News reports attribute moonlighting to teachers’ need to supplement their income.22

The total teacher pay penalty is even greater for teachers in high-poverty schools.23 Relative to teachers in low-poverty schools, teachers in high-poverty schools are paid about 10% less (teachers in low-poverty schools are paid $53,300 vs. $58,900), and also earn slightly less from moonlighting ($4,000 vs. $4,300) (García and Weiss 2019c).

Among individuals taking a college entrance examination, low salaries for teachers is the most often cited reason for a lack of interest in teaching (Croft, Guffy, and Vitale 2018). Parents, too, cite low teacher pay as a concern: They say that despite their continued respect for the profession, they would not want their children to become teachers, given the low pay (PDK 2018).

In addition to affecting recruitment and retention, low pay affects the relative qualifications of the teaching workforce.24 Research has shown how lower pay can reduce the share of highly credentialed teachers in schools by changing both who applies to become a teacher and who stays in the profession.25

The following reforms aim to increase teacher pay in a way that will strengthen recruitment and retention—particularly in high-poverty schools. Each reform listed below is necessary because it plays a specific role in addressing the challenges associated with compensation.

1) Increase teacher base pay across the board

At the outset, strengthening base salaries by giving every teacher a flat increase would help close the gap in pay between teachers and their peers in other professions, keep more teachers in the school and in the profession, and strengthen the credentials of the teaching workforce.26 In fact, our regression analysis using the teacher salary information in García and Weiss 2019c shows that higher salaries correlate with increased teacher retention (see Appendix Table 1). We find that a $1,000 increase in a teacher’s base salary is associated with a 0.4 percentage-point increase in the probability of that teacher staying in the same school and with a 0.3 percentage-point increase in the probability of that teacher staying in teaching. These associations are modest, but any proposed increase in teacher base salary is likely to be much larger than $1,000 (a $1,000 increase would represent less than 2% of the average public school teacher salary in 2015). As our research also shows, increases have larger effects at lower levels of base salary: i.e., increases would be particularly helpful in keeping novice teachers in schools long enough to gain critical experience, and to retain teachers in high-poverty schools where teacher salaries are lower and the shortages are the most acute.

Moreover, in addition to their positive effects on recruitment and retention, higher salaries could strengthen the workforce’s credentials and potentially help diversify the workforce. Higher salaries would provide incoming teachers from traditionally underrepresented groups, and those who take on substantial student debt to finance their bachelor’s degree, with a career option that is financially more attractive and that will enable them to more easily pay off that student debt.27

2) Enact higher increases to teacher base pay in high-poverty schools

Due in large part to a heavily local public education funding system that channels more money into districts serving children from higher-income families, teachers in high-poverty schools face a double disadvantage: They are not only paid less than comparable workers in nonteaching jobs (Allegretto and Mishel 2019; Allegretto 2019), but also are paid, on average about $5,600 less than teachers in low-poverty schools—a gap representing close to 10% of teachers’ base salary in 2015‒2016 (García and Weiss 2019c). Because across-the-board raises would thus still leave teachers in high-poverty schools worse off, on average, relative to teachers in low-poverty schools, additional, targeted raises are needed to equalize salaries across high- and low-poverty school districts.

Other targeted raises—such as compensation for teachers who teach in high-poverty schools or a rural school or choose to teach math—would help address teacher shortages in specific settings or for specific subjects. Such increases for teachers who meet certain criteria would not eliminate the teaching pay penalty (only a generalized increase can) but could—especially in the short run—be seen as a first step28 to increasing the attractiveness of teaching in contexts in which the shortages are most worrisome.29 These kinds of targeted raises could still help strengthen collaboration and other aspects of the teacher workforce in a way that merit pay increases and other incentives-based strategies do not.30

3) Adequately fund pension benefits and remove obstacles to accessing them

Pensions have been a critical factor in both attracting teachers to the profession and retaining them as teachers, and to improving their retirement security (Morrissey 2017; Keefe 2018; Rhee and Joyner 2019). But there are hurdles to accessing these benefits—including portability across states (Rhee and Joyner 2019), cuts during the last recession (Doherty, Jacobs, and Madden 2012), and more general underfunding (Pew 2019)—which diminish their positive contribution to teachers’ total compensation and teacher retention (Morrissey 2017).

It is important to readdress the changing needs of the current pension systems, and move away from misguided descriptions of their burden, generosity, and coverage.31 In order for the existing pension system “to serve the goals of attracting and retaining teachers, promoting orderly retirement, and providing retirement security” we must safeguard them by ensuring transparency and adequate investments, and by removing obstacles to portability and vesting (Morrissey 2017).

4) Consider programs that reduce the major financial burdens that are barriers to entering and staying in the teaching profession

By increasing teachers’ capacity to repay their student debts and handle other expenses, higher salaries would address key barriers to entering and remaining in the teaching profession. Other policies that address major expenses burdening teachers financially are also worth consideration, particularly if designed to help teachers that make a long-term commitment to the profession.32

Given the high cost of quality teacher preparation programs, initiatives such as service scholarships and student loan forgiveness programs can support recruitment. Indeed, the majority of states have established some sort of service scholarship and/or loan forgiveness program. These programs “underwrite the cost of teacher preparation in exchange for a number of years of service in the profession. Research has found that effective service scholarship and loan forgiveness programs leverage greater recruitment into professional fields and locations where individuals are needed, and support retention” (Espinoza et al. 2018).

Research assessing such programs, which have long been supported by federal and state governments, shows them to be effective in attracting and preparing candidates to meet the demand for teachers in certain settings (Podolsky and Kini 2016; Podolsky et al. 2019; Feng and Sass 2015).33 Effective program models emphasize preparation, offer a sufficiently substantial award (i.e., covering all or a large percentage of a student teacher’s tuition), target high-need fields and/or locations, recruit and select candidates who are academically strong and committed to teaching, encourage recipients to teach for a number of years (i.e., the financial consequences if recipients do not fulfill their commitment are strong but not so punitive that students avoid the scholarship entirely), and are bureaucratically manageable for participating teachers, districts, and higher education institutions (Podolsky et al. 2019).

A recent initiative designed to increase the number of teachers in officially designated “TSAs” (i.e., teacher shortage areas) is the TEACH grant, which offers $4,000 per year to undergraduate and graduate students who enroll in coursework to become a teacher in a high-need field and commit to teach for at least four years (Barkowski et al. 2018).34 Nearly one in five TEACH grant recipients said that the grant was very influential in their decision to pursue teaching as a career, and almost a quarter, 23%, said the grant strongly influenced their decision to pursue teaching in a high‐need field at a high-need school.35

One recent legislative proposal to extend loan forgiveness to teachers is the Supporting the Teaching Profession Through Revitalizing Investments in Valuable Educators (STRIVE) Act, which would provide incremental loan forgiveness each year to public school teachers who teach in low-income schools, and cancel student loans completely after seven years. The STRIVE Act, which was introduced by Sen. Cory Booker (D-NJ), “would be retroactive, so current teachers who have been teaching for at least seven years would also have their loans canceled.” (Booker 2019). A number of initiatives—designed not to expire—that complement national-level initiatives such as the Perkins Loan Forgiveness and Teacher Loan Forgiveness programs have been implemented in Connecticut, Illinois, Oklahoma, Tennessee, Virginia, Washington, and Wisconsin (American Federation of Teachers 2019).

A non-exhaustive look at some initiatives implemented or underway reveals several types of housing-related interventions. For example, in many locations, especially cities, high housing costs and the lack of available affordable housing make teaching even more burdensome and unattractive. Local initiatives in Massachusetts and New Jersey offer reduced interest rates, or fixed rates, and 100% financing for teacher and other municipal employee homebuyers (Cisneros et al. 2007). In states such as Mississippi, Illinois, and Maryland, teachers who commit to teaching for a number of years are offered grants to cover a down payment, closing costs, mortgage insurance, or prepaid items for buying a home. In rare cases, where the lack of housing is particularly severe, districts have built dedicated housing for teachers, including in McDowell County, West Virginia, and several locations in California (Balingit 2019; Picchi 2019; Lambert and Willis 2019; and Richards and Wynn 2019).

5) Acknowledge and take steps to address other financial burdens that arise when teachers in under-resourced schools must take on safety net roles

In some districts, teachers are increasingly called on to serve as first responders when it comes to children’s basic needs, whether it is connecting them up with laundry services or a place to shower (Weiss and Reville 2019; Kirp 2019; Da Costa 2019). That generosity extended by teachers includes filling the gap when schools, districts, and states fail to provide all the needed educational goods. The nation’s K–12 public school teachers personally spend, on average, $459 annually on school supplies for which the overwhelming majority—more than nine out of ten—will not be reimbursed (García 2019).36

Spending on school supplies and students’ basic needs can be considered a (negative) component of the total compensation package. In addition to addressing the inequitable funding system and adequately budgeting for all expenses, districts should ensure that teacher salaries are not further diminished by out-of-pocket expenses they should not be expected to incur.

Elevate teacher voice, and nurture stronger learning communities to increase teachers’ influence and sense of belonging

Teacher influence, autonomy, and the role of teachers in creating learning communities play a key role in the teacher shortage. Comparing teachers who stay in their schools with teachers who quit the profession, we found that larger shares of those who quit reported that cooperation among staff was not great, said they lacked influence over school policy or what takes place in their classrooms, and said that they were not fully satisfied with teaching at the schools than among teachers who stayed, with gaps ranging from 3 to 17 percentage points. We also showed that larger shares of teachers who quit said there was not a great deal of cooperative effort among staff members and said they did not feel strongly supported by the school’s administration (García and Weiss 2019d and 2019e).

The two recommendations that follow seek to address the lack of consideration and respect for teachers’ voices that discourage teachers from remaining in the profession. It is important to note that giving teachers a greater say and increasing cooperation among teachers are also foundational to establishing positive working environments and enhancing professional development, which are addressed in subsequent sections of this report. The fact that the factors behind the teacher shortage straddle several policy buckets underscores our argument that the factors are so closely associated and interwoven that they must all be addressed comprehensively.

1) Increase teacher autonomy and influence

Our analyses showed that only small shares of teachers report having a great deal of influence or control over school policy. A scant 3.2%, for example, report having a great deal of influence over how teachers are evaluated; only 5.3% have a great deal of influence in hiring new teachers; and only 8.9% have a great deal of influence setting discipline policy. The shares of teachers reporting a great deal of influence setting performance standards for students or establishing the curriculum are higher, but at 17.6% and 20.4% respectively, still are at or under one in five (García and Weiss 2019d). This alone should put those concerned about teachers on notice: setting the curriculum is among the most fundamental responsibilities of teachers, and the fact that four in five lack such authority speaks volumes about how much teachers’ voices have been ignored.

Although teachers report much more influence in individual classroom planning and teaching than on schoolwide policies, they still indicate a surprisingly low level of control over their daily activities. The shares of teachers who report a great deal of control is 61.2% when it comes to evaluating and grading students and is 68.3% when it comes to assigning the amount of homework. Again, the fact that these shares are not closer to 100% is certainly surprising. Also concerning is the much lower share (under 30%) of teachers who report a great deal of control selecting textbooks and other instructional materials and selecting topics and skills to be taught. And just 11.1% of teachers report having a great deal of influence in determining the content of professional development programs (García and Weiss 2019d). All these findings from García and Weiss 2019d indicate very low regard for teachers’ knowledge and judgment. Teachers sense that they are not treated as professionals, and their increasing exit from education should thus come as no surprise. All of these findings point to the need for urgent and comprehensive changes to how we treat teachers and support their profession.

Given this feedback, a comprehensive strategy to address the teacher shortage must ensure that teachers have a say in the components of teaching that they are trained to master and that shape their daily activities and their professional lives. These components include the curriculum they teach, the classroom practices they follow, and the materials they use with their students, as well as the type of professional development they participate in. Their expertise should also be tapped when decisions about school policies are made: Research has pointed to the need for efforts to retain more experienced teachers by giving them “shared decision-making roles” (Sorensen and Ladd 2020). Top-down policies that ignore teacher expertise, misguided accountability policies that make teachers feel disrespected, and lack of attention to what teachers have to say about the policies in their schools and classrooms are critical obstacles in the way of the professionalism of the teaching profession.

The regression analyses included in Appendix Table 2 confirm the practical and critical importance of giving teachers a greater voice. Reporting more influence on school policy and classroom activities is associated with an increase in the probability that a teacher stays at his or her school.37 The effect of influence on a teacher’s probability of staying is higher in high-poverty schools than in low-poverty schools. As with cooperation and support (see below), measures of influence are among the largest predictors of retention.

2) Nurture stronger learning communities through acknowledging and fostering teacher collaboration

The majority of teachers are not working in the kinds of learning communities that would support their teaching and their career growth. The data from teacher surveys analyzed in García and Weiss 2019d clearly signal limited opportunities to cooperate and coordinate, and weak learning communities surrounding teachers. Less than half of teachers strongly agree that their administrations’ behavior is supportive and encouraging (49.6%) or that there is a great deal of cooperative effort among the staff members (38.4%). One in 20 teachers (4.9%) say that the stress and disappointments involved in teaching are not worth it; more than one-fourth of teachers say they think about leaving teaching at some point (27.4%); nearly half express some level of dissatisfaction with being a teacher in their school (48.7%); and more than half say they are not certain that they would still become teachers if they could go back to college and make a decision again (57.5%).38

School districts must heed this feedback and change policies and procedures to ensure that teachers have sufficient time for collaboration, cooperation, observation, and feedback. Peer learning opportunities must be incorporated into classroom and schoolwide operations. Teachers are more likely to stay in their schools and in the profession when they are satisfied with the school and its management. Quint (2011) emphasizes the idea of “a broader conception of teaching learning that involves all teachers in a school in a professional learning community that is engaged in a continuous and collegial cycle of learning, practice, reflection, and improvement.” These collaborative practices, if made systematic, would also serve as a way to build career ladders for teachers that leverage their experience and expertise.

The regression analyses included in Appendix Table 2 confirm the practical and critical importance of a more cooperative and supportive learning community. Reporting a greater level of cooperation and support is associated with an increase in the probability that a teacher stays at his or her school.39 The effect of cooperation and support on a teacher’s probability of staying is higher in high-poverty schools than in low-poverty schools. As we saw around the “influence index,” measures of cooperation and support are among the largest predictors of retention.

Lower the barriers to teaching that affect teachers’ ability to do their jobs and their morale

Another factor behind the exodus of teachers from the profession and the shrinking supply of future teachers is teachers’ working environments (or school climates). These difficult working environments reflect the poor funding systems and the lack of resources and supports available to students and schools (described above), as well as the lack of consideration for teachers’ professional judgment, disregard of teachers’ voices, and poor learning communities. These difficult school climates are byproducts of broader problematic societal forces—poverty, segregation, inequality, and insufficient public investments—that administrators and policymakers have failed to tackle. Rather than expecting educators, students, and families to bear the burden of the challenging conditions (described below), officials charged with overseeing these systems must assume responsibility for remaking school climates.40 Among the factors that make school climates so challenging are widespread barriers to teaching and learning and extensive threats to teachers’ emotional well-being and physical safety.41 Students are coming to school unprepared to learn (as reported by 27.3% of teachers), and parents are struggling to be involved (as reported by 21.5% of teachers), conditions that are exacerbated by greater needs among our students and insufficient supports to them in schools and in their communities.42 Teachers report stress and a lack of safety. More than one in five teachers (21.8%) report that they have been threatened and one in eight (12.4%) say they have been physically attacked by a student at their current school (García and Weiss 2019d).43

The recommendations that follow seek to address the challenging working conditions—many springing from a lack of investment in schools and communities, as well as in children and their families—that make it extremely difficult to persuade young people to choose teaching as a career and that discourage teachers from remaining in the profession.

1) Hire support personnel with the right qualifications to help mitigate barriers to teaching and learning

Often, teachers wear far too many hats over their main one as educators: they are asked to be first responders, social workers, physicians, counselors, and nurses, especially when schools offer insufficient numbers of these professionals. Several interconnected interventions are needed to meet children’s needs so that so much does not fall on teachers and add to the difficulties they have doing their jobs. These supports include investing in more counselors, nurses, librarians, and paraprofessionals, all of whom make schools healthier, more enriching places, and reduce behavioral issues (see below); and investing in strategies to meaningfully engage parents and families.44 Before entering education professions and while on the job, administrators, teachers, and support personnel would benefit from receiving targeted training in trauma-informed practices, restorative practices, culturally responsive pedagogy, community schools strategies, or reflective discipline. School administrators can also play a significant role in promoting more positive school climate outcomes.45

Mitigating the barriers to teaching and learning also requires ensuring the physical spaces are safe. School districts and policymakers must identify buildings with infrastructure problems that have led to unsafe or unhealthy schools and come up with plans for the necessary upgrades. Investing in these upgrades would enhance the working environments and would further minimize the shortage of teachers.46

2) Revisit disciplinary policies

There are several reasons why student behavior problems may lead to tense relationships and even threats to teachers’ physical and mental health. Effectively reducing these problematic situations requires, first, understanding their drivers. For example, in some cases, children may act out due to trauma or toxic stress that originated outside of school. In other cases, the reactions and aggressions may result from children’s disengagement or disconnection from school. Each of the drivers demands a very different response.

In all cases, adequate supports for children will reduce unsafe episodes in school and make it much easier for teachers to teach and for children to learn effectively. In school, these supports include investments in school support personnel (mentioned above) and supportive policies. Evidence points to the efficacy of shifting from so-called zero-tolerance to preventive and supportive policies—“restorative” approaches such as peer mediation, group responsibility, and counseling that support and promote safe learning environments. Current disciplinary measures that focus mainly on punishing wrongdoing are unlikely to improve school climate. Rather, these measures should be rooted in schools’ ability to support and promote better behavior, and to prevent misbehavior, which are much more effective ways to keep teachers safe physically and emotionally (García and Weiss 2016; García 2014). In addition, shifting to restorative approaches benefits minority children because they are disproportionately more likely to be impacted by school disciplinary policies (RPWG 2014).

Sustained lack of attention to these realities has real consequences for teachers’ odds of staying in the profession (see summary in Katz 2018). The added analyses in Appendix Table 3 also shows that facing higher stress and threats to safety, as measured by the “Stress and safety index,” decreases the probability that a teacher stays at her school a year later by about 4 percentage points overall, with greater reduction in high-poverty schools than in low-poverty schools (columns 4-6).47

Design professional supports that strengthen teachers’ sense of purpose, career development, and effectiveness

Teachers who receive strong professional supports will be more able to share their know-how, exert positive influence, and grow learning communities. But the systems of professional supports and development currently offered to teachers offer significant room for improvement. While some supports are broadly available, there are multiple weaknesses to address if we want to help teachers do their jobs better and advance in their careers (García and Weiss 2019e). Making the systems of supports effective and meaningful requires all of the following: that professional development opportunities are widely available early and during teachers’ entire careers, that all teachers have financial resources and time allotted to participate, and that the opportunities are targeted to their particular needs and are of the necessary quality. One without the others will only lead to an unsatisfactory system of supports in need of further fixing.

On the positive side, large shares of first-year teachers work with a mentor (79.9%) and participate in teacher induction programs (72.7%), and the vast majority of teachers of all experience levels access certain types of professional development such as workshops or training sessions (91.9%) and activities focused on the subjects that teachers teach (85.1%) (García and Weiss 2019e). On the negative side, the seemingly widespread availability of basic professional development opportunities is negated by most teachers’ inability to actually take advantage of them. Novice and veteran teachers largely do not have the time and resources they need to study, reflect, and prepare their practice. Only a minority of first-year teachers are released from classroom instruction to participate in support activities for new or beginning teachers (37.1%) or have aides to enhance classroom management and one-on-one attention for students (26.9%). Moreover, among all teachers, only half have released time from teaching to participate in professional development (50.9%), less than a third are reimbursed for conferences or workshop fees (28.2%) or receive a stipend for professional development accessed outside of regular work hours (27.3%), and only one in 10 teachers (9.4%) receives full or partial reimbursement of college tuition.

Moreover, teachers have very limited access to the kinds of professional development that are highly valued and more effective: Just slightly over one-fourth of teachers attend university courses related to teaching and less than one-fourth serve as workshop presenters or make observational visits to other schools. And barely 1 in 10 teachers report having a great deal of influence in determining the content of professional development programs (11.1%).48 This is in contrast with what teachers report wanting and with the norm in other professions—a more direct role in selecting the types of professional development offered and the content of that professional development (Quint 2011; Warner-Griffin, Cunningham, and Noel 2018; Schwartz 2019; Ingersoll and Collins 2018; OECD 2016, 2019). Given the critical importance of establishing a proper system of supports, we offer the following recommendations.

1) Ensure that teachers have access to coherent, high-quality, lifelong systems of supports, and that they are engaged in designing these systems

A proper system of professional supports—early career supports and meaningful continuous professional development—is key to profession-building and to effective teaching. Strong professional supports guided by educators enhance teachers’ practices, satisfaction, and sense of purpose; contribute to the professionalization of the teaching profession; and advance career development. In such a knowledge-based profession, in which the demands are constantly changing and teachers must continually adapt their knowledge and practice, adequate and effective professional development are critically important. All in all, adequate career supports elevate the prestige and appeal of teaching for novice and experienced teachers, and will thus help reduce the shortage of teachers.49

Though there is a network for the provision of these supports in most schools, that network needs substantial improvement. The current system of supports must be made more coherent and purposeful, program quality must be improved, and teachers must have access to the resources that enable them to participate in the system of supports.

For novice teachers, the system of supports should ensure their early preparation and improve their adaptation to the profession and their schools in the first years.50 One effective initiative for beginning teachers and teachers-in-training is a teacher residency (a program that provides new teachers with a year-long apprenticeship teaching alongside an expert mentor teacher).51 Also effective are mentoring and induction programs (programs in which the mentor offers orientation, useful knowledge, and skills to the mentee through modeling, feedback, and other support for the mentee).52 Residencies and mentoring and induction programs all combine to facilitate young teachers’ adaptation to the profession, foster cooperation and collegiality, and strengthen teachers’ professional knowledge and skills. These programs also would improve teacher retention: as our data show, the probability that teachers stay in the profession increases if teachers participate in teacher mentor or teacher induction programs.53

Evaluations of teaching apprenticeships have demonstrated a positive influence on both the mentees and the mentors, as measured by the performance of their students, especially later in teachers’ careers and especially in math (Goldhaber, Krieg, and Theobald 2018a and 2018b; Papay et al. 2016).54 Existing residency programs shown to enhance retention, effectiveness, and teacher diversity include the Boston Teacher Residency,55 the Denver Teacher Residency,56 the San Francisco Teacher Residency Program, and teaching residency programs funded through the U.S. Department of Education’s Teacher Quality Partnership grants program (Papay et al. 2012; Eisner et al. 2015; Guha, Hyler, and Darling-Hammond 2016; Silva, McKie, and Gleason 2015).

Throughout teachers’ careers, continued professional learning and access to continuous preparation and to meaningful career ladder options are of critical importance.57 Teachers must have resources (in both financial resources and time allotted to participate) and access to continuous training opportunities that are relevant and effective, and teachers should participate in determining the package of offerings.

Researchers note that effective professional development programs are content-focused, they support collaboration and job-embedded practice,58 they are intensive and of sustained duration, they focus on discrete skill sets, they offer opportunities for feedback and reflection, and they are characterized by active learning and collaboration (Darling-Hammond, Hyler, and Gardner 2017; Kraft, Blazar, and Hogan 2018; OECD 2019).59 Evidence also shows that curriculum development and lesson study, teacher research, teacher-led professional development (i.e., professional development that is more self-directed by teachers and more actively informed and overseen by them), and appraisal and feedback are key components of solid systems of professional supports (Darling-Hammond et al. 2017).60 Across all of these programs, educators should provide crucial input in their design and be part of the solution.61

Our regression results also confirm how these systems of professional supports would substantially improve teacher retention rates, especially in high-poverty schools (Appendix Table 5). More resources (time and economic resources), more professional development, and being more satisfied with the professional development opportunities are associated with increased retention in the school (associations vary between about 2 to 5 percentage points, with generally larger coefficients for regressions run for high-poverty schools).62

2) Provide teachers with the option of meaningful second jobs that offer career advancement, not just survival

Beyond accessing professional development programs, teachers can enhance their professional growth by taking on additional roles and responsibilities within the school system. These meaningful “second jobs” include mentoring or coaching other teachers, teaching evening classes, or leading teaching induction programs. (Note that while these activities do constitute an additional source of pay for teachers, they do not replace the need for stronger pay in the professions, and they are most relevant because they provide meaningful additional incentives to teachers and have a large set of positive repercussions.)63 For example, more senior teachers can play leading roles in mentoring and induction programs. Reducing churn and increasing the number of experienced teachers helps peer-to-peer mentoring and collaboration.64 All of these within-school activities advance the positive professional development cycle described above; increased staff stability and the building of a more experienced workforce in turn improve the ability of the school to carry through on plans and medium-term strategies. At the core, lower turnover and attrition would thus help to not only build effective learning communities, but to make teaching a better respected and broadly valued profession.

Our regression results emphasize that teachers who work second jobs in their schools value the professional support and career- or profession-building aspect of those opportunities, and that these activities boost their retention (above and beyond the financial boost they provide). For example, we find positive retention effects of working inside the school system, and positive effects of the extra compensation received. For teacher retention, engaging in the activities seems to matter more than the amount of income received from the activities, particularly in high-poverty schools.65 Thus when designing policy, it is important not to offer these activities as a substitute for raising salaries.66

Conclusion

There is no question that the agenda that we have laid out is ambitious, but it is intentionally so.

The specific policy recommendations are designed to address, and sometimes, redress, the negative roles the different factors—low relative pay, the challenging school environment, the lack of respect and recognition for teachers’ knowledge and judgment, and uneven access to and resources for useful types of professional development—play in the teacher labor markets, and the severe, costly consequences the factors have in the education system. The overarching policy recommendations also address the context in which teachers operate and how to improve the context so that the teaching workforce crisis would lessen. The persistent inaction on this collective set of issues, accompanied by pressing needs for more (not less) teachers led up to a national teacher shortage crisis. We have shown during this series of reports that the teacher shortage crisis was “worse than we thought” not only because of the spiking shortage of highly qualified teachers, especially in high-poverty schools, but also because of the lack of attention and unacceptable inaction on the shortage and, more broadly, on public education.

This sixth and final report in our series is calling for immediate policy steps that are not only necessary but eminently doable. For one, several states and several peer nations have already implemented solid models for many of these steps, which could serve as model interventions in particular districts or states. Every school district and state is different but multiple examples show that we can take concepts and translate them into policy and action. In addition, we can expect that research, through more controlled experiments, and a broadened look at all factors prompting teachers to quit and dissuading people from entering the profession would provide further guidance informing additional policy steps for tackling the teacher shortage.

If this comprehensive policy agenda is followed, it would confront the teacher shortage in the nation’s public schools. Furthermore, in observation of the ongoing realities posed by the COVID-19 pandemic, the proposed agenda would strengthen the foundations and protections of our education system so that the next time we enter a crisis—in the system, in the country, or around the globe—our education system is in a “better than we could have expected” position to help children and educators emerge stronger.

About the authors

Emma García is an education economist at the Economic Policy Institute, where she specializes in the economics of education and education policy. García’s research focuses on the production of education (cognitive and noncognitive skills), evaluation of educational interventions (early childhood, K–12, and higher education), equity, returns to education, teacher labor markets, and cost-effectiveness and cost–benefit analysis in education. She has held research positions at the Center for Benefit-Cost Studies of Education, the Campaign for Educational Equity, the National Center for the Study of Privatization in Education, and the Community College Research Center; she has consulted for MDRC, the World Bank, the Inter-American Development Bank, and the National Institute for Early Education Research; and she has served as an adjunct faculty member at the McCourt School of Public Policy, Georgetown University. García received her Ph.D. in economics and education from Columbia University Teachers College.

Elaine Weiss is the lead policy analyst for income security at the National Academy of Social Insurance, where she spearheads projects on Social Security, unemployment insurance, and workers’ compensation. Prior to her work at the academy, Weiss was the national coordinator for the Broader, Bolder Approach to Education (BBA), a campaign launched by the Economic Policy Institute, from 2011–2017. BBA promoted a comprehensive, evidence-based set of policies to allow all children to thrive in school and life. Weiss has authored and co-authored EPI and BBA reports on early achievement gaps and the flaws in market-oriented education reforms. She is co-author, with former Massachusetts Secretary of Education Paul Reville, of Broader, Bolder, Better, published by Harvard Education Press in 2019. Weiss came to BBA from the Pew Charitable Trusts, where she served as project manager for Pew’s Partnership for America’s Economic Success campaign. She has a Ph.D. in Public Policy from the George Washington University Trachtenberg School and a J.D. from Harvard Law School.

Acknowledgments

The authors are grateful to EPI Publications Director Lora Engdahl for her extraordinary contributions to this series of reports. We appreciate her help structuring the contents of the series and her edits to the six reports in the series, the summary report, and the several accompanying infographic and dissemination pieces. We want to express our gratitude for her help ensuring the consistency of the contents in each of the reports as well as across the six installments of the series. We also appreciate EPI Vice President John Schmitt’s supervision of this project and EPI President Lawrence Mishel’s guidance in earlier stages of this research. We acknowledge EPI Research Assistant Daniel Perez for his assistance with the tables and figures in this report and EPI’s communications staff for their contributions to the different components of this report and the teacher shortage series including infographics, assistance with media, proofreading, technical features, and coordination of all the steps required for the publication of all the reports of the series.

Appendix tables

Appendix Figure A

Appendix Figure B

Appendix Figure C

Appendix Figure D

Appendix Table 1

Appendix Table 2

Appendix Table 3

Appendix Table 4

Appendix Table 5

Endnotes

1. According to Sutcher, Darling-Hammond, and Carver-Thomas (2016), the gap between the number of qualified teachers needed in the nation’s K–12 schools and the number available for hire in the 2017–2018 school year was about 110,000 teachers, up from 54,000 in school year 2015–2016 and from no shortage earlier in the decade. Carver-Thomas and Darling-Hammond (2017) and the Learning Policy Institute (2017) estimate that filling a vacancy costs $21,000 on average, and Carroll (2007) estimates the total annual cost of turnover at $7.3 billion per year. According to Strauss (2017), that estimated annual cost of turnover would exceed $8 billion at present. A lack of sufficient, qualified teachers threatens students’ ability to learn (Darling-Hammond 1999; Ladd and Sorensen 2016). Instability in a school’s teacher workforce (i.e., high turnover and/or high attrition) negatively affects student achievement and diminishes teacher effectiveness and quality (Ronfeldt, Loeb, and Wyckoff 2013; Jackson and Bruegmann 2009; Kraft and Papay 2014; Sorensen and Ladd 2020).

2. A teacher is considered to be in a high-poverty school if 50% or more of the students in his/her classroom are eligible for free or reduced-price lunch programs. A teacher is considered to be in a low-poverty school if less than 25% of the students in his/her classroom are eligible for those programs.

3. From the 2011–2012 to the 2015–2016 school year, there were increases in the shares of teachers who were not fully certified (from 8.4% to 8.8%), who had not taken the traditional route into teaching (14.3% to 17.1%), who had five years or less of experience (20.3% to 22.4%), and who did not have an educational background in the subject they were teaching (31.1% to 31.5%). In high-poverty schools, the shares of teachers without these credentials were even higher: 9.9% were not fully certified, 18.9% took an alternative route into teaching, 24.6% had five years or less of experience, and 33.8% didn’t have an educational background in the subject they were teaching. See García and Weiss 2019a.

4. As we documented in García and Weiss 2019b, “turnover and attrition have been increasing over time (Goldring, Taie, and Riddles 2014) and are higher for U.S. teachers than among teachers in other countries (Darling-Hammond et al. 2017).” Teachers see much higher attrition than their peers in most occupations: About 30% of college graduates who became teachers were not in the profession five years later, compared with 14% of pharmacists, 16% of engineers, and 19% of nurses and lawyers (Ingersoll 2014). From 2008 to 2016, the number of people enrolled in teacher preparation programs fell 37.8% and the number of people completing teacher preparation programs fell 27.4% (see Appendix Figure C).

5. Our research compared factors such as pay, working conditions, and professional supports reported by teachers who stayed in the school with those who quit the profession and found greater shares of leaving teachers reported low pay and negative conditions. In this report we include an examination of these associations using a regression framework in this report. See Appendix Tables 1–5 for regression analyses of the correlations between these factors and retention in the school (unlike in the previous reports, we now also include information from teachers who left to teach in another school or were on short-term leave and planned to return to the school). Various meta-analyses by other researchers have also reviewed how some of these factors correlate with teacher attrition (Nguyen et al. 2019; Borman and Dowling 2008). More limited information is available regarding how these factors influence potential teachers’ decisions to enter into the profession.

6. Larger shares of teachers in schools serving large shares of poor students turned over or quit, and those remaining thus tend to work in schools with relatively more unstable workforces than teachers in low-poverty schools.

7. Teachers in high-poverty schools work in more challenging environments, are paid less, have less autonomy and influence in their jobs, and receive weaker supports both early in their careers and throughout their careers. These findings are consistent across all factors examined even though the low- and high-poverty schools in our study are not particularly extreme schools, but rather fairly common schools (i.e., low-poverty schools are those with less than 25% of students in poverty and high-poverty schools are those with more than 50% of students in poverty). Moreover, the resulting professional, personal, and material discomfort for teachers affects not only teachers themselves but their students; as such, low-income students’ classroom experiences are disproportionately disrupted by staff instability or inadequate preparation.

8. See García and Weiss 2019a, 2019b, 2019c, 2019d, 2019e, and 2020b.

9. The student population has continuously increased since 2008 (by 1.2 million, according to NCES 2019c). While the number of teachers in public schools has also increased between 2000 and 2015—by 210,000—it hasn’t done it in a continuous manner. See Appendix Figure A. The number of teachers reached its highest point in 2008, and dropped significantly between 2008 and 2010 as a consequence of the recession (Gould 2017; Gould 2019; Berry and Shields 2017).

10. According to national Schools and Staffing Survey data over multiple years on teachers who had stayed in their schools, moved to different schools, or left teaching altogether from one survey year to the next, the rate of teacher retention decreased by 0.6 percentage points between 2001 and 2013, while teacher turnover (teachers moving to different schools) and teacher attrition (teachers leaving teaching) both increased 0.3 percentage points in that period. See Appendix Figure B. The decrease in the number of potential new teachers is shown in Appendix Figure C, reproduced from García and Weiss 2019b, Figure A.

11. For example, a common theme documented in our series was that the different factors behind the teacher shortage were systematically more acute in high-poverty schools. If we tackled the underlying system-level problems that create worse problems in high-poverty schools, than we may no longer need additional compensation for teachers in high-poverty schools. Or, if teachers were accorded respect and recognition as the professionals they are, then their salaries would be higher.

12. For example, giving teachers bonuses or a one-time chance to weigh in on how they are evaluated and what kinds of professional development they get are not solutions to the interconnected and systemic problems of low relative teacher pay, lack of teachers’ say over working conditions, and inadequate professional development opportunities over the course of a teaching career.

13. For example, educators have shared that policies such as holding teachers accountable for student test scores and the associated excessive focus and pressure on testing, or letting charter schools reject the most difficult to serve students, play a role in making teaching less attractive, but none of the studies or data examined in this report address those factors so we don’t discuss them here. Likewise, the experimental evidence on the impacts of specific policies that have been implemented to tackle some of the drivers is limited because the policies might have been tried in a scattered, rather than cohesive, manner, or may have been tried in only a few locations.

14. It is complex to model the individual and interacting behaviors of the roughly 4 million teachers represented in our data, especially given that the factors driving shortages are interrelated and there may be other factors that are not accounted for in our regression framework. For example, two teachers with very similar characteristics and views may react differently to policy changes for reasons not controlled for in our models.

15. For example, state governments would likely have a much larger role in those states that have a larger attrition rate than others. As another example, the role of teachers unions may be larger under certain circumstances. The literature on the role teachers unions play emphasizes their positive contribution to teachers’ pay, working conditions, school climates, and in elevating teachers’ voices (Allegretto and Tojerow 2014; Allegretto and Mishel 2019; Han 2019; Moore-Johnson et al. 2007; Jones, Bettini, and Brownell 2016; Lyon 2020; NEA 2012). As such, we would plan to explore the role unions and bargaining play in the teacher shortage more in depth in future research.

16. Research finds that 24 states—including half of the states with over a million enrolled students—still spend less on education per student than they were spending prior to the recession a decade ago (Leachman and Figueroa 2019). In most states, high-poverty districts receive either the same amount of funding or less than low-poverty districts, showing a glaring lack of progressivity in the schools’ funding formulae. With respect to how much states would have to spend in order to achieve national average test scores, a majority of states spend only a fraction of estimated requirements especially in the highest-poverty districts, again revealing sharp inadequacies (Baker, Di Carlo, and Weber 2019; Morgan and Amerikaner 2018; Chingos and Blagg 2017; Urban Institute 2017).

17. Appendix Figure D, reproduced from Baker, Di Carlo, and Weber 2019, shows the degree to which states and districts are allocating resources so that districts can achieve national average test scores, by level of poverty in the school districts. In moderate- and higher-poverty districts, the gaps between what is spent and what would be required to achieve at the national level are significant and increase with the level of poverty. In the middle group, districts are spending $1,300 per student less than what would be required (about 88%). In the highest-poverty quintile, the gap exceeds $6,600. This means that, on average, the highest-poverty districts are spending fully one-third less than what they need to in order to deliver a sound education to their students.

18. A recent meta-analysis widens the examination of the drivers of attrition to include principals’ autonomy and comprehensive school reform, among others, but no category representing teachers’ influence, autonomy, etc., appeared in the widened framework (Nguyen et al. 2019). The authors classify the determinants under personal correlates, school correlates, and external correlates (which include broader categories such as teacher evaluation systems, teacher merit pay, school accountability, principal effectiveness, teacher-principal race/gender matching, teacher-student race matching, comprehensive school reform, and research-practice partnership, not explored in earlier studies). See Borman and Dowling 2008 for an earlier meta-analysis focused on teacher characteristics and school characteristics. This reflects how research has overlooked most of these factors, due to lack of data but also lack of acknowledgment of their relevance. See more information in García and Weiss 2020a.

19. As we wrote in the first report in the series and pointed out earlier, “The teacher shortage has serious consequences. A lack of sufficient, qualified teachers threatens students’ ability to learn (Darling-Hammond 1999; Ladd and Sorensen 2016). Instability in a school’s teacher workforce (i.e., high turnover and/or high attrition) negatively affects student achievement and diminishes teacher effectiveness and quality (Ronfeldt, Loeb, and Wyckoff 2013; Jackson and Bruegmann 2009; Kraft and Papay 2014; Sorensen and Ladd 2020). And high teacher turnover consumes economic resources (i.e., through costs of recruiting and training new teachers) that could be better deployed elsewhere. Filling a vacancy costs $21,000 on average (Carver-Thomas and Darling-Hammond 2017; Learning Policy Institute 2017) and Carroll (2007) estimated that the total annual cost of turnover was $7.3 billion per year, a cost that would exceed $8 billion at present (note that this is an estimate of the cost of turnover/attrition, not an estimate of the cost of the shortage). The teacher shortage also makes it more difficult to build a solid reputation for teaching and to professionalize it, further perpetuating the shortage” (García and Weiss, 2019a).

20. The analysis accounted for education, experience, and other factors known to affect earnings. The pay compares weekly compensation, which avoids measurement issues regarding differences in annual weeks worked and thus accounts for summers “off.” For some explanations of some of the reasons why the teaching wage penalty exists (historically low pay, pink-collar occupation, etc.), see Allegretto and Mishel 2019, and Folbre and Smith 2017.

21. In most states, mid-career teachers who head families of four or more are eligible for government subsidies such as subsidized children’s health insurance or free or reduced-price school meals (Boser and Straus 2014).

22. See Erasmus 2019; Long 2019; Nieto 2019; Reilly 2018; Talley et al. 2018; Zdanowicz 2019.

23. We remind the reader that a teacher is in a high-poverty school if 50% or more of the student body in his or her classroom is eligible for free or reduced-price lunch programs.

24. The relative qualifications of the teaching workforce refer to the credentials of teachers in the teaching workforce.

25. The research indicates that fewer people are willing to make the choice to be in a profession that puts them at a financial disadvantage, a disadvantage that is known as raising the opportunity cost to be (stay or become) a teacher (see Loeb and Page 2000 and Murnane and Olsen 1989, and summary in García and Han 2019). In general, lower salaries correlate with more attrition and higher salaries correlate with retention and with more interest in becoming teachers (see Podolsky et al. 2019 and Katz 2018 for higher rates of turnover in schools with lower salaries, across a number of contexts). García and Weiss (2019c) compared characteristics of teachers who quit their jobs with teachers who stayed at their schools and found that the quitting teachers had received, on average, lower salaries, participated less in the kinds of paid extracurricular activities that complement professional development (activities like coaching students or mentoring teachers), and engaged more in working options outside the school system than did teachers who stayed. A broad consensus has emerged around the evidence that higher pay should be part of the solution to problems in the teaching profession and labor markets (see Podolsky et al. 2016; Espinoza et al. 2018; PDK 2018; Croft, Guffy, and Vitale 2018; García and Weiss 2019b).

26. Existing research finds higher rates of turnover in schools with lower salaries, across a number of contexts (see summaries in Podolsky et al. 2019 and Katz 2018). In general, lower salaries correlate with more attrition and higher salaries correlate with more interest in becoming teachers. Recent research using international data shows that countries that pay teachers more “tend to draw their teachers from higher parts of the college skill distribution” (Hanushek, Piopiunik, and Wiederhold 2019). Research also indicates that when states have raised and equalized salaries (at the same time as they have raised standards for preparation), they have increased quality and retention as well (Podolsky et al. 2016).

27. The problem of low diversity in the teaching profession, both by gender and by race, and its potential connection to the teacher shortage is an increasingly dominant part of education policy discussions (Ingersoll 2015; Ingersoll and May 2011; Ingersoll, Merrill, and Stuckey 2014; Carver-Thomas 2018). A recent study explains that the teaching pay penalty represents a particular barrier for teacher candidates of color because they are more likely to have student loan debt upon graduating and are less likely to pursue lower-salary public interest jobs (Fiddiman, Campbell, and Partelow 2019; see also Darling-Hammond 2019). It is argued that changes in the racial and ethnic composition of the teacher workforce are not keeping pace with changes in the demographic makeup of the student population. The share of K‒12 teachers who are Black, Hispanic, Asian, American Indian, or other nonwhite groups increased from 12.4% to 17.3% from 1987‒1988 to 2011‒2012, while the share of students in those groups increased from 27.3% in 1987‒1988 to 44.1% in 2011‒2012 (Ingersoll 2015). The argument that students’ learning improves when they are taught by teachers of their same race/ethnicity (see Figlio 2017) is used to advocate for increasing the diversity of the teacher workforce at public schools. Also, increasing the share of Latino teachers could boost the share of teachers with the proper qualifications—such as English Language Learner (ELL) training—to meet the needs of the nation’s growing share of students who are Hispanic (Carnoy and García 2017; NASEM 2017; WHIEEH 2015).

28. Though the cost of some of these increases hasn’t been calculated as it is a function of numerous variables, it is reasonable to estimate that the total investment for an across-the-board, significant, salary increase would exceed the investment required to give targeted raises. Still the targeted raises may have significant costs too: Dee and Goldhaber (2017) note that “it might be necessary to offer quite large monetary incentives to induce teachers to take positions in hard-to-staff schools.”

29. The literature sometimes argues that shortages are not general or universal but that they particularly affect some subjects/specialties, schools, groups of teachers, or even states (see Dee and Goldhaber 2017). For example, shortages are reported worse for STEM (science, technology, engineering, and math) and special education teachers (Ingersoll, Merrill, and Stuckey 2014; Dee and Goldhaber 2017; U.S. Department of Education 2017c, 2019). Shortages are also reported worse for so-called hard-to-staff schools, mainly high-poverty, high-minority, and rural schools. Sources discussing worse shortages for teachers at high-poverty schools include Sutcher, Darling-Hammond, and Carver-Thomas 2016; Podolsky et al. 2016, Loeb, Darling-Hammond, and Luczak 2005, Ingersoll, Merrill, and Stuckey 2014; Darling-Hammond 2010; Simon and Johnson 2015; Loeb, Darling-Hammond, and Luczak 2005; Sutcher et al. 2016; Podolski et al. 2016. Sources discussing urban teachers’ retention include Papay et al. 2017; while Ingersoll 2015; Ingersoll and May 2011; and Carver-Thomas 2018 discuss shortages for minority teachers. Levin et al. (2015) find teacher surpluses in Massachusetts; Berg-Jacobson and Levin (2015) find teacher deficits in Oklahoma; Keefe (2018) notes teacher shortages in Pennsylvania. Officially, the Department of Education occasionally publishes the “States’ Reports of Teacher Shortage Areas” (TSAs) (U.S. Department of Education 2017c, 2019). A summary of different direct and indirect metrics used to measure a shortage of teachers is available upon request. See also Behrstock-Sherratt 2017; Ingersoll 2015; Guarino, Santibáñez, and Daley 2006; Ingersoll 2004; Sutcher, Darling-Hammond, and Carver-Thomas 2016; Podolsky et al. 2016; Loeb, Darling-Hammond, and Luczak 2005; Ingersoll 2001; Moore-Johnson, Kraft, and Papay 2012; Viadero 2002.

30. Merit pay raises, i.e., raises based on some selected (and often incomplete) measure of teacher contribution to student learning) are not recommended for the goals discussed in our series. Importantly, by design, they are only offered temporarily and to subgroups of teachers, failing to meet the foundational principle that the teacher shortage is a problem in need of multifaceted, long-term solutions. Research indicates that the effects of merit pay for teachers based on student performance fade once the incentives end (Glazerman and Seifullah 2012). Also merit pay is based on promoting “competition” among teachers as a strategy to boost their productivity, which can induce perverse incentives by diminishing collaboration and peer supports (Gius 2013 etc.). Finally the effectiveness of merit pay relies on strict assumptions about its components and functioning (Imberman 2015).

31. Teacher pensions account for a larger share of total expenditures in education now than they did in the past, not because of their increased generosity but because of two negative trends that make them relatively strong: salaries have been decreasing and the costs of health care have spiked (Baker 2018). Startz (2016) also challenges the idea that pensions are as high or generous as argued given that not all teachers will become eligible for a pension and that a large share of teachers are not covered by Social Security. Salaries and benefits combined have accounted for about the same share, or a slightly smaller share, of total expenditures for public elementary and secondary education and other related programs as was the case in the 1990s (NCES 2017).

32. Care must be taken when comparing the costs and benefits of service incentives to ensure that resources are not redirected to those who are not making long-term commitments to the profession.

33. According to Feng and Sass (2015), loan forgiveness and service scholarship programs date back to 1958, when the National Defense Education Act created National Defense Student Loans. Since then, the federal government and more than 40 states have at various points offered loan forgiveness and/or service scholarship programs to individuals interested in teaching. While the student debt problem is not unique to teachers, research focused on the overwhelming weight of student loans suggests that teachers who receive loan forgiveness are more likely to remain in the profession (see Podolsky et al. 2019, citing Feng and Sass 2015). Podolsky et al. (2019) also highlight the role of these programs in increasing their overall compensation.

34. Other federal student financial aid programs are available for those teaching in a TSA. See Strauss 2017 and U.S. Department of Education 2017c.

35. In 2019, a news report found that the program’s opaque administrative requirements caused a large number of teachers who should have benefited from the aid to, instead, see their grants converted into loans, some with high interest rates, compounded by extensive red tape (Arnold and Turner 2019). Implementing the best practices of effective program models is thus key for such programs. Overall, close to 270,000 grant recipients have benefited since the program was launched (AACTE 2017).

36. Adjusted for inflation to 2018 dollars, according to the NCES 2011–2012 Schools and Staffing Survey (SASS). This figure does not include the dollars teachers spend that are reimbursed by their school districts. The $459-per-teacher average is for all teachers, including the small (4.9%) share who do not spend any of their own money on school supplies.

37. We construct an “Influence index” that captures the degree of influence teachers report having over a range of 11 factors, from establishing the curriculum at the school to selecting textbooks and other instructional materials. An increase of one standard deviation in this index is associated with a 4.3 percentage-point increase in the probability the teacher is still in the same school the following year after the teacher reported their level of influence on the factors. The 11 factors are establishing curriculum; setting performance standards for students; determining the content of in-service professional development programs; setting discipline policy; hiring new full-time teachers; evaluating teachers; determining the amount of homework to be assigned; evaluating and grading students; disciplining students; selecting contents, topics, and skills to be taught; and selecting textbooks and other instructional materials.

38. All data are from NCES 2015–2016 except for the share of teachers who are not sure they would become teachers if they could start over again, which is from NCES 2011–2012.

39. We build a “Cooperation and support index” that captures the degree to which teachers believe that their school environment matches the attributes of a cooperative and supportive environment. An increase of one standard deviation in this index is associated with a 4.4 percentage-point increase in the probability the teacher is still in the same school the following year after the teacher reported their level of agreement with the factors. The seven statements teachers were asked to agree or disagree with were: “The principal knows what kind of school he or she wants and has communicated it to the staff,” “School administration’s behavior is supportive and encouraging,” “I make a conscious effort to coordinate the content of my courses with that of other teachers,” “There is a great deal of cooperative effort among the staff members,” “Most of my colleagues share my beliefs and values about what the central mission of the school should be,” “In this school, staff members are recognized for a job well done,” and “I receive a great deal of support from parents for the work I do.”

40. In some cases, schools are suffering because administrators, superintendents, and boards of education are not enforcing existing codes of conduct, and are dismissing educators’ and other staff members’ inputs and consensus regarding improving school working conditions.

41. Working environments are also shaped by a troubling lack of teacher influence over school policy and over what and how they teach in their classrooms (see below for this point).

42. As the shares of low-income students and of minority students have increased substantially, school districts have not adopted a response sufficient to lift up these new students (NCES 2019a and 2019b).

43. Larger shares of these teachers had reported in the year before they left teaching that they faced significant barriers to teaching than was true among teachers who stayed at their schools. More of the quitting teachers reported feeling stressed (the share of teachers who quit who felt that stress at work was not worth it was 3.5 times greater than among those who stayed). And while teachers who stay are happier than those who leave, the numbers above do not paint a pretty picture about the morale of the current teaching workforce either. Even among those who stayed, over half had reported planning to leave teaching at some point, and nearly half reported being dissatisfied with their jobs. Over a quarter had students who were not prepared to learn, and nearly as many were frustrated by the challenge of engaging their students’ parents (see Figure D in García and Weiss 2019d).

44. Weiss and Reville (2019) explore a dozen diverse communities across the United States that have built comprehensive school-community partnerships to enable supportive, enriching school environments. As they show, such investments pay off in substantial improvements to student achievement, behavior, and well-being; in narrowed achievement gaps; and in improved school climate and morale.

45. For example, under the professional learning communities model groups of administrators, teachers, and staff are dedicated to improving the school culture. See DuFour 2004. For teachers’ training on these different aspects, see Blitz, Yull, and Clauhs 2020; Gay 2002; McInerney and McKlindon 2014; Thomas, Crosby, and Vanderharr 2019; among others).

46. As shown in the regression analyses included in Appendix Table 3, a greater value in the “Barriers to teaching and learning index” (the greater the value, the more barriers the teacher perceives) is associated with a decrease in the probability that a teacher stays at his or her school. The coefficient is higher in high-poverty schools (it is actually not statistically significant in low-poverty schools). The “Barriers to teaching and learning index” is composed by combining the information on variables measuring whether teachers saw certain barriers as causing problems in their schools, based on their responses to the question, “To what extent is each of the following a problem in this school?” The potential responses were “serious problem, “moderate problem,” “minor problem,” and “not a problem.” The barriers were: student tardiness, student absenteeism, student class cutting, student apathy, lack of parental involvement, poverty, students come to school unprepared to learn, poor student health.

47. The “Stress and safety index” is calculated using four variables measuring the share of teachers who are asked to rate their agreement or disagreements with two statements: “The stress and disappointments involved in teaching at this school aren’t really worth it” and “The level of student misbehavior in this school (such as noise, horseplay or fighting in the halls, cafeteria, or student lounge) interferes with my teaching.” The index also relies on the share of teachers who reported that they had ever been threatened or physically attacked by a student from their school.

48. For the statistics in these paragraphs, the data for first-year teachers come from NCES 2015–2016; the data for all teachers come from NCES 2011–2012.

49. Although the link between professional supports and the teacher staffing crisis is less direct than the link between other factors such as school climates and the teacher shortage, these supports are nonetheless critical aspects of the teaching profession. Our data suggest a relationship between the system of professional supports and teacher retention. When we compare teachers who stayed in their school with those who quit teaching, we observe that larger shares of staying teachers had received early support in the form of an assigned mentor (77.0% vs. 69.2%), had found their subject-specific professional development activities very useful (27.4% vs. 19.5%), and had worked in highly cooperative environments (38.7% vs. 33.9%) (García and Weiss 2019e). Strengthened systems of supports have the potential to help teachers do their jobs better, progress in their profession, and gain satisfaction with and a sense of ownership of their careers. These supports are essential to guaranteeing the quality of the teaching workforce and to professionalizing teaching.

50. Large shares of novice teachers report that key aspects of their adaptation to teaching were problematic. For example, at least one in four teachers did not participate in teacher mentoring or induction programs, and only about one in three said that working with the mentor improved their first year teaching by “a lot.”

51. Typically, “residents simultaneously complete credential coursework that is tightly integrated with their clinical placement. Residents are paid a stipend and/or receive tuition remission to enable them to devote the full year to their preparation, and in exchange commit to teach for three to five years in the districts’ schools” (Podolsky et al. 2019). As Podolsky et al. (2019) point out in their report, “studies of the longest-standing teacher residency programs have found higher retention rates of residency graduates” (the studies referenced are Guha, Hyler, and Darling-Hammond 2016; Papay et al. 2012; and Silva et al. 2014). Residencies hold “promise for both recruiting diverse individuals and retaining effective teachers” (Podolsky et al. 2019). As Podolsky et al. (2019) explain, work by Ingersoll, Merrill, and May 2014 using prior SASS data concluded that “new recruits who had a semester or more of practice teaching prior to employment were more than three times less likely to leave teaching after a year than those who had no practice teaching.” Podolsky et al. (2019) continue, “Further analysis found that beginning teachers who had received comprehensive preparation (i.e., observing others teaching; student teaching a full semester; receiving feedback; taking courses in teaching methods, learning theory and selecting instructional materials) were two-and-a-half times less likely to leave teaching after a year in the profession than teachers with little or no pedagogical training. About 37% of first-year teachers had received comprehensive preparation, while about 15% received little or no pedagogical training before entry in 2004-05. Moreover, this research shows that, having courses in teaching methods in addition to student teaching reduced attrition, and, regardless of the coursework package, teachers who received at least one semester of practice teaching were half as likely to leave teaching than those who had not received such training.”

52. Mentoring and induction programs normally entail a “hierarchical relationship in which the mentor is more experienced than the mentee and has useful knowledge and skills that can be conveyed to the mentee through role modeling, feedback, and support” (Goldhaber, Krieg, and Theobald 2018b), and are “designed to support inexperienced teachers in their early years, and to keep those who have potential from leaving the school or the profession before they have a chance to master the art of teaching” (Sorensen and Ladd 2020; Ingersoll and Strong 2011).

53. In producing these estimates we control for a variable that measures how well prepared teachers felt to perform different classroom activities, see Appendix Table 4. The influence of these early supports is larger in low-poverty schools, which in part could can be linked to the greater availability of these supports in these schools or to the intensity or quality of these programs in these schools (Goldhaber, Krieg, and Theobald 2018b).

54. This influence is largely driven by mentors who are “highly effective” in the first place. Goldhaber, Krieg, and Theobald (2018a) explain that in the process of mentoring, the mentor is given the possibility to reflect on her own practices in ways that lead to “self-improvement” and that “peer-learning” can induce improvements in both the mentor and mentee. Goldhaber, Krieg, and Theobald (2018b) emphasize that the effectiveness of the mentee is also driven by the effectiveness of the mentor (as measured by value-added), and that the influence is particularly large in math.

55. Papay et al. (2012) find that graduates of the Boston Teacher Residency program are more likely to remain in teaching in the district in the first five years, rapidly improve their effectiveness, and are more racially diverse.

56. Data examining the first six cohorts of the program showed that graduates were more likely to remain in teaching roles in Denver Public Schools and more likely than other teachers who started at the same time to remain in Title I schools and received slightly higher ratings of their teaching in their first year (Eisner et al. 2015).

57. As discussed in García and Weiss 2019e, while there is a clear understanding that continuous professional supports are incredibly important (indeed, the highest-performing systems in the world provide professional development as part of the regular daily and weekly experience of teaching and continuous training, which are “inextricably linked together,” as noted in Darling-Hammond et al. 2017), there are no universal benchmarks for what an optimal set of professional development activities looks like, and a lot of flexibility may be required.

58. According to Croft et al. (2010), job-embedded professional development “refers to teacher learning that is grounded in day-to-day teaching practice and is designed to enhance teachers’ content-specific instructional practices with the intent of improving student learning” (Croft et al. 2010, citing work by Darling-Hammond and McLaughlin 1995 and Hirsh 2009).

59. The research by Kraft, Blazar, and Hogan (2018) to identify these features of effective professional development builds on Darling-Hammond et al. 2009; Desimone 2009; Desimone and Garet 2015; Garet et al. 2001; and Hill 2007. Darling-Hammond et al. (2017) review the availability of supports in other contexts.

60. As mentioned above, to help ensure that teachers have access to these opportunities, they must be provided with the necessary resources, which include time off for scheduled professional development and reimbursements for time, travel, and other expenses.

61. As cited earlier, just 11.1% of teachers report having a great deal of influence in determining the content of professional development programs (García and Weiss 2019e).

62. Appendix Table 5 presents the relationship between staying in the school and the different supports teachers receive. There are three indices: the “Supports for professional development index”; the “Professional development activities index”; and the “Satisfaction and hours in professional development index.” They are defined as follows:

The “Supports for professional development index” is built using the responses of teachers who received the following types of supports: scheduled time in the contract year for professional development, released time from teaching, reimbursement for conference or workshop fees, stipend for professional development activities that took place outside regular work hours, reimbursement for travel and/or daily expenses, full or partial reimbursement of college tuition, and credits toward re-certification or advanced certification in their main teaching assignment or other teaching field(s).

The “Professional development activities index” is built using information on professional development activities the teacher participated in during the last 12 months. Among the options, there are four standard types of professional development: workshops, conferences, or training sessions where the teacher was not the presenter; workshops, conferences, or training sessions where the teacher was the presenter; university courses related to teaching; and observational visits to other schools. There were also five types of professional development focused on a specific area: content of the subject(s) they teach, use of computers for instruction, student discipline and management in the classroom, how to teach students with disabilities, and how to teach English language learners (ELL). There were three other activities that offer professional development opportunities, in which teachers were asked whether they engaged in research on a topic of interest to them professionally; participated in regularly scheduled collaboration with other teachers on issues of instruction (excluding administrative meetings), and observed, or were observed by, other teachers in their classrooms.

Finally, the “Satisfaction and hours in professional development index” is built using the underlying responses about how useful were the following different professional development activities: offerings that were specific to the subject the teacher teaches, offerings that involve using computers for instruction, offerings that address student discipline and management in the classroom, and offerings that help teachers teach students with disabilities and ELL students. The index also incorporates total hours in specific professional development activities.

The results indicate that an increase of one standard deviation in the supports index is associated with a 1.7 to 3.7 percentage-point increase in the probability that a teacher stays at the school (the increase was not statistically significant in low-poverty schools). An increase of one standard deviation in the professional development activities index is associated with a 2.7 to 5.0 percentage-point increase in the probability that a teacher stays in the school. And being more satisfied with the professional development opportunities also positively correlates with both staying at the school and in the profession (an increase of one standard deviation in the satisfaction and hours in professional development index is associated with a 1.8 to 4.1 percentage-point increase in the probability that a teacher stays in the school).

63. In García and Weiss 2019c, we discussed positive and the negative repercussions of teachers working multiple jobs—or moonlighting in and out of the school system during the school year. We distinguished between teachers who may choose to take on a second job voluntarily versus those who do so due to financial stress, and between moonlighting activities that are career-building opportunities that increase teachers’ sense of belonging and lead to learning communities and moonlighting activities that do not offer these benefits.

64. When too many of the school’s teachers are novices, there are too few veterans available to serve as mentors or to help with induction and avert turnover (Jackson and Bruegmann 2009; Sorensen and Ladd 2020; García and Weiss 2019b).

65. Working in the school system increases the probability that a teacher stays at the same school by 3.6 percentage points, on average, with a larger association in high-poverty schools than in low-poverty schools (4.0 percentage points versus 3.1 percentage points respectively).  Indeed, test-based bonuses, which do serve to substitute for salary increases, negatively correlate with teacher retention, as do jobs working outside the school system, in all schools and low-poverty schools; note that none of these associations are statistically significant.

66. Available upon request, we have also examined the significance of teachers’ satisfaction and motivation with retention. Unsurprisingly, teachers who have higher satisfaction and motivation are more likely to stay in their schools. The association is higher for teachers in high-poverty schools. Because satisfaction and motivation can be the reflection of multiple other factors besides the main channels described in this report, we do not include those results in the appendix.

References

Allegretto, Sylvia. 2019. “For Teachers, the Money Keeps Getting Worse.” Atlantic, September 6, 2019.

Allegretto, Sylvia, and Lawrence Mishel. 2019. The Teacher Weekly Wage Penalty Hit 21.4 Percent in 2018, a Record High: Trends in the Teacher Wage and Compensation Penalties Through 2018. Economic Policy Institute and the Center on Wage and Employment Dynamics at the University of California, Berkeley, April 2019.

Allegretto, Sylvia A., and Ilan Tojerow. 2014. “Teacher Staffing and Pay Differences: Public and Private Schools.” Monthly Labor Review, September 2014.

American Association of Colleges for Teacher Education (AACTE). 2017. “TEACH Grants” (web page). Last updated August 2017.

American Federation of Teachers. 2019. “Loan Forgiveness & Funding Opportunities” (interactive web page). Accessed October 2019.

Arnold, Chris, and Cory Turner. 2019. “Teachers Begin to See Unfair Student Loans Disappear.” Morning Edition (National Public Radio), May 3, 2019.

Baker, Bruce D. 2018. Educational Inequality and School Finance: Why Money Matters for America’s Students. Cambridge, Mass.: Harvard Education Press.

Baker, Bruce D., Matthew Di Carlo, and Mark Weber. 2019. The Adequacy and Fairness of State School Finance Systems: Key Findings from the School Finance Indicators Database, School Year 2015–2016, First Edition. Albert Shanker Institute and Rutgers Graduate School of Education, April 2019.

Balingit, Moriah. 2019. “Rebuilding the Village: A West Virginia School System Strives to Lift Up Its Children by Tackling Poverty.” Washington Post, October 19, 2019.

Barkowski, Elizabeth, Evan Nielsen, HarmoniJoie Noel, Melissa Dodson, Kathy Sonnenfeld, Cong Ye, Elizabeth DeMonte, Brianne Monahan, and Megan Eccleston. 2018. Study of the Teacher Education Assistance for College and Higher Education (TEACH) Grant Program. Prepared for the U.S. Department of Education, Office of Planning, Evaluation and Policy Development, Policy and Program Studies Service, by the American Institutes for Research, March 2018.

Behrstock-Sherratt, Ellen. 2017. “Teacher Shortages: Top 10 Ideas from the First State ESSA Plans” (blog post). Policy Center at American Institutes for Research, May 22, 2017.

Berg-Jacobson, Alex, and Jesse Levin. 2015. Oklahoma Study of Educator Supply and Demand Trends and Projections. American Institutes for Research, September 2015.

Berry, Barnett, and Patrick M. Shields. 2017. “Solving the Teacher Shortage: Revisiting the Lessons We’ve Learned.” Phi Delta Kappan 98, no. 8: 8–18. https://doi.org/10.1177/0031721717708289.

Blitz, Lisa V., Denise Yull, and Matthew Clauhs. 2020. “Bringing Sanctuary to School: Assessing School Climate as a Foundation for Culturally Responsive Trauma-Informed Approaches for Urban Schools.” Urban Education 55, no. 1: 95–124. https://doi.org/10.1177/0042085916651323.

Booker, Cory A., U.S. Senate Office of. 2019. “Booker, Norcross, Pascrell Introduce Legislation to Address Growing Teacher Shortage” (press release). June 19, 2019.

Borman, G.D., and N.M. Dowling. 2008. “Teacher Attrition and Retention: A Meta-Analytic and Narrative Review of the Research.” Review of Educational Research 78, no. 3: 367–409.

Boser, Ulrich, and Chelsea Straus. 2014. Mid- and Late-Career Teachers Struggle with Paltry Incomes. Center for American Progress, July 2014.

Carnoy, Martin, and Emma García. 2017. Five Key Trends in U.S. Student Performance: Progress by Blacks and Hispanics, the Takeoff of Asians, the Stall of Non-English Speakers, the Persistence of Socioeconomic Gaps, and the Damaging Effect of Highly Segregated Schools. Economic Policy Institute, January 2017.

Carroll, T.G. 2007. Policy Brief: The High Cost of Teacher Turnover. National Commission on Teaching and America’s Future.

Carver-Thomas, Desiree. 2018. Diversifying the Teaching Profession: How to Recruit and Retain Teachers of Color. Palo Alto, Calif.: Learning Policy Institute, April 2018.

Carver-Thomas, Desiree, and Linda Darling-Hammond. 2017. Teacher Turnover: Why It Matters and What We Can Do About It. Learning Policy Institute, August 2017.

Chetty, Raj, John N. Friedman, Nathaniel Hilger, Emmanuel Saez, Diane Whitmore Schanzenbach, and Danny Yagan. 2011. “How Does Your Kindergarten Classroom Affect Your Earnings? Evidence from Project STAR.” Quarterly Journal of Economics 126, no. 4: 1593–1660. https://doi.org/10.1093/qje/qjr041.

Chingos, Matthew M., and Kristin Blagg. 2017. Making Sense of State School Funding Policy. Urban Institute, November 2017.

Cisneros, Henry G., Jack F. Kemp, Nicolas P. Retsinas, and Kent W. Colton. 2007. Our Communities, Our Homes: Pathways to Housing and Homeownership in America’s Cities and States. Cambridge, Mass.: Joint Center for Housing Studies of Harvard University.

Croft, Andrew, Jane G. Coggshall, Megan Dolan, and Elizabeth Powers (with Joellen Killion). 2010. Job-Embedded Professional Development: What It Is, Who Is Responsible, and How to Get It Done Well. National Comprehensive Center for Teacher Quality, Mid-Atlantic Comprehensive Center, and National Staff Development Council.

Croft, Michelle, Gretchen Guffy, and Dan Vitale. 2018. Encouraging More High School Students to Consider Teaching. ACT Policy Research, June 2018.

Da Costa, Pedro. 2019. “‘Schools Are No Longer Just Institutions of Learning—We Are the Primary Hub of Care Outside the Family.’” Working Economics Blog (Economic Policy Institute), May 2019.

Darling-Hammond, Linda. 1999. Teacher Quality and Student Achievement: A Review of State Policy Evidence. Center for the Study of Teaching and Policy, University of Washington, December 1999.

Darling-Hammond, Linda. 2010. The Flat World and Education: How America’s Commitment to Equity Will Determine Our Future. New York, N.Y.: Teachers College Press.

Darling-Hammond, Linda. 2019. “Burdensome Student Loan Debt Is Contributing to the Country’s Teacher Shortage Crisis.” Forbes, November 17, 2019.

Darling-Hammond, Linda, Dion Burns, Carol Campbell, A. Lin Goodwin, Karen Hammerness, Ee-Ling Low, Ann McIntyre, Mistilina Sato, and Kenneth Zeichner. 2017. Empowered Educators: How High-Performing Systems Shape Teaching Quality Around the World. San Francisco: Jossey-Bass.

Darling-Hammond, Linda, Maria E. Hyler, and Madelyn Gardner. 2017. Effective Teacher Professional Development. Learning Policy Institute, June 2017.

Darling-Hammond, Linda, and Milbrey W. McLaughlin. 1995. “Policies That Support Professional Development in an Era of Reform.” Phi Delta Kappan 76, no. 8: 597–604.

Darling-Hammond, Linda, Ruth Chung Wei, Alethea Andree, Nikole Richardson, and Stelios Orphanos. 2009. Professional Learning in the Learning Profession: A Status Report on Teacher Development in the United States and Abroad. National Staff Development Council and School Redesign Network at Stanford University, February 2009.

Dee, Thomas S., and Dan Goldhaber. 2017. “Understanding and Addressing Teacher Shortages in the United States.” Policy Proposal 2017-05, The Hamilton Project at the Brookings Institution.

Desimone, Laura M. 2009. “Improving Impact Studies of Teachers’ Professional Development: Toward Better Conceptualizations and Measures.” Educational Researcher 38, no. 3: 181–199. https://doi.org/10.3102/0013189X08331140.

Desimone, Laura M., and Michael S. Garet. 2015. “Best Practices in Teachers’ Professional Development in the United States.” Psychology, Society and Education 7, no. 3: 252–263.

Doherty, Kathryn M., Sandi Jacobs, and Trisha M. Madden. 2012. No One Benefits: How Teacher Pension Systems Are Failing Both Teachers and Taxpayers. National Council on Teacher Quality.

DuFour, Richard. 2004. “What Is a Professional Learning Community?” Educational Leadership 61, no. 8: 6–11.

Eisner, Ryan, Eleanor Fulbeck, Glenance Green, Kelly Hallberg, David Manzeske, Bo Zhu, and Martyna Citkowicz. 2015. Lessons From AIR’s Ongoing Evaluation of the Denver Teacher Residency. AIR. https://www.air.org/project/denver-teacher-residency-program-evaluation.

Erasmus, Estelle. 2019. “New Economic Report Shows Most Teachers Have Side Hustles to Make Ends Meet.” Forbes, May 27, 2019.

Espinoza, Danny, Ryan Saunders, Tara Kini, and Linda Darling-Hammond. 2018. Taking the Long View: State Efforts to Solve Teacher Shortages by Strengthening the Profession. Learning Policy Institute, August 2018.

Feng, Li, and Tim R. Sass. 2015. “The Impact of Incentives to Recruit and Retain Teachers in ‘Hard-to-Staff’ Subjects.” Working Paper no. 141, National Center for Analysis of Longitudinal Data in Education Research (CALDER), American Institutes for Research, September 2015.

Ferlazzo, Larry. 2018. “Response: Do Professional Development ‘With’ Teachers, Not ‘to’ Them.” Education Week, June 7, 2018.

Fiddiman, Bayliss, Colleen Campbell, and Lisette Partelow. 2019. Student Debt: An Overlooked Barrier to Increasing Teacher Diversity. Center for American Progress, July 2019.

Figlio, David. 2017. The Importance of a Diverse Teaching Force. The Brookings Institution, November 2017.

Folbre, Nancy, and Kristin Smith. 2017. “The Wages of Care: Bargaining Power, Earnings and Inequality.” Political Economy Research Institute and the Washington Center for Equitable Growth, Working Paper, February 2017.

García, Emma. 2014. The Need to Address Noncognitive Skills in the Education Policy Agenda. Economic Policy Institute, December 2014.

García, Emma. 2019. “It’s the Beginning of the School Year and Teachers Are Once Again Opening Up Their Wallets to Buy School Supplies.” Working Economics Blog (Economic Policy Institute), August 2019.

García, Emma, and Eunice S. Han. 2019. Teachers’ Base Salary and Districts’ Academic Performance: Evidence from National Data. Unpublished manuscript.

García, Emma, and Elaine Weiss. 2016. Making Whole-Child Education the Norm: How Research and Policy Initiatives Can Make Social and Emotional Skills a Focal Point of Children’s Education. Economic Policy Institute, August 2016.

García, Emma, and Elaine Weiss. 2019a. The Teacher Shortage Is Real, Large and Growing, and Worse Than We Thought: The First Report in ‘The Perfect Storm in the Teacher Labor Market’ Series. Economic Policy Institute, March 2019.

García, Emma, and Elaine Weiss. 2019b. U.S. Schools Struggle to Hire and Retain Teachers: The Second Report in ‘The Perfect Storm in the Teacher Labor Market’ Series. Economic Policy Institute, April 2019.

García, Emma, and Elaine Weiss. 2019c. Low Relative Pay and High Incidence of Moonlighting Play a Role in the Teacher Shortage, Particularly in High-Poverty Schools: The Third Report in ‘The Perfect Storm in the Teacher Labor Market’ Series. Economic Policy Institute, May 2019.

García, Emma, and Elaine Weiss. 2019d. Challenging Working Environments (‘School Climates’), Especially in High-Poverty Schools, Play a Role in the Teacher Shortage: The Fourth Report in ‘The Perfect Storm in the Teacher Labor Market’ Series. Economic Policy Institute, May 2019.

García, Emma, and Elaine Weiss. 2019e. The Role of Early Career Supports, Continuous Professional Development, and Learning Communities in the Teacher Shortage: The Fifth Report in ‘The Perfect Storm in the Teacher Labor Market’ Series. Economic Policy Institute, July 2019.

García, Emma, and Elaine Weiss. 2020a. “How Teachers View Their Own Professional Status: A Snapshot.” Phi Delta Kappan 101, no. 6: 14–18. https://doi.org/10.1177/0031721720909581.

García, Emma, and Elaine Weiss. 2020b. Examining the Factors That Play a Role in the Teacher Shortage Crisis: Key Findings from EPI’s ‘Perfect Storm in the Teacher Labor Market’ Series. Economic Policy Institute, forthcoming.

Garet, Michael S., Andrew C. Porter, Laura Desimone, Beatrice F. Birman, and Kwang Suk Yoon. 2001. “What Makes Professional Development Effective? Results from a National Sample of Teachers.” American Educational Research Journal 38, no. 4: 915–945. https://doi.org/10.3102/00028312038004915.

Gay, Geneva. 2002. “Preparing for Culturally Responsive Teaching.” Journal of Teacher Education 53, no. 2: 106–116.

Gius, Mark. 2013. “The Effects of Merit Pay on Teacher Job Satisfaction.” Applied Economics 45, no. 31: 4443–4451. https://doi.org/10.1080/00036846.2013.788783.

Glazerman, Steven, and Allison Seifullah. 2012. An Evaluation of the Chicago Teacher Advancement Program (Chicago TAP) After Four Years. Final Report. Mathematica Policy Research, March 2012.

Goldhaber, Dan, John Krieg, and Roddy Theobald. 2018a. “Exploring the Impact of Student Teaching Apprenticeships on Student Achievement and Mentor Teachers.” CALDER Working Paper no. 207-1118-1, November 2018.

Goldhaber, Dan, John Krieg, and Roddy Theobald. 2018b. “Effective Like Me? Does Having a More Productive Mentor Improve the Productivity of Mentees?” CALDER Working Paper no. 208-1118-1, November 2018.

Goldring, Rebecca, Soheyla Taie, and Minsun Riddles. 2014. Teacher Attrition and Mobility: Results from the 2012–13 Teacher Follow-Up Survey (NCES 2014-077). U.S. Department of Education, National Center for Education Statistics.

Gould, Elise. 2017. Local Public Education Employment May Have Weathered Recent Storms, but Schools Are Still Short 327,000 Public Educators (economic snapshot). October 6, 2017.

Gould, Elise. 2019. Back-to-School Jobs Report Shows a Continued Shortfall in Public Education Jobs (economic snapshot). October 10, 2019.

Guarino, Cassandra M., Lucrecia Santibáñez, and Glenn A. Daley. 2006. “Teacher Recruitment and Retention: A Review of the Recent Empirical Literature,” Review of Educational Research 76, no. 2: 173–208.

Guha, R., M.E. Hyler, and L. Darling-Hammond. 2016. The Teacher Residency: An Innovative Model for Preparing Teachers. Learning Policy Institute, September 2016.

Han, Eunice S. 2019. “The Impact of Teachers Unions on Teachers’ Well-Being Under Various Legal Institutions: Evidence from District–Teacher Matched Data.” AERA Open 5, no. 3. https://doi.org/10.1177/2332858419867291.

Hanushek, Eric A., Marc Piopiunik, and Simon Wiederhold. 2019. “Do Smarter Teachers Make Smarter Students?” Education Next 19, no. 2: 57–64.

Hill, Heather C. 2007. “Learning in the Teacher Workforce.” Future of Children 17, no. 1: 111–127. https://doi.org/10.1353/foc.2007.0004.

Hirsh, Stephanie. 2009. “A New Definition.” Journal of Staff Development 30, no. 4: 10–16.

Hussar, W.J., and T.M. Bailey. 2020. Projections of Education Statistics to 2028 (NCES 2020-024). U.S. Department of Education, Washington, D.C.: National Center for Education Statistics. https://nces.ed.gov/pubs2020/2020024.pdf

Imberman, S.A. 2015. “How Effective Are Financial Incentives for Teachers?” IZA World of Labor (Institute of Labor Economics newsletter), article no. 158. https://doi.org/10.15185/izawol.158.

Ingersoll, Richard M. 2001. “Teacher Turnover and Teacher Shortages: An Organizational Analysis.” American Educational Research Journal 38, no. 3: 499–534; https://doi.org/10.3102/00028312038003499.

Ingersoll, Richard M. 2004. “Revolving Doors and Leaky Buckets.” In Letters to the Next President: What We Can Do About the Real Crisis in Public Education, edited by Carl D. Glickman, 141–150. New York, N.Y.: Teachers College Press.

Ingersoll, Richard M. 2014. “Why Do High-Poverty Schools Have Difficulty Staffing Their Classrooms with Qualified Teachers?” Presentation for the panel discussion How Do We Get Experienced, Accomplished Teachers into High-Need Schools? Albert Shanker Institute, October 8, 2014.

Ingersoll, Richard M. 2015. The State of Teacher Diversity in American Education. Albert Shanker Institute.

Ingersoll, Richard M., and Gregory J. Collins. 2018. “The Status of Teaching as a Profession.” In Schools and Society: A Sociological Approach to Education, 6th ed., edited by Jeanne H. Ballantine, Joan Z. Spade, and Jenny M. Stuber, 199–213. Los Angeles: SAGE.

Ingersoll, Richard M., and Henry May. 2011. “The Minority Teacher Shortage: Fact or Fable?” Kappan Magazine 93, no. 1: 62–65.

Ingersoll, Richard M., Lisa Merrill, and Henry May. 2014. What Are the Effects of Teacher Education and Preparation on Beginning Teacher Attrition?  Consortium for Policy Research in Education, University of Pennsylvania.

Ingersoll, Richard M., Lisa Merrill, and Daniel Stuckey. 2014. Seven Trends: The Transformation of the Teaching Force. Consortium for Policy Research in Education, University of Pennsylvania, CPRE Report (#RR-80).

Ingersoll, Richard M., Elizabeth Merrill, Daniel Stuckey, and Gregory Collins. 2018. Seven Trends: The Transformation of the Teaching Force—Updated October 2018. CPRE Research Reports.

Ingersoll, Richard M., and Michael Strong. 2011. “The Impact of Induction and Mentoring Programs for Beginning Teachers: A Critical Review of the Research.” Review of Educational Research 81, no. 2: 201–233. https://doi.org/10.3102/0034654311403323.

Jackson, C.K., and E. Bruegmann. 2009. “Teaching Students and Teaching Each Other: The Importance of Peer Learning for Teachers.” American Economic Journal: Applied Economics 1, no. 4: 85–108. https://doi.org/10.1257/app.1.4.85.

Jones, N.D., E. Bettini, and M.T. Brownell. 2016. Can Collaborative School Reform and Teacher Evaluation Reform Be Reconciled? Albert Shanker Institute.

Katz, V. 2018. Teacher Retention: Evidence to Inform Policy. Policy Brief. Curry School and Batten School, EdPolicyWorks, University of Virginia, Charlottesville. Prepared for Virginia’s Teacher Retention Summit, Charlottesville, October 23, 2018.

Keefe, Jeffrey. 2018. Pennsylvania’s Teachers Are Undercompensated—and New Pension Legislation Will Cut Their Compensation Even More: Undercompensation Is Likely a Factor in Pennsylvania’s Growing Teacher Shortage. Economic Policy Institute, February 2018.

Kirp, David. 2019. “The Community School Comes of Age: The Model Is Expanding Rapidly. Is It a Fad, or the Future?” New York Times, January 10, 2019.

Kraft, Matthew A., David Blazar, and Dylan Hogan. 2018. “The Effect of Teacher Coaching on Instruction and Achievement: A Meta-Analysis of the Causal Evidence.” Review of Educational Research 88, no. 4: 547–558. https://doi.org/10.3102/0034654318759268.

Kraft, Matthew A., and John P. Papay. 2014. “Can Professional Environments in Schools Promote Teacher Development? Explaining Heterogeneity in Returns to Teaching Experience.” Educational Evaluation and Policy Analysis 36, no. 4: 476–500.

Ladd, Helen F., and Lucy C. Sorensen. 2016. “Returns to Teacher Experience: Student Achievement and Motivation in Middle School.” Education Finance and Policy 12, no. 2: 241–279.

Lambert, Diana, and Daniel J. Willis. 2019. “In Need of Teacher Housing, More California School Districts Building Their Own.” EdSource, April 17, 2019.

Leachman, Michael, and Eric Figueroa. 2019. K–12 School Funding Up in Most 2018 Teacher-Protest States, But Still Well Below Decade Ago. Center on Budget and Policy Priorities, March 2019.

Learning Policy Institute (LPI). 2017. What’s the Cost of Teacher Turnover? (calculator). September 2017.

Levin, Jesse, Alex Berg-Jacobson, Drew Atchison, Katelyn Lee, and Emily Vontsolos. 2015. Massachusetts Study of Teacher Supply and Demand Trends and Projections. American Institutes for Research.

Loeb, Susanna, Linda Darling-Hammond, and John Luczak. 2005. “How Teaching Conditions Predict Teacher Turnover in California Schools.” Peabody Journal of Education 80, no. 3: 44–70.

Loeb, Susanna, and Marianne E. Page. 2000. “Examining the Link Between Teacher Wages and Student Outcomes: The Importance of Alternative Labor Market Opportunities and Non-Pecuniary Variation.” Review of Economics and Statistics 82, no. 3: 393–408. https://doi.org/10.1162/003465300558894.

Long, Cindy. 2019. “Money Still Matters: A Tale of Two Teachers.” National Education Association. Accessed April 30, 2019.

Lyon, M. 2020. “Teacher’s Political Voices: The Effects of Policies Hindering Unionization on Teachers, Students, and State Politics.” Columbia University Working Paper.

McInerney, Maura, and Amy McKlindon. 2014. Unlocking the Door to Learning: Trauma-Informed Classrooms & Transformational Schools. The Education Law Center, December 2014.

Mishel, Lawrence, and Richard Rothstein, eds. 2002. The Class Size Debate. Washington, D.C.: Economic Policy Institute.

Moore-Johnson, S., M.L. Donaldson, M.S. Munger, J.P. Papay, and E.K. Qazilbash. 2007. Leading the Local: Teacher Union Presidents Speak on Change, Challenges. Education Sector.

Moore-Johnson, Susan, Matthew A. Kraft, and John P. Papay. 2012. “How Context Matters in High-Need Schools: The Effects of Teachers’ Working Conditions on Their Professional Satisfaction and Their Students’ Achievement.” Teachers College Record 114, no. 10: 1–39.

Morgan, Ivy, and Ary Amerikaner. 2018. Funding Gaps 2018: An Analysis of School Funding Equity Across the U.S. and Within Each State. Education Trust, February 2018.

Morrissey, Monique. 2017. Teachers and Schools Are Well Served by Teacher Pensions. Economic Policy Institute, October 2017.

Murnane, R.J., and R.J. Olsen. 1989. “The Effect of Salaries and Opportunity Costs on Duration in Teaching: Evidence from Michigan.” Review of Economics and Statistics 71, no. 2: 347–352. https://doi.org/10.2307/1926983.

National Academies of Sciences, Engineering, and Medicine (NASEM). 2017. Promoting the Educational Success of Children and Youth Learning English: Promising Futures. Washington, D.C.: National Academies Press. https://doi.org/10.17226/24677.

National Center for Education Statistics (NCES) (U.S. Department of Education). Various years.  “Percentage Distribution of Public School Teachers by Stayer, Mover, and Leaver Status for Selected Teacher and School Characteristics in the Base Year: 1994–95, 2000–01, 2004–05, and 2008–09.” Schools and Staffing Survey (SASS), “Public School Teacher Data Files,” 1993–94, 1999–2000, 2003–04, and 2007–08; Teacher Follow-up Survey (TFS), “Current and Former Teacher Data Files,” 1994–95, 2000–01, 2004–05, and 2008–09, 2012-2013, U.S. Department of Education.

National Center for Education Statistics (NCES) (U.S. Department of Education). 2011–2012. Licensed microdata from the 2011–2012 Schools and Staffing Survey (SASS).

National Center for Education Statistics (NCES) (U.S. Department of Education). 2012–2013. Licensed microdata from the 2012–2013 Teacher Follow-Up Survey (TFS).

National Center for Education Statistics (NCES) (U.S. Department of Education). 2015–2016. Licensed microdata from the 2015–2016 National Teacher and Principal Survey (NTPS).

National Center for Education Statistics (NCES) (U.S. Department of Education). 2017. “Table 236.20. Total Expenditures for Public Elementary and Secondary Education and Other Related Programs, by Function and Subfunction: Selected Years, 1990-91 Through 2013-14.” Digest of Education Statistics: 2016, Institute of Education Sciences, U.S. Department of Education, NCES, July 2016

National Center for Education Statistics (NCES) (U.S. Department of Education). 2018. Digest of Education Statistics: 2017.

National Center for Education Statistics (NCES) (U.S. Department of Education). 2019a. “Table 203.50. Enrollment and Percentage Distribution of Enrollment in Public Elementary and Secondary Schools, by Race/Ethnicity and Region: Selected years, Fall 1995 Through Fall 2028.” Digest of Education Statistics: 2018, Institute of Education Sciences, U.S. Department of Education, NCES no. 2020-009, December 2019.

National Center for Education Statistics (NCES) (U.S. Department of Education). 2019b. “Table 204.10. Number and Percentage of Public School Students Eligible for Free or Reduced-Price Lunch, by State: Selected Years, 2000-01 through 2016-17.” Digest of Education Statistics: 2018, Institute of Education Sciences, U.S. Department of Education, NCES no. 2020-009, December 2019.

National Center for Education Statistics (NCES), U.S. Department of Education. 2019c. “Table 208.20. Public and Private Elementary and Secondary Teachers, Enrollment, Pupil/Teacher Ratios, and New Teacher Hires: Selected Years, Fall 1955 through Fall 2028.” Digest of Education Statistics: 2018, Institute of Education Sciences, U.S. Department of Education, NCES no. 2018-070, April 2019.

National Education Association Foundation (NEA Foundation). 2012. Peer Assistance and Review: All Teachers on the Road to Instructional Leadership in Columbus (OH) “100% Project Schools.” May 2012.

Nguyen, Tuan D., Lam Pham, Matthew Springer, and Michael Crouch. 2019. “The Factors of Teacher Attrition and Retention: An Updated and Expanded Meta-Analysis of the Literature.” EdWorkingPaper no. 19-149, October 2019. Retrieved from Annenberg Institute at Brown University. https://doi.org/10.26300/cdf3-4555.

Nieto, Ana B. 2019. “Los Verdaderos ‘Influencers’ a Veces no Llegan a Fin de Mes.” La Raza, September 4, 2019.

Organisation for Economic Co-operation and Development (OECD). 2016. Supporting Teacher Professionalism: Insights from TALIS 2013. Paris: TALIS, OECD Publishing. https://dx.doi.org/10.1787/9789264248601-en.

Organisation for Economic Co-operation and Development (OECD). 2019. TALIS 2018 Results (Volume I): Teachers and School Leaders as Lifelong Learners. Paris: TALIS, OECD Publishing. https://doi.org/10.1787/1d0bc92a-en.

Papay, John P., Andrew Bacher-Hicks, Lindsay C. Page, and William H. Marinell. 2017. “The Challenge of Teacher Retention in Urban Schools: Evidence of Variation From a Cross-Site Analysis.” Educational Researcher 46, no. 8: 434–448. https://doi.org/10.3102/0013189X17735812.

Papay, John P., Eric S. Taylor, John H. Tyler, and Mary Laski. 2016. “Learning Job Skills from Colleagues at Work: Evidence from a Field Experiment Using Teacher Performance Data.” National Bureau of Economic Research Working Paper no. 21986, February 2016. https://doi.org/10.3386/w21986.

Papay, John P., Martin R. West, Jon B. Fullerton, and Thomas J. Kane. 2012. “Does an Urban Teacher Residency Increase Student Achievement? Early Evidence from Boston.” Educational Evaluation and Policy Analysis 34, no. 4: 413–434. https://doi.org/10.3102/0162373712454328.

Pew Charitable Trusts (Pew). 2019. The State Pension Funding Gap: 2017. Overall Debt Decreases but Disparity Grows Between Well-Funded and Fiscally Challenged Public Worker Retirement Plans. June 2019.

Phi Delta Kappan (PDK). 2018. Teaching: Respect but Dwindling Appeal. The 50th Annual PDK Poll of the Public’s Attitudes Toward the Public Schools. Supplement to Kappan magazine, September 2018.

Phi Delta Kappan (PDK). 2019. Frustration in the Schools: Teachers Speak Out on Pay, Funding, and Feeling Valued. The 51st Annual PDK Poll of the Public’s Attitudes Toward the Public Schools. Supplement to Kappan magazine, September 2019.

Picchi, Aimee. 2019. “New Teachers Can’t Afford the Rent in All but One Major U.S. City.” CBS News, August 27, 2019.

Podolsky, Anne, and Tara Kini. 2016. How Effective Are Loan Forgiveness and Service Scholarships for Recruiting Teachers? Learning Policy Institute, April 2016.

Podolsky, Anne, Tara Kini, Joseph Bishop, and Linda Darling-Hammond. 2016. Solving the Teacher Shortage: How to Attract and Retain Excellent Educators. Learning Policy Institute, September 2016.

Podolsky, Anne, Tara Kini, Linda Darling-Hammond, and Joseph Bishop. 2019. “Strategies for Attracting and Retaining Educators: What Does the Evidence Say?” Education Policy Analysis Archives 27, no. 38. http://dx.doi.org/10.14507/epaa.27.3722.

Quint, Janet. 2011. “Professional Development for Teachers. What Two Rigorous Studies Tell Us.” MDRC Paper, July 2011.

Reilly, Katie. 2018. “‘I Work 3 Jobs and Donate Blood Plasma to Pay the Bills.’ This Is What It’s Like to Be a Teacher in America.” Time, September 13, 2018.

Restorative Practices Working Group (RPWG). 2014. Restorative Practices: Fostering Healthy Relationships and Promoting Positive Discipline in Schools: A Guide for Educators. Advancement Project, American Federation of Teachers, National Education Association, and National Opportunity to Learn Campaign, March 2014.

Rhee, Nari, and Leon F. Joyner Jr. 2019. Teacher Pensions vs. 401(k)s in Six States: Connecticut, Colorado, Georgia, Kentucky, Missouri, and Texas. UC Berkeley Labor Center and the National Institute on Retirement Security, January 2019.

Richards, Erin, and Matt Wynn. 2019. “‘Can’t Pay Their Bills with Love’: In Many Teaching Jobs, Teachers’ Salaries Can’t Cover Rent.” USA Today, June 5, 2019.

Ronfeldt, Matthew, Susanna Loeb, and James Wyckoff. 2013. “How Teacher Turnover Harms Student Achievement.” American Educational Research Journal 50, no. 1: 4–36.

Schultz, Katherine. 2019. “The Distrust Beneath the Recent Teacher Strikes.” The Blog of Harvard Education Publishing, April 23, 2019.

Schwartz, Sarah 2019. “What Do Teachers Really Want from Professional Development? Respect.” Education Week, May 14, 2019.

Silva, Tim, Allison McKie, and Philip Gleason. 2015. New Findings on the Retention of Novice Teachers From Teaching Residency Programs. Prepared for the National Center for Education Evaluation and Regional Assistance, Institute for Education Sciences, by Mathematica Policy Research, August 2015.

Silva, T., A. McKie, V. Knechtel, P. Gleason, and L. Makowsky. 2014. Teaching Residency Programs: A Multistate Look at a New Model to Prepare Teachers for High-Need Schools. Prepared for the National Center for Education Evaluation and Regional Assistance, Institute for Education Sciences, by Mathematica Policy Research, November 2014.

Simon, Nicole S, and Susan Moore-Johnson. 2015. Teacher Turnover in High-Poverty Schools: What We Know and Can Do. Teachers College Record 117: 1–36.

Sorensen, Lucy C., and Helen Ladd. 2020. “The Hidden Costs of Teacher Turnover.” AERA Open. https://doi.org/10.1177/2332858420905812.

Startz, Dick. 2016. “Are Teacher Pensions Really That High? Pension Puzzles Part I.” Brown Center Chalkboard (Brookings Institution blog), Thursday, September 29, 2016.

Strauss, Valerie. 2017. “Why It’s a Big Problem That So Many Teachers Quit—and What to Do About It.” Washington Post, November 27, 2017.

Sutcher, Leib, Linda Darling-Hammond, and Desiree Carver-Thomas. 2016. A Coming Crisis in Teaching? Teacher Supply, Demand, and Shortages in the U.S. Learning Policy Institute, September 2016.

Talley, Tim, Melissa Daniels, Michael Melia, and John Raby. 2018. “‘I Just Have to Do It.’ Teachers Struggle With Second Jobs.” AP News, April 15, 2018.

Thomas, M. Shelley, Shantel Crosby, and Judi Vanderhaar. 2019. “Trauma-Informed Practices in Schools Across Two Decades: An Interdisciplinary Review of Research.” Review of Research in Education 43, no. 1: 422–452. https://doi.org/10.3102/0091732X18821123.

U.S. Department of Education. 2017a. “Completers, By State, By Program Type” [data table]. Data from the Higher Education Act Title II State Report Card (SRC) Reporting System. Spreadsheet downloadable at https://title2.ed.gov/Public/DataTools/NewExcels/CompletersProgramType.aspx.

U.S. Department of Education. 2017b. “Enrollment, By State, By Program Type” [data table]. Data from the Higher Education Act Title II State Report Card (SRC) Reporting System. Spreadsheet downloadable at https://title2.ed.gov/Public/DataTools/NewExcels/EnrollmentProgramType.aspx.

U.S. Department of Education. 2017c. Teacher Shortage Areas Nationwide Listing. 1990–1991 through 2017–2018. June 2017.

U.S. Department of Education. 2019. “Office of Postsecondary Education: Teacher Shortage Areas” (web page). Accessed March 2019.

Urban Institute. 2017. “School Funding: Do Poor Kids Get Their Fair Share?” (interactive data web page). May 2017.

Viadero, Debra. 2002. “Researcher Skewers Explanations Behind Teacher Shortage.” Education Week, April 10, 2002

Warner-Griffin, Catharine, Brittany C. Cunningham, and Amber Noel. 2018. Public School Teacher Autonomy, Satisfaction, Job Security, and Commitment: 1999–2000 and 2011–12. Institute of Education Statistics, National Center for Education Statistics, U.S. Department of Education, Stats in Brief no. 2018-103, March 2018.

Weingarten, Randi. 2019. “The Freedom to Teach.” Keynote address at “The Teaching Profession in Crisis: The Reasons and Remedies” conference at the National Press Club, Washington, D.C., April 18, 2019.

Weiss, Elaine, and Paul Reville. 2019. Broader, Bolder, Better. How Schools and Communities Help Students Overcome the Disadvantages of Poverty. Cambridge, Mass.: Harvard Education Publishing Group.

White House Initiative on Educational Excellence for Hispanics (WHIEEH). 2015. Hispanic Teacher Recruitment: Increasing the Number of Hispanic Teachers. April 2015.

Zdanowicz, Christina. 2019. “Denver Is So Expensive That Teachers Have to Get Creative to Make Ends Meet.” CNN, February 11, 2019.

Introduction

The COVID-19 pandemic is overwhelming the functioning and outcomes of education systems—some of which were already stressed in many respects. This is true across the world and affects all children, though to differing degrees depending on multiple factors—including the country/region where they live, as well as their ages, family backgrounds, and degree of access to some “substitute” educational opportunities during the pandemic. In early spring as the pandemic was hitting its first peak, the virus consigned nearly all of over 55 million U.S. school children under the age of 18 to staying in their homes, with 1.4 billion out of school or child care across the globe (NCES 2019a; U.S. Census Bureau 2019; Cluver et al. 2020). Not only did these children lack daily access to school and the basic supports schools provide for many students, but they also lost out on group activities, team sports, and recreational options such as pools and playgrounds.

(COVID-19 & Education Webinar: Join us Wednesday for a discussion on this report, including opening remarks from Randi Weingarten, the president of the 1.7 million-member American Federation of Teachers, AFL-CIO, about the state of COVID-19 and education and what needs to be done now to support educators and mitigate the damage to student performance, especially the most vulnerable children. Register here.)

The shutdown of schools, compounded by the associated public health and economic crises, poses major challenges to our students and their teachers. Our public education system was not built, nor prepared, to cope with a situation like this—we lack the structures to sustain effective teaching and learning during the shutdown and to provide the safety net supports that many children receive in school. While we do not know the exact impacts, we do know that children’s academic performance is deteriorating during the pandemic, along with their progress on other developmental skills. We also know that, given the various ways in which the crisis has widened existing socioeconomic disparities and how these disparities affect learning and educational outcomes, educational inequities are growing (Rothstein 2004; Putnam 2015; Reardon 2011; García and Weiss 2017). As a consequence, many of the children who struggle the hardest to learn effectively and thrive in school under normal circumstances are now finding it difficult, even impossible in some cases, to receive effective instruction, and they are experiencing interruptions in their learning that will need to be made up for.

The 2020–2021 school year is now underway, and with many schools remaining physically closed as the 2020–2021 year begins, there is more we need to understand and think through if we are to meet the crisis head-on. If students are to not see their temporary interruptions become sustained and are to regain lost ground, if teachers are to do their jobs effectively during and after the pandemic, and if our education system is to deliver on its excellence and equity goals during the next phases of this pandemic, it will be critical to identify which students are struggling most and how much learning and development they have lost out on, which factors are impeding their learning, what problems are preventing teachers from teaching these children, and, very critically, which investments must be made to address these challenges. For each child, this diagnostic assessment will deliver a unique answer, and the system will have to meet the child where he or she is. A strengthened system based on meeting children where they are and providing them with what they need will be key to lifting up children.

This report briefly reviews the relevant literature on educational settings that have features in common with how education is occurring during the crisis and emerging evidence on opportunity gaps during the COVID-19 pandemic in order to propose a three-pronged plan. The plan covers the three Rs: (immediate) relief for schools, (short-term) recovery, and (long-term) rebuilding for schools and the education system as a whole.

Children are not in their schools: What should we expect the consequences to be?

The current downturn is unique, and in most ways it is much more severe than any we have experienced in recent history. Almost overnight, the pandemic forced the cancellation of the traditional learning that takes place in school settings. It imposed substantial alterations in the “inputs” used to produce education—typically all the individual, family, teacher, school, etc., characteristics or determinants that affect “outcomes” like test scores and graduation rates. The pandemic has affected inputs at home too, as families and communities juggling health and work crises are less able to provide supports for learning at home.1 Because there are no direct comparisons to past events or trends, we are without fully valid references for assessing the likely impacts of the COVID-19 crisis on children. There are, however, specific aspects of this crisis that have arisen in other contexts and been studied by education researchers, and we can derive from them some guidance on topics such as the loss of learning time and use of alternative learning modes.

Here we thus summarize research findings on aspects of education that appear most pertinent to the current crisis. We selected this set of studied conditions because they represent situations in which children are out of school in large numbers or using the unusual learning tools that have become typical in recent months. As discussed in the sections below, however, the sudden, severe, and universal nature of this crisis means that the current contexts in which students are currently “absent,” engaged in “remote learning,” or “homeschooled” are very different during the pandemic. However, while these findings are only partially applicable to the situations arising during this pandemic, if we dig into why various modes of learning worked or did not work well, it can help guide how to improve learning as education continues under the pandemic—and how to lift children up once schools recover their normal mode of operation.2

Decreased learning time has likely impeded student learning

The school lockdowns that started in the spring of 2020 reduced instructional and learning time, which are known to impede student performance, with disparate impacts on different groups of students.

Research on time in school anticipates the consequences of having learning interrupted

International and U.S. data provide a benchmark of what can be considered usual educational progress over a given school year. Here we look at data on reading, math, and science test results of 15-year-old students in countries all over the world from the Programme for International Student Assessment (PISA) run by the Organisation for Economic Co-operation and Development (OECD 2009) and data on a cohort of U.S. children who entered kindergarten in 2010 for the 2010–2011 school year from Early Childhood Longitudinal Study, Kindergarten Class of 2010–2011 (ECLS-K-2010–2011), run by the U.S. Department of Education, National Center for Education Statistics (NCES 2010–2011). From these studies, it has been estimated how much children learn over a school year (to make the estimates of how far the group’s average score on skills were at the end of the year from their skill levels at the beginning of a year comparable across studies, we use standard deviations). On average, students advance in their academic performance by between about 0.3 standard deviations (SD) and 0.5 SD to 0.7 SD per year, depending on their age and subject/skill (OECD 2009; own analysis based on NCES 2010–2011).3 The 2019–2020 school year was cut by at least one third relative to its normal length, which, assuming linear increments in growth over the year and no major other obstacles, suggests a loss of at least 0.1 SD across the board, and larger in earlier grades. These benchmarks will be helpful as we look at the various ways that students have seen their learning interrupted and disrupted this year, and they will continue to do so in 2020–2021.

It is useful as well to examine the research on the length of the school day, which has identified a causal relationship between the amount of (high-quality) instructional time and student performance (Figlio, Holden, and Özek 2018; Goodman 2014; Kidronl and Lindsay 2014; Jin Jez and Wassmer 2013; Marcotte and Hansen 2010). Challenges, though, arise in most evaluations because it is difficult to disentangle the effects of the length of the school day from the effects of starting the school day earlier, or switching to a four-day school week, or to year-round instruction.4

Figlio, Holden, and Özek (2018) find that extending the school day by an hour to provide literacy instruction increases reading scores by 0.05 SD in elementary schools. Thompson (2019) explains that school days lost due to weather-related cancellations negatively impact performance (citing Marcotte 2007; Marcotte and Hemelt 2008), and that the positive impact of a four-day school week on performance is due to the longer school day, the increased flexibility, and the expanded total learning time over the year. He finds a negative effect (0.03–0.05 SD) of four-day school weeks on performance in Oregon, where weekly instructional time was lower in the districts adopting this model.

Research on summer learning losses and gains show that these vary widely

Another body of research that speaks to potential lost learning time arises from studies of so-called summer learning loss. In earlier research, researchers consistently found that test scores for low-income students would decrease over the summer, while test scores for better-off students would stay constant or increase slightly (Kuhfeld 2019 based on Cooper et al. 1996).5 (This pattern has also been referred to in some studies as “slide” or “setback”). A limitation of this earlier research, however, was that the samples represented students who were in school in the 1970s and 1980s—and thus were exposed to very different circumstances than their current counterparts.6

The findings from more recent evidence on summer learning are less consistent. One study reveals a substantial learning loss over the summer of about one to two months in reading and from one to three months of school-year learning in math (Kufheld 2019). Others find that, on average, the change in scores over the summer is near zero—which von Hippel, Workman, and Downey (2018) have renamed “summer slowdown” or “summer stagnation.” Researchers tend to agree, though, on the fact that there is a large variation in summer learning among students, and on the fact that gaps between students of differing socioeconomic status (SES)—specifically high- and low-SES students—widen (Atteberry and McEachin 2020; Kuhfeld 2019; von Hippel, Workman, and Downey 2018).7

Multiple factors are used to explain the variation in these findings. In addition to differences in the educational resources that families provide children across the year, there are a large number of factors that appear to affect learning and are of particular relevance in the current context when trying to gauge the level of learning that has taken place during the pandemic: these findings on summer learning (loss or gain) reflect the great range of learning styles that students exhibit during the summer, or when schools are not in session, i.e., learning styles and outcome levels vary greatly because students have different innate individual characteristics and their learning and development is shaped by multiple factors and circumstances, in and out of school. This fact will be critically important when schools are back in session in the following two ways. First, when educators measure and assess children’s learning, they will need to consider that there are many ways that children learn and many types of knowledge that they acquire beyond math and reading. In other words, teaching and assessing children needs to be done within a framework that understands that each child may have learned differently and may have learned different things. Second, when designing how to best lift children up to make up for the extended out-of-school sessions and disruptions, it will be critical to create more personalized instruction and extend learning (see the policy section at the end of the report).

Research on chronic absenteeism reinforces the urgency of tending children at risk of becoming disengaged

The literature on student absenteeism also sheds light on the relationship between learning and instructional time. The evidence indicates that the negative relationship between absenteeism and student outcomes becomes more intense the more school days that a student misses. Using data from public schools in Chicago, Allensworth and Evans (2016) noted that each week of absence per semester in ninth grade is associated with a more than 20% decline in the probability of graduating from high school. With respect to performance, the disadvantage associated with absenteeism grows as the number of days missed increases: students who missed 1–2 school days, 3–4 days, 5–10 days, or more than 10 days scored, respectively, 0.10, 0.29, 0.39, and 0.64 SD below students who missed no school on mathematics performance for eighth graders (García and Weiss 2018; see Figure A reproduced below).

As this correlation between days absent and declining test scores indicates, there also seems to be a point after which the disadvantage becomes much larger. Indeed, researchers put a strong emphasis on “chronic absenteeism” as the critical indicator, as students who are chronically absent are at serious risk of falling behind in school, having lower grades and test scores, exhibiting behavioral issues, and, ultimately, dropping out (Balfanz 2017; U.S. Department of Education 2016; Gottfried and Ehrlich 2018).8 Indeed, the risk of dropping out is of particular concern for students for whom the pandemic may act as the revolving door but one that ushers them away from the school period (IES 2020; Dorn et al. 2020; Stancati, Brody, and Fontdeglòria 2020; Torres 2020). The United Nations has recently defined this as a “generational catastrophe” (United Nations 2020).

A final point to highlight from this body of research is the range of reasons for, and thus strategies needed to reduce, student absenteeism. There are multiple reasons why students miss classes, as well as large differences in the absenteeism rate among both individual students and student subgroups. Those seeking to develop effective policies to reduce absenteeism, especially chronic absenteeism, understand the need to examine the root causes—academic disengagement, socioemotional distress, economic challenges, health problems, and others. Initiatives that have been rigorously evaluated show that it is critical both to identify the specific reason(s) why a student is missing school and to respond with targeted, relevant supports.9 This point is particularly relevant in the current context, in which so many students are frequently absent for a variety of reasons that may be difficult for teachers and schools to know or address.

Figure A

Of course, the various approaches examined by the research on learning time assume two groups of students: those who are missing some learning time in school and those who are not. (In general, they compare “treatment” versus “nontreatment” groups to estimate impacts.) This comparison does not hold during the lockdown. Instead, all students are missing out on in-class instruction, and instead have been attending school remotely via various online arrangements that in some ways resemble homeschooling or online education. As discussed below, the evidence about homeschooling and remote education presents serious limitations, given their very different context, but nonetheless uncovers many issues that we will need to address in post-pandemic education.

Lacking the needed requirements for effectiveness, remote and alternative learning and online instruction during the pandemic has likely affected teaching and learning

The two main tools for education available to children during the lockdowns have been remote and alternative learning and, at least technically, a homeschooling environment. Evidence on these two modes make clear the conditions that would be needed in order for children to effectively learn under these conditions and for teachers to effectively teach under these conditions. As the following subsections show, most of these conditions have been lacking in recent months.

Research on effective online learning indicates it is critical that students have the tools and the experience

Online learning means, first and fundamentally, the shift from face-to-face learning to the use of devices of various sorts to deliver that learning. Successful online learning thus requires that students (and teachers) be familiar and proficient in their uses of those devices for learning. Of course, even more fundamentally, it requires that the devices exist. Here we discuss the needs of students.

We have limited knowledge about how much and for which purposes students have used devices and technology at home up to this point. An estimated 1.5 million K–12 students participated in some online learning in 2010 (Bettinger and Loeb 2017, based on Wicks 2010).10 Figure B uses PISA data from 2018 for the United States to show that, while students spent extensive time online prior to the pandemic, that time was heavily spent on social activities, browsing or seeking information, playing games, or accessing email. Students spent less time on educational activities, such as school work or communicating with other students or teachers. These findings suggest that over the past few months as children transitioned suddenly to online learning, they did so without necessarily having the practice or experience to learn well online, and that the transition required them to shift their device-use habits from leisure to studying. What we also know is that remote learning demands that children ignore the distractions that are now in front of their faces all the time and to which they, like all of us, are naturally drawn.11

Figure B

In addition to assessing quality and time, the literature on the use of devices assumes that all students have access to appropriate digital devices—i.e., it assumes no digital divide. As has been extensively documented, however, that is not the case. For example, García, Weiss, and Engdahl (2020) show that nearly 16% of eighth graders, or one in six who participated in the National Center for Education Statistics’ National Assessment of Educational Progress (NAEP) for 2017, do not have a desktop or laptop computer at home on which to follow their classes. And a small fraction of eighth graders, 4.2%, lack home internet, the other essential instrument for remote study. (It’s important to note that the survey questions do not ask about the quality or coverage of the internet access, or the number of computers in the house, and that the information predates the pandemic’s arrival. Devices once available for homework may now be shared with siblings or be used by parents for work.12)

A final caveat is that there is still limited evidence on the effectiveness of online education. A critical aspect highlighted by Bettinger and Loeb (2017) is that online courses are difficult, especially for the students who are least prepared.13 Research on performance of children attending virtual charter schools confirms the importance of self-engagement and parental supervision for success with this mode of education. Also, selection into these schools (students disengaged with  traditional schools enter these schools); worse inputs (teacher-to-student ratios, one-on-one instruction, etc.) than in traditional schools; and other features of these schools translated into negative effects on performance.14 Later in the report we discuss the requirements for successful online education from the perspective of teachers.

Research on home schooling makes clear that it works well for students under narrow circumstances

According to the NCES, close to 1.7 million students, or about 3.3% of K–12 students, were home-schooled in 2016 (NCES 2018).15 Parents who home-schooled their children cited the following as the most important reasons for doing so: concerns about the school environment, such as safety, drugs, or negative peer pressure; dissatisfaction with academic instruction at available schools; and a desire to provide religious instruction (Grady 2017).

In terms of its effectiveness, performance of home-schooled students is generally higher than that of their non-home-schooled peers. A review of 14 studies found consistent positive results in 11, mixed results in another study (some positive and some negative results), zero impact in another study, and neutral and negative effects in a final one. The estimate of the effects (based on eight of the 14 studies for which this information was available) ranged from very small (0.05 SD) to extremely large (1.13 SD) (Ray 2017a). Using percentile metrics, home-schooled students scored, on average, at or above the 84th percentile in all subject areas (Ray 2017b).16

While these findings may look promising, however, it is important to keep in mind two key considerations when interpreting these results. First, many more resources are devoted to home-schooled children, so they would be expected to perform higher, all else equal. Also, higher performance among home-schooled students may be due more to their selection into the category than the “treatment”/type of education they receive.17

Belfield (2004), for example, suggests that the improved outcomes among students who are home-schooled could be due to flexible instruction (without age-tracking), small “class sizes,” and dedicated parent-teachers who should make home schooling more effective than other forms of education. He also notes that “educational outcomes may be skewed toward those on which the family has competence, and educational progress may be slow if there is no formative assessment or peer-pressure to learn (although home-school parents may exert more pressure or have higher expectations as a result of their supervision).” More recent studies suggest that parameters such as structured or unstructured instruction may also be important drivers of the results (Neuman and Guterman 2016).

These underlying factors could be particularly relevant in the current crisis. Many of the same stark distinctions between effective and ineffective online education and home schooling would apply to the “emergency remote learning” done at home under a pandemic: students who entered the pandemic better off and those whose parents have been trained in instruction or have a particular ability teach would likely perform better than students whose parents have not been able to develop (or as successful at developing) those skills. In general, parents who were suddenly thrust into the role of home-schoolers had no such preparation; most are taking on that new task while juggling the full range of other home-care responsibilities as well as, in many cases, full-time remote jobs. That said, students whose parents have more formal education likely also have an advantage in this context—as they do in nonpandemic contexts—further compounding the disparities that low-income students are accruing (see, for example, Dinarski 2020; Rothstein 2020; Belfield 2004; Goldstein 2020a).18

Evidence on online instruction emphasizes that teachers also need training and supports

As the discussion of successful versus unsuccessful remote and online learning reveals, there are multiple requirements needed for online education to work as intended and deliver positive results. Just as the requirements for effective student learning have largely not been met during the pandemic, the same is true for effective online instruction.

First, there was little time to design and develop instructional tools for wide deployment.19 As a recent analysis of research on the subject details,

Online education, including online teaching and learning, has been studied for decades. Numerous research studies, theories, models, standards, and evaluation criteria focus on quality online learning, online teaching, and online course design. What we know from research is that effective online learning results from careful instructional design and planning, using a systematic model for design and development. The design process and the careful consideration of different design decisions have an impact on the quality of the instruction. And it is this careful design process that will be absent in most cases in these emergency shifts.20 (Hodges et al. 2020)

Moreover, it is hard to plan and to design effective instruction for the COVID-19 era when teachers and school districts don’t have a framework (or even the right language) to accommodate what they are doing. As Hodges et al. (2020) emphasized when exploring how colleges and universities were coping with the sudden and rapid shift to remote learning (in March 2020), understanding the current circumstances required distinguishing between online or remote learning generally. For our current context, they suggested the term “emergency remote teaching,” which helps signal the uncertainties and unknowns that could affect teachers’ instruction.

Second, weak systems of support, including lack of professional development on how to integrate computers into instruction, have left teachers less than optimally equipped to teach during the pandemic.21

Slightly over two in three public school teachers report having participated in professional development activities on the use of computers for instruction in the past 12 months, as shown in Figure C, based on García and Weiss 2019 using data from the 2011–2012 Schools and Staffing Survey (SASS).22 But those who participated in these activities were not broadly satisfied with them. Among these teachers, one in four found the activity very useful, with about one in three finding it either not useful or just somewhat useful. And teachers who participate in such activities have to surmount barriers to do so, as access to work time and supports to participate in professional development are very limited. Among all teachers, only half have released time from teaching to participate in professional development (50.9 percent), and less than a third are reimbursed for conferences or workshop fees (28.2 percent).23

Figure C

The limited training pre-pandemic is compounded by the limited technical support during the pandemic. Most K–12 teachers did not contemplate online instruction until being forced to do so by the pandemic. As a result, teachers have had to come up with a variety of options on the fly, from assigning daily or weekly coursework that students turn in online to full classes conducted via Zoom and a range of approaches in between. We can expect that some of these online strategies launched during the COVID-19 crisis did not lead to optimal outcomes.

Third, inadequate systems for tracking attendance online leave teachers in the dark on a key “input” of education: student learning time. Even the most well-trained teacher when it comes to online instruction won’t be effective if his or her students are not online and following instruction. At the most basic level, schools are trying to assess how broadly and consistently students are interacting with teachers and receiving instruction. One ambitious effort has been in Southern Florida, where districts rigorously track attendance and contact parents when students are absent. Quickly recognizing that relying on student log-ins failed to capture much of the activity taking place, districts in Palm Beach County and the Florida Keys ask teachers to log student participation in online forums and completion of assigned work. In general, schools in this system are seeing attendance that is only modestly lower than normal, with the biggest drop-offs among the youngest and oldest students (who, respectively, need parents’ help to get online and are least motivated to take part). However, while the system helps monitor potential race- and class-based disparities in attendance, concerns remain (Bakeman 2020). Attesting the importance of attendance, some school districts that have chosen online instruction for the beginning of the 2020–2021 school year are making registering attendance compulsory through their platforms.24

Fourth, the emotional bonds critical to any kind of learning are just as important for remote learning or home schooling but hard to attain in the current crisis. Even more so than college professors, K–12 teachers also need to retain emotional bonds with their students, especially younger ones, that can be extremely difficult to attain remotely. Many of these teachers are also parents and so must juggle their children’s activities, such as helping their children with homework, with their own job responsibilities. And teachers working with particularly vulnerable students face additional challenges as some of these students lack access to computers to work or even enough internet bandwidth (see barriers to access described below).

The “whole-child” development that occurs at school was also interrupted during the pandemic

For children, going to school is not just about learning reading and math: it’s also about developing the social and emotional skills critical to succeeding in life. School closures eliminated some of these critically important aspects of school beyond academic activity, such as the development that occurs through personal relationships among students and between students and teachers, after-school activities that support children’s mental and emotional well-being and skills development, and a sense of routine. In addition to the cessation of their normal activities at school, during the pandemic, children have lost in-person contact with relatives and friends and have witnessed many sobering daily life realities, from parents who may be unsure where the next meal or rent payment will come from or who are working risky jobs in order to make ends meet, to family members fearing that loved ones are in danger of serious illness or even death. Overall, the crisis has helped highlight the importance of other skills that are often overlooked in the school context, but that should be nurtured as part of going to school and that will merit more attention in the aftermath of the pandemic.

A range of skills often referred to as socioemotional or noncognitive skills—including creativity, tolerance, persistence, empathy, resilience, self-control, and time management—have long been neglected in education policy, which has tended to follow the so-called cognitive hypothesis (Tough 2012; Ravitch 2011, 2020; Rothstein, Jacobsen, and Wilder 2008).25 These noncognitive skills are deemed lower priorities in academic contexts—including skills that children typically lagging behind could have an edge in—and their integration in the usual components of learning and teaching is far from standard. As a result, when decisions about curriculum, standards, and evaluation are made, socioemotional skills tend to be the last on the priority list and the first on the chopping block, while testing highly on math and reading—skills that tend to be correlated with having more educated parents and higher household incomes—is richly rewarded in school, furthering “deficit” narratives (faulty messages about who can and cannot succeed in school, and about what succeeding in school means).

For sure, parents and teachers have long been attuned to the broad range of life skills that their students need to develop, but this crisis has sharpened that focus. The sudden need for children across the board to adapt to uncertain and rapidly changing circumstances and to cope with new levels of trauma make it all the more urgent to address this disparity between what parents and teachers understand about the breadth of skills critical to child development and systems that focus on testing a narrow set of cognitive skills. For example, resilience—the ability to adapt to and thrive in different situations—along with persistence and self-control have gained new recognition as important life skills during these months of the pandemic. Children transitioned to online learning overnight and have had to follow classes without the direct supervision of the teacher or the interactions with other students, which requires a higher than usual degree of self-control and persistence. Creativity is another skill that likely is serving children well during this crisis: Students who find new ways to keep themselves engaged and to make forced isolation productive are benefiting, while their peers who are easily bored are losing ground.

As we slowly move forward during the pandemic and we return to “normal,” it is going to be more important than ever that we do not let this recognition of whole-child development fall away and revert to a narrow focus on academics. Doing so would cause harm on several fronts. First, it would ignore and potentially exacerbate the trauma that many children are experiencing. Second, it would put low-income students even further behind—both by weighing heavily the areas of learning that they have been least able to access and by failing to recognize the natural variation in students’ strengths across a broader range of skills, or “patterns of thoughts, feelings, and behavior” (Borghans et al. 2008). And finally it would miss a unique opportunity to better balance what schools can do. Noncognitive skills are demonstrably as important as other cognitive skills when it comes to ensuring that children will thrive both in school and later in life. Moreover, since academic and socioemotional skills develop in tandem, and in recognition of the added challenges during the pandemic, it will be more critical to approach skills development holistically and make teaching and nurturing the whole child central, rather than marginal (see García 2014 and García and Weiss 2016 for a summary of this literature).

Recessions, natural disasters, and pandemics disrupt learning the most when there is no contingency planning

As noted above, prior research on circumstances somewhat similar to the shutdown during the pandemic is important to review—findings from this research may not be directly applicable due to substantial differences in the circumstances, but understanding the mechanisms through which learning occurs under these circumstances, as well as how to be prepared for the upheaval, is critical to informing our way out of this current crisis and our readiness for future ones. This is particularly the case regarding evidence from the research on “education in emergencies,” which examines the provision of education in emergency and post-emergency situations due to pandemics, other natural disasters, and conflicts and wars, generally in poor countries around the world.26 The practical recommendations from this field have been largely ignored in the education policy arena until now, because they have not seemed to apply in the rich countries.27 However, there are some exceptions overall and for the United States in particular, including cases of natural disasters such as Hurricanes Katrina and Maria.

The following lessons can be extracted from this research: Emergencies lead to undeniably negative impacts on educational processes and outcomes; the most disadvantaged population subgroups experience the largest, and most lasting, negative consequences; and contingency plans—absent during the ongoing pandemic—are of critical importance. Providing education, often made available because of these plans, leads to positive outcomes to children and societies. Moreover, emergencies tend to strain existing resources, adding additional challenges.

We summarize here a few key findings. For example, by the end of the school year following the devastation that Hurricanes Katrina (August 2005) and Rita (September 2005) brought to New Orleans, the performance of students who were displaced dropped by 0.07 to 0.22 standard deviations relative to what their performance would have been without the hurricanes (this range includes an average across subjects and grades calculated by Pane et al. [2008] and estimates by Sacerdote [2012] on math and reading). Principals reported that students who were displaced were judged more likely than students in the control groups to engage in negative behaviors, such as fighting, violating school rules, arguing, bullying, playing in isolation, and eating in isolation, and more likely to need mental health counseling; they were also judged less likely to engage in positive behaviors, such as participating in before- or after-school clubs or activities, school-sponsored social events outside the school day, or sports teams (Pane et al. 2008). Sacerdote (2012) also found longer-run effects, including rates of college attendance that were one to four percentage points lower relative to trends measured in cohorts not affected by the natural disasters.28 Importantly, Özek (2020) finds that some of the negative effects of disasters on students mostly vanish after the first year when there is an “adequate compensatory allocation of resources.” Among the resources he cites as critical to compensating the negative effects of emergencies on learning are teachers—specifically ensuring that the most effective teachers are working with the most vulnerable students. Although, as noted, Özek (2020) found that first-year effects tend to decline, effects persist in the second year in high-poverty schools and in low-performing schools.

Natural disasters and recessions also create economic shocks. Research exploring the consequences of recessions such as the Great Recession sheds light on ways today’s economic crisis is likely affecting children’s education. For example, Irons (2009) discusses the ways that “unemployment and income losses can reduce educational achievement by threatening early childhood nutrition; reducing families’ abilities to provide a supportive learning environment (including adequate health care, summer activities, and stable housing); and by forcing a delay or abandonment of college plans.” Shafiq (2010) also discusses potential negative effects from economic shocks, such as long hours worked by parents, which “reduces the time that parents can devote to assisting their child with homework, reading, and other educational activities.”

Economic shocks in turn lead to cuts in education budgets. Jackson, Wigger, and Xiong (2018) show that spending cuts enacted during the last recession had detrimental effects on education outcomes: the per-pupil spending cuts that states made during the Great Recession (by roughly 7% overall, by over 10% in seven states, and by more than 20% in two states) reduced college enrollment and test scores, particularly for children in poor neighborhoods, and the impacts of these cuts were greater for Black and white students than for Latino students. Jackson, Wigger, and Xiong (2018) estimated that the impacts of such large-scale and persistent education budget cuts are very significant: a $1,000 reduction in per-pupil spending led to a reduction in test scores of about 0.045 standard deviations and a roughly 3 percentage point decline in the share of high school students who go to college. Often, recovery after a shock never fully happens, as explored in more detail later in our report.

The education-in-emergencies research underscores that “contingency plans” are critical to dealing with emergency and post-emergency situations. Specifically during crises arising from war, conflicts, natural disasters, and pandemics, children are displaced often as homes, neighborhoods, and schools are destroyed—and this may threaten survival or inflict some level of trauma upon children.29 A certain level of preparedness is critical in order to provide an effective response at the onset of a crisis, and to “prepare, cope, and recover” (UN IASC 2007, 2015; Anderson 2020; Azzi-Hucktigran and Shmis 2020).

Although it is expected that countries and their education agencies have a plan to deal with short-run disruptions (i.e., snow days, flu season, etc.), such expectations are uncommon when it comes to contingency plans for larger, longer emergencies. Most information including guidance on planning for education in emergencies comes from several international organizations involved in major, longer-term emergencies. One exception is a reference in a White House publication reviewing assistance provided after Katrina; these words should be heeded in the aftermath of this pandemic:

Individual local and state plans, as well as relatively new plans created by the federal government since the terrorist attacks on September 11, 2001, failed to adequately account for widespread or simultaneous catastrophes.…The President made clear that we must do better in the future. The objective of this report is to identify and establish a roadmap on how to do that, and lay the groundwork for transforming how this Nation—from every level of government to the private sector to individual citizens and communities—pursues a real and lasting vision of preparedness. To get there will require significant change to the status quo, to include adjustments to policy, structure, and mindset. (The White House 2006)

As has been evident in the past few months, there was no national education plan in place to deal with medium-run or long-run emergencies for the scale of COVID-19. Existing plans (as indicated, outlined by international organizations) offer “contingency planning tools” to ensure appropriate arrangements are made to analyze the impact of potential crises and to respond in a timely and effective way. The strategies suggested are characterized as flexible learning approaches, which reflect the reality that the circumstances and needs vary widely. Continued provision of education is expected to support both learning and the psychosocial well-being of both students and educators (Anderson 2020). Some strategies aim at promoting cognitive, emotional, and social development through structured, meaningful, and creative activities in a school setting or in informal learning spaces that replace the unavailable traditional schools. In other words, these programs are designed to provide support similar to that provided by good school systems on a regular basis.30

Clearly, there are potentially relevant aspects of research on emergency education that, where emergency education resembles the COVID-19 situation, could help policymakers identify what needs to be done immediately and going forward to help schools and students recover. Before we discuss these, we devote much of the next section to assessing how this crisis is expected to have worsened impacts on vulnerable subgroups, and to exacerbate inequities overall.

How is COVID-19 exacerbating opportunity gaps (and what steps are schools taking in response)?

The COVID-19 crisis has exacerbated the well-documented opportunity and enrichment gaps that put low-income students at a disadvantage relative to their better-off peers. By opportunity and enrichment gaps, we mean gaps in access to the conditions or resources that enhance learning and development between low-income students and their higher-income peers (with low-income students less likely than their better-off peers to access these conditions and resources). Before we delve into the details, it is important to state that this should not come as a surprise. The baseline operating status of the education system in the United States before the pandemic had severe problems with regard to equity. Put simply, as a nation, we have structured the education system to deliver the disparate outcomes that it delivers, i.e., outcomes that differ by social class, minority status, and other student characteristics: “It’s not a coincidence or accident” (ASI 2020).31 Here we briefly describe a few of the gaps that are most directly relevant to students’ abilities to learn during the pandemic: basic needs, economic relief, and support for families and health. We also discuss how the pandemic has exacerbated the limitations of standardized assessments, especially when used to measure performance gaps in education.

There are two important caveats to this discussion. First, any recent statistics are preliminary (and likely quite conservative). Second, there are, of course, other gaps that we are not able include here—for example, in wealth through homeownership or toxic stress linked to structural racism (Lerner 2020; Morsy and Rothstein 2019)—but that are interacting with and compounding those factors that we are able to examine. As leading education and civil rights organizations summarizing the breadth of the opportunities and enrichment gaps note, “the transition to educating students in their homes or shelters has exposed and exacerbated inequities in education, food security, and housing that have long existed” (AFT, LDF, and Leadership Conference 2020). We add health and mental health to that list, and we emphasize the critical role schools play as part of the social safety net and as the first responders to children’s basic needs (Kirk 2019; Weiss and Reville 2019; ASI 2020).

The pandemic has exacerbated opportunity gaps associated with uneven access to food and nutrition, shelter, health insurance, and financial relief measures

The disruption caused by the pandemic and the interruption of the normal operation of schools continue to pose barriers to meeting the most basic of children’s needs (access to food and nutrition and shelter). Families’ resources also have been largely impacted by the economic downturn that followed the disruption. There is overwhelming evidence that low-income children and their families have much less access to nutrition and shelter, that children of color and children from immigrant families are disproportionately affected, and that this lack of access has palpable consequences for their development. It is no secret that the inequities are built into our economic and policy setups, and that these inequities affect children’s development as well. The school shutdowns and economic crisis caused by the pandemic are exposing and exacerbating these challenges.

Evidence on expanded opportunity gaps due to lack of access to food and nutrition

In 2013, as the United States was still recovering from the recession of 2007–2008, half of all public-school students were eligible for free or reduced-price school meals (SEF 2015; Carnoy and García 2017). In other words, years into the economic recovery, a record share of one in two public-school students lived in a household that was unable, absent government support, to consistently feed them. With millions of adults newly out of work due to the economic shock of the coronavirus pandemic—and federal relief insufficient, slow, and difficult to access—many more children are now in food-insecure homes (i.e., they have limited or uncertain access to adequate food, as measured by responses to survey questions about access to food).

Using data from the new Household Pulse Survey (HHPS) from the U.S. Census Bureau, 29.8% of respondents with children were food insecure (Schanzenbach and Tomeh 2020). Bauer (2020) estimates that there were almost 14 million children living in a household characterized by child food insecurity during the week of June 19–23, 2020, “5.6 times as many as in all of 2018 (2.5 million) and 2.7 times as many as during peak of the Great Recession in 2008 (5.1 million).”32

The data about food insecurity is backed up by news reports showing record levels of visits to food banks during the early part of the pandemic and the shortage of resources to meet the demand for food. According to Feeding America, one in seven Americans relied on food pantries before the pandemic, with demand doubling or tripling in many places in the first weeks of the crisis. By late April, less than two months into the pandemic, food pantries in Chicago and Houston were almost out of staples, and one third of New York City’s food banks had closed due to lack of supplies, donations, and/or volunteers (Conlin, Baertlein, and Walljasper 2020).

Schools continuously tried to fill the void to the extent they could, with buildings that were closed for instruction reopening as places to collect, prepare, and distribute meals. Some schools were serving breakfast or dinner or are giving out weekend meal “packs” for students, and many provide meals for older and younger siblings as well. For example, schools in Anne Arundel County, Maryland, served an average of 8,000 meals—three per day—for the first 39 days of the pandemic, hitting the one million mark on May 12. District Superintendent George Arlotto said of the importance of supporting his students, “We know if we’re not serving meals they might not be getting fed, at least certainly not three meals a day” (Streicher 2020).

However, difficulty matching meals to parents’ schedules and lack of sufficient transportation to deliver meals limited many districts’ ability to serve the students they normally serve. Across the Denver metro region, district capacity during the first month of school closure starting in March spanned a wide range, serving just 12% of students in the largest and lowest-income district, Denver; 16% in Jeffco; 34% in Aurora; and 57% in the Adams 12 Five Star Schools district (Meltzer, Robles, and LaMarr LeMee 2020).

Across the country overall, the networks set up to provide meals left out a large proportion of children. “Only 61.0% of parents whose families received free or reduced-price meals during the school year reported receiving school meal assistance during closures,” noted Waxman, Gupta, and Karpman (2020), who also found that 17.2% of parents living with children under age 19 reported receiving charitable food in May 2020.

Evidence on expanded opportunity gaps due to lack of access to shelter

In addition to children who are especially vulnerable during the pandemic because they rely on schools for basic food and nutrition are children who are homeless. Data show that before the pandemic began, large numbers of students in districts across the country were homeless.33 For this numerous group of students, getting an education remotely is unthinkable. With millions of adults newly out of work due the economic shock of the coronavirus pandemic—and eviction bans expired or expiring in localities around the country—unstable housing is putting the challenges of educating homeless students into starker relief. Some school districts are paying attention to the needs of their homeless students. In San Jose, California, for example, some schools are expected to be open for counseling and in-person instruction for homeless and special needs students (Lambert, Burke, and Tadayon 2020). The United States Interagency Council on Homelessness (USICH 2020) has issued some general guidelines as to how school districts can work with local public health officials and community partners to identify temporary, safe, and stable shelter options for families or youth experiencing homelessness who must quarantine. The agency also provides guidance on ensuring homeless children’s access to remote education while schools are closed.

Evidence on expanded opportunity gaps due to unstable employment and lack of access to financial relief and health insurance

Loss of work has hit families across the board, as initial unemployment shocks in the travel and entertainment industries expanded to shut down restaurants, retail, and even some of the health care sector shortly after the pandemic started. While some of those jobs have returned, we still have extremely elevated rates of unemployment and loss of health insurance. And low-income parents are in particularly tough situations because of the low-paying and unstable nature of their jobs. Those who lost already-precarious non-standard jobs (like “gig” work and other independent contracting work) don’t qualify for unemployment insurance (and many had trouble accessing emergency unemployment benefits because of outdated state systems). Further, many workers around the country who had job-related health insurance lost it just when they needed it most (Cooper and Worker 2020; Bivens and Zipperer 2020). While Congress passed relief measures earlier in the pandemic, some key components of relief—such as the extended unemployment benefits—have expired, and further measures are at this writing stalled in Congress (Gould 2020a; Shierholz 2020). Not granting the needed economic relief and not granting more support for families is going to add to the challenges of parents who have dual responsibilities of supervising children’s learning and putting food on the table and providing them with health protection.

Evidence on expanded opportunity gaps due to health challenges for families

The pandemic obviously also raises the possibility that children’s families and children themselves are grappling with illness and even death. Research shows that the health risks are higher for workers in low-paying professions than for workers in high-paying professions because the former are much less likely to be able to work remotely (Gould and Shierholz 2020). Moreover, essential workers—such as warehouse stockers, home health aides, and delivery and trash truck drivers—now risk contracting COVID-19 while still struggling to survive on low wages.34

Thus it is not surprising that this crisis has also resulted in an increase in the number of children who face the serious illness or death of a relative. It seems likely that a large share of low-income students and Black and Hispanic students now resuming schooling have suffered major trauma. With Black students losing family members in disproportionate numbers, the pandemic is exacting a particular toll on these communities (Harper 2020). For example, in Georgia, where African Americans make up just 30% of the state’s population, they represent over 80% of COVID-19-related hospitalizations and more than 50% of deaths (Weiner 2020). When New York City was the epicenter of the pandemic in the United States, the heavily white borough of Manhattan had a hospitalization rate of 3.31% and a death rate of 1.22%—the city’s lowest—despite having the oldest residents of any of the city’s five boroughs, while the heavily low-income, African American borough of the Bronx had the highest rates, 2.24% and 6.34%, roughly double those of Manhattan (Wadhera et al. 2020).35

Evidence on expanded opportunity gaps due to health challenges for students

These same groups of students—Black and Hispanic students, and low-income students— suffer academically due to physical and mental health problems that are less likely to be addressed in a timely and consistent manner (Ghandour et al. 2018; Menas 2019; Morsy and Rothstein 2019). Many rely on school-based health clinics, a critical resource that is no longer available in schools where teaching is not occurring on site. Earlier in the pandemic when access to doctors’ offices was severely limited (with many serving only urgent cases) and hospitals were overwhelmed (and perceived as unsafe), problems from toothaches and ear infections to emotional breakdowns went untreated and, in many cases, became much worse. When the state of Florida shut down in late March, for example, it banned all nonemergency medical and dental services, leading to questions as to whether even check-ups conducted prior to procedures were permitted (Boca News Now 2020).36

With both physical and mental health on the line for stressed-out students, school districts are trying to leverage newly available resources to compensate. These include additional Medicaid resources provided in the first federal COVID-19 relief legislation, the Families First Coronavirus Response Act. That act temporarily increases the federal Medicaid match to states that agree to maintain current eligibility standards and cost-sharing requirements and limit disenrollment. Relaxed guidelines enable states to use some of that money for telehealth services without additional authorization, so students can see doctors remotely as needed. The federal CARES Act that was enacted in March provides $13.2 billion for K–12 schools as part of Title I funding, and it includes several aspects of student health in allowable uses. The Los Angeles Unified School District has used some of that funding to launch a mental health hotline for students. Superintendent Austi Beutner notes, “Their world has been turned upside down and we need to make sure students have the support they need [during this crisis]” (Jordan 2020a).

All of the above challenges, of course, mean more stress. And for children who were already living in cramped and less-than-ideal situations, having all family members in the house makes the regular challenges of daily life much greater. Increased incidences of abuse due to confinement, stress, and lack of access to outside support further affirm the urgency of addressing the stressors that are affecting families and, in turn, their children’s development and ability to learn (Stratford 2020; Greeley 2020; Tolerance Trauma 2020).

The pandemic has exacerbated opportunity gaps in teaching and learning

It is in these challenging contexts of economic insecurity and housing instability that students (and teachers) were suddenly transitioning to remote learning, adding another class- and race-based disparity in education opportunity: the “digital divide.” The “digital divide” refers to the fact that some children do not have access to the devices or internet services needed to operate online—and there is a double digital divide that arises from the fact that low-income children and Black and Hispanic children are more likely to lack this access (García, Weiss, and Engdahl 2020; Tinubu Ali and Herrera 2020). Research on the digital divide counters the idea that all children can access online instruction and the education system shifted to online education. Given the resurgence of COVID-19 cases over the summer and the growing number of school districts announcing plans to begin the 2020–2021 school year totally remotely, the divide would only continue in the imminent future. Some low-income families are struggling to obtain a computer or other device for each child, with a share of families lacking an internet connection enabling children to do assigned work online or a quiet space to do solo work (let alone attend the Zoom calls that classrooms are now conducting; see Hodges et al. 2020).

Our analysis of data from the 2017 National Assessment of Educational Progress shows that digital devices are not universally available or used at home for school-related purposes. Our findings are presented in Figure D. Specifically, 84.4% of eighth graders overall, and 76.3% of poor eighth graders have a laptop or computer, which means that about 16% of eighth graders and 25% of poor eighth graders have no desktop or laptop at home. In addition, only about half of eighth graders had experience using the internet at home frequently for homework, with a much larger share of non-poor students (56.1%) than poor students (46.4%) accustomed to using the home internet frequently for homework (a gap of 10 percentage points). (We define poor students as students who are eligible for the federal free or reduced-price lunch programs, and non-poor students as students who are ineligible for those programs.)37

Figure D

Our analysis of 2017 NAEP data also shows that teachers are not universally prepared to teach online, as also shown in Figure D. Just about a third (32.5%) of eighth graders overall have teachers who consider themselves proficient in using software applications, and only a fifth (19.3%) have teachers who consider themselves proficient in integrating computers into instruction. The shares of students overall with teachers who don’t consider themselves proficient but who have received some training in applications and in computer use in instruction are higher (43.4% and 69.2% respectively). Yet that still leaves nearly a quarter (100% minus 43.4% minus 32.5%, or 24.1%) of eighth graders with teachers who are neither proficient in nor trained in software applications, and close to one in eight (100% minus 69.2% minus 19.3%, or 11.5%) with teachers who are neither proficient in nor trained in how to integrate computers into instruction.

A Southern Education Foundation report on class- and race-based disparities during the COVID-19 crisis finds similar disparities in access to the resources needed for online learning. It notes that nearly one in five African American children and a slightly greater share of children in low-income households have no access to the internet at home (Tinubu Ali and Herrera 2020). These disparities mirror those reported by superintendents who responded to a survey by AASA, the School Superintendents Association, in late March as schools across the country were closing down (Rogers and Ellerson Ng 2020).38 Numerous news outlets reporting on the digital divide have also noted these disparities by race and ethnicity (for example, see Kamenetz 2020b). School shutdowns and associated internet- and device-access challenges have been occurring at a time when many of the public libraries that have been a resource for families without computers or home internet access are closed due to the pandemic.

School districts are trying hard to take these challenges into consideration and to make up for the large disparities they know their students face. Some, like Montgomery County, Maryland, are sending home Chromebooks and tablets, prioritizing students who are eligible for free- and reduced-price lunches or are known not to have devices at home (St. George 2020). Others, like New York City, are lending iPads to students who need them (NYC Department of Education 2020). All of this takes time, however, and many districts lack the resources. (Montgomery County provided paper packets to students for the first few weeks of closures, until it could distribute the Chromebooks.) Some districts are making online work optional, as a way to not further disadvantage students who physically cannot do it, but of course that can weaken schools’ capacity to continue to instruct.

Tinubu Ali and Herrera (2020) also report on dozens of innovative strategies districts have employed to overcome some of these disparities. These strategies include deploying roving school buses that add Wi-Fi coverage in South Carolina, the purchase of thousands of additional hotspots in Texas, and two months of free internet in Caldo Parish in Louisiana thanks to a partnership between Comcast and the local NAACP. (Comcast is also providing free access in Montgomery County, Maryland.) In Tennessee, Staples is printing and distributing printed materials free of charge to students who cannot afford the cost, and public schools in Jackson, Mississippi, are developing a package of learning materials that are paper-based or online and shared via the state’s educational programming television channels. South Carolina’s public television network is providing free virtual professional development sessions on home learning and technology best practices. In Miami-Dade, one of the most diverse school districts in the country, instructions for families are provided in English, Spanish, and Creole.

The pandemic has exacerbated the limitations of standardized tests

Digital divides and disparities in parental resources are fueling the growth of opportunity gaps that likely will make it harder for disadvantaged students to engage with their schoolwork and easier for these students to lose interest in school. If so, the pandemic will also widen performance gaps between disadvantaged students and their better-off peers and increase graduation and school dropout rates among disadvantaged students, particularly if districts don’t adjust practices to reconnect with these students.

Thus, one practice that may need adjusting or revisiting is testing. During the pandemic, traditional assessments—which have limited value even in normal contexts—are much less useful in capturing what students know and have learned. These assessments could feel “overwhelming or condemning to children” at a time when it is necessary to create opportunities for students to show what they know and to demonstrate where they are, and for teachers to adjust instruction to students’ current development in order to advance their development and potential (RESEARCHED 2020, NPE 2020). As set forth above, students have very uneven access to the online resources they need to take tests, let alone complete them effectively. Similarly, students have uneven access to the special instruction and supervised practice that help students pass these tests—with lower income students and Black and Hispanic students less likely to have access than their higher income and white peers. This means that standardized testing during the pandemic will deliver results that are, by design, going to be even more closely correlated with life circumstances than is true during periods of regular classroom instruction. Compounding all of the barriers to meaningful and equitable monitoring and testing during the pandemic, teachers in remote settings lack the tools that they have when they are in their classrooms to interpret test results. In other words, in a classroom, teachers are more able to distinguish between a low score likely due to the student’s lack of understanding of the material versus a low score due to the student’s frequent absences, emotional distress, or other factors. As a result, teachers working remotely are hard-pressed to respond to a test score with an appropriate strategy to support the student.

For all of these reasons, traditional standardized tests have limited value in this context and may do more harm than good.39 Rather, school districts should be using tests that are designed to assess where students are across a range of areas and to help teachers meet students there. These tests include diagnostic tests, formative tests, SEL assessments, and assessments that can be performed remotely such as project-based assessments and capstone projects.40 These types of tests will be critical to helping students and teachers alike start to dig out of the academic hole dug by the COVID-19 shovel.

Going forward: Translating what we have learned into a plan for the “three Rs” of relief, recovery, and rebuilding

Throughout the coronavirus pandemic, we have made choices about how to sustain, or provide relief to, the education system. We have also had the opportunity to consider how best to proceed as we start to recover, and how to rebuild the system by taking more decisive action on substantial, long-needed changes. Indeed, how well we rebuild the education system will determine how well we address the impacts the pandemic has had on our human capital and how prepared we are for shocks of this nature in the future.

As noted above, students have seen their normal learning and development interrupted and disrupted. Inevitably, this will lead to lost ground during the pandemic, with disadvantaged students particularly vulnerable given the way that the pandemic has compounded large existing opportunity gaps. We propose a set of targeted education interventions and comprehensive services to lift up disadvantaged children and reduce inequities as we move out from this pandemic. This plan tackles today’s three Rs—relief, recovery, and rebuilding—with a phased three-stage process that must be properly funded at each stage.

Specifically, this three-pronged plan requires making the necessary investments to 1) put school systems on a solid footing to provide effective remote instruction and supports at scale as the crisis continues to play out (the “relief” phase); 2) make new investments to help schools and students compensate for lost time and ground during the period of quarantine (during the “recovery” phase); and 3) lay the foundations for a shift toward an education system that understands the complexity of education production and its multiple components, untaps children’s talents, works equally for all students, and reflects the value we place on education as a society (in the “rebuilding” phase). This plan will require substantial amounts of resources and strong collaboration and effort.

If the Great Recession is any indicator, competition for resources will be fierce. In fact, early indicators are that this public health crisis will pose enormous challenges for states and local governments, those responsible for over 90% of the school systems’ revenue.41 Moreover, we entered this crisis in a more difficult position than in the Great Recession (based on a comparison with what we learned from the 2009 federal stimulus, and from the fact that about half of the states as of 2016 had yet to return to the level of per-student spending that they had attained prior to the Great Recession).42

With state budgets at historic crisis levels and the economy continuing to struggle,43 the prevailing narrative will likely be an even more severe version of “we can’t afford that” than what we experienced in the aftermath of the Great Recession. It will therefore be more important than ever to meet that assertion with the fact that “we can’t afford not to.” All of the evidence we have amassed demonstrates that not spending costs far more, and delivers far less, in the long run, than making the needed investments.44

Underlying the fiscal barriers to making the needed investments in education is a lack of leadership at the federal level that makes it very difficult for states to do what is needed. So far, there has been insufficient, scattered attention to education from policymakers, but even that has had a marked political tone that fails to acknowledge challenges or provide required resources.45

Relief: Give schools urgent resources so that they can provide effective remote instruction and supports at scale during the pandemic

During the pandemic, schools have been challenged with not only fulfilling their main roles of educating our children but also serving as a key part of the safety net: Specifically, to some degree, schools have provided not just remote education but also supports like meals, health services, counseling, and, in some cases, housing. Given the fact the schools are not universally going to be resuming standard operating procedures in the foreseeable future, policies must be enacted to enable all schools to provide effective remote instruction and supports consistently, and at scale.

While states and school districts are critical players in the relief stage, most of the calls for action involve the federal government because states and school districts are not only overstrained but also facing imminent budget cuts caused by the pandemic, with an inability to incur deficit spending.

Congress must resume consideration of additional relief measures and pay more attention to schools and associated public supports, including child care, social services, food and nutrition supports, and physical and mental health care—devoting substantially larger shares of, and sufficient, funding to these needs. At a minimum:

• Every school must be equipped and have the necessary resources, in conjunction with both public and private community institutions, to feed children (and, as relevant, their families) for as long as the current crisis demands.
• Federal aid that enables schools to provide counseling and other mental health supports should be expanded and extended to meet the large and growing needs of our students.
  • These needed services include the various wraparound supports specific to physical and mental health services, and to countering the various negative impacts of the crisis on the mental and emotional health of both students and educators.
• Every school district must develop a system to monitor what its students need in order to be able to learn while remote education prevails.
  • During the first months of the pandemic, the lack of preparation to cope with the lockdowns meant that many children lost access to the most basic needs. School districts must coordinate with state and local agencies and partner organizations to assess students’ needs so that districts understand their students’ situations and can respond accordingly.
• Every school district must be provided with resources and technical support to narrow the digital divide, in terms of both internet access and equipment (computers, tablets, etc.)
  • Unlike during the first months into the pandemic, access to online education must be universal.
  • Schools must be equipped to do needs-based monitoring of students’ status in terms of internet access; their access to computers and other technology tools for online learning; and students’ capacity to make effective use of the tools they have. This type of diagnostic assessment of technology and access is critical to understanding the degree to which students can engage with instruction on a regular basis and is foundational to their ability to learn.
• The remote instruction students receive needs to be of high-quality, and to attend to unique needs including those of special-needs students and English learners.46
  • District and school leaders should provide teachers with the necessary training and preparation to avoid unstructured instruction and the kind of “trial-and-error” instruction many had to employ during the first months of the pandemic.
  • District and school leaders should survey teachers as to the specific professional development and other supports they need to teach effectively in these adapted contexts, and Congress should allocate federal aid to ensure that all teachers obtain the needed support.47
  • Given that many teachers, like other “essential workers,” must balance instruction with attending to other household realities, including parenting their own children, Congress should ensure that support for child care is included in key relief measures.48

In the “relief” phase, schools must also have the resources they need to safely operate with partial on-site instruction if the health protocols allow for doing so.

• Every school district must have established a plan to meet the COVID-19 required safety measures, following the guidance from public health experts and educators.49
  • These plans at the very least must include communicating, educating, and reinforcing appropriate hygiene and social distancing practices in ways that are developmentally appropriate for students, teachers, and staff; maintaining healthy environments (e.g., cleaning and disinfecting frequently touched surfaces); repurposing unused or underutilized school (or community) spaces to increase classroom space and facilitate social distancing, including outside spaces, where feasible; developing a proactive plan for when a student or staff member tests positive for COVID-19; conducting case tracing in the event of a positive case; etc.
• Every school district must receive the resources to ensure the safety guidelines are disseminated, understood, and followed. Ensuring that guidelines are followed includes providing the financial resources and the equipment so that members of the school community are protected, the facilities are cleaned, and staff members have what they need to be safe.50

Recovery: Provide extra investments to help students and schools make up lost ground as they return to in-school operations

When schools resume their operations back in the classroom, it will be critical to fully understand which students have been engaged and to what degree, how much they have learned, and where they have fallen behind. But for meaningful teaching and learning to take place, educators must first be able to assess their students’ well-being and readiness to learn. Once they achieve that, educators will need sufficient, appropriate resources and tools to enable students to catch up and continue their development.

• School districts and the broader education system must prioritize diagnostic- and curriculum-embedded assessments, pausing or waiving upcoming state and other performance assessments to allow teachers to meet students where they are.51
  • Careful use of well-designed diagnostic tests will be critical to preparing and equipping schools and teachers to do their jobs, which will include adjusting instruction as necessary, and thus to helping students make up for disrupted education.
  • These assessments can also provide critical help to teachers who are trying to prevent disengaged students from ending up dropping out of school.
    • Using diagnostic assessments to assess the needs of the pandemic can provide a model for using assessments more appropriately in the future—i.e., as formative and informative tools of teaching and learning, rather than as evaluative tools of judgment.52
  • Teachers must receive training in interpreting diagnostic assessments and using them to enhance instruction.
    • Educators must receive training not just on diagnostic testing but also on benchmark testing, project-based learning, capstone projects, and performance assessments, with a focus on remote instruction and trauma-based instruction.53
• School districts must implement strategies to retain highly credentialed teachers, especially those at high risk of not returning because of the coronavirus.
  • COVID-19 is expected to boost early retirements, especially among teachers who are closer to retirement and among those in the highest-risk groups, and voluntary attrition, especially among those teachers who faced major obstacles in their work during the first months of the pandemic. These risks could also affect other staff at schools (e.g., nurses, paraprofessionals, principals) and come at a time when more personnel are needed. Budget constraints could further deplete the teaching and education workforces.54

• Schools and teachers must anticipate the need for more personalized learning, especially in 2020–2021, with a particular emphasis on the students who experienced the most interrupted learning time and the greatest challenges during the coronavirus crisis.
  • Flexible approaches will be necessary: Children learn differently, and they underwent different challenges during the pandemic. Remote learning is less effective for children who are less prepared (i.e., without full access to computers and other equipment, without experience using devices for school work, with fewer supports, and with less likelihood of being engaged).
  • More intensive interventions and strategies will be needed for students identified as at heightened risk of dropping out altogether.
  • Providing more flexible and personalized interventions for students will require more, better, and targeted investments in professional development for teachers so that they are equipped to deliver personalized learning.
• Systems must be redesigned to support students’ and teachers’ social and emotional learning.
  • The coronavirus crisis created serious challenges to students’ well-being and development that require a response focusing on their social and emotional learning, health, and well-being.55
  • Through their positive relationships with students, and through more specialized knowledge about social and emotional learning (SEL), teachers can contribute to the social and emotional learning of students. Therefore, improving training and support for teachers, teachers’ aides, and other school staff members in SEL will be critical to helping students regain their footing after the coronavirus crisis.
  • Supporting students’ social and emotional development will also require increasing the number of school nurses (clinics), counselors, social workers, paraprofessionals, etc., with a focus on both students’ social and emotional learning and their mental and physical health. Other practices at school (curriculums, etc.) can be enhanced to support social and emotional learning.
• The education system must explore other strategies that will allow children to make up for their interrupted education, drawing on the literature review presented earlier in this report. For the 2020–2021 school year and summer:
  • Schools should consider increasing both the amount and quality of learning time through a number of options, including extended schedules (in particular for those students lagging behind), summer enrichment programs that support the whole child, and staffing strategies that reduce class sizes and staff schools with sufficient and highly credentialed educators,56 including teachers’ aides and tutors, whether in person or online.
  • Schools should also consider ensuring access to and quality of online instruction, if online education is going to be used on its own or in conjunction with traditional instruction. In keeping with the recommendations in the “relief” section above, online instruction needs to be better tailored (especially for those who are least prepared), of high-quality, and accessible to all students. Similarly, schools need to provide supports for teachers who had not been prepared on how to use technology for instruction. Teachers should be enlisted in helping to create online instructional tools and policies.57 Finally, districts and teachers must apply “an equity lens,” to target tools and resources to students who experience the biggest opportunity gaps (i.e., students who lack digital access or who suffer more from nutrition challenges or housing instability).

Rebuilding: Redesign the system to focus on nurturing the whole child and on equal provision of opportunities

Major crises provide unique opportunities to rethink the status quo. In the aftermath of the coronavirus crisis, policymakers must seize the opportunity to address structural problems in the educational system and invest new and different approaches. This should be a pathway toward establishing a system that ensures we meet the student, teacher, and school needs that we have been neglecting and make delivering excellence and equity in education the norm. Delivering equity in education requires addressing the major disparities in student outcomes by race and social class that arise in a system designed to deliver disparities in educational opportunities. The bottom line is, we must seize this moment to redesign the system to deliver the excellence and equity needed for every child to be able to thrive.58

• Education policy and systems must embrace a whole-child approach to education. The pandemic has crystallized the lack of sufficient balance in the types of instruction and supports that schools prioritize.
  • Going forward, the education system must better balance what we teach, how we teach it, and how we reward the full range of skills that matter for and define a child’s development and education. The institutions that create education policy and practice must make many changes to ensure that schools teach and reward the development of cognitive and socioemotional skills. The shift begins with recognizing that skills of both types are mutually supportive, not mutually exclusive.59
  • A whole-child approach to education would include policies and practices that also close both opportunity and enrichment gaps.
    • For example, a whole-child approach that embraces and employs a broader range of assessments, and uses these assessments for “formative and informative” purposes, rather than for judging and sorting students, would also go a long way to closing the gaps. This shift recognizes that traditional tests are designed to capture only a narrow slice of what children know and can do, and that these tests are biased toward the types of skills that are closely correlated with parents’ socioeconomic status, not necessarily, and not exclusively, children’s potential.
• Education policy must also acknowledge and address the impacts of poverty and of racial and economic segregation on students’ capacity to learn and on teachers’ abilities to do their jobs.
  • School districts must conduct a detailed needs assessment of the district overall and of each school in the district, identifying where poverty and all other stressors that are intertwined with poverty impact the ability of children to learn, and mapping out community resources that can be leveraged to meet those needs. And it means working through a variety of channels (and with a variety of partners) to close the opportunity and enrichment gaps that have long impeded progress for low-income students, students of color, and students from immigrant families and communities.60
  • Education systems must tackle head-on the school- and district-based disparities that mirror and compound the disparities that children experience at home. In high-poverty schools, and in schools serving larger shares of minority students, there is generally less access to the education “inputs” that lead to good outcomes, whether it is highly credentialed teachers, access to after-school programs, access to AP classes, positive ways of dealing with discipline issues, etc. A broad range of tools and resources must be deployed to close gaps by types of school on all fronts, making education funding more adequate and more equitable.
  • School systems and their community partners must also establish a flexible set of strategies to offer wraparound supports—such as health clinics, community gardens, and parenting classes—tailored to the specific features of the community and the diversity of the communities serving our 55 million students across the country.
• Administrators and institutions must treat teachers as professionals whose knowledge and experiences are a valuable resource for improving education.
  • All the institutions in the education system and society at large must value education and educators and treat teachers as professionals. Teachers’ judgement is critical to identifying what children and educators need. School districts and education institutions must improve the types and usefulness of the professional development and supports offered to teachers, to allow them to keep up with advances in research on effective teaching and face the challenges of the job. Teachers must also be given more of a say in the decisions affecting their jobs and careers, from the materials they use in their classrooms to the types of training they receive. Valuing educators also includes paying them at a level commensurate with what similar college-educated workers earn in other professions. Research shows that taking these steps can help attract professionals to teaching as a career and help prevent them from retiring or quitting their schools and the profession.61
• Policymakers must recognize that education policy alone cannot ensure that all children have the foundation they need to get a good education. We need an economic agenda to accompany the rebuilding that lifts all children up and closes the opportunity gaps that are educational and not educational in nature. Children in low-income families—often children of color—lack many of the resources that their higher income and white peers have, which puts them at a disadvantage before they even enter their classrooms. Some opportunity gaps can be addressed by strengthened education policies. But the ones of a different nature would call for better public policies and a stronger economic agenda.62

• Finally, policymakers at all levels must establish and fund contingency plans for the next time we experience a crisis as disruptive and overwhelming as the coronavirus pandemic, whether that occurs in the next handful of years or further into the future.

Conclusion

Despite the fact that we do not know exactly how the COVID-19 pandemic is affecting children’s needs and academic performance, we know enough from existing research on learning during somewhat comparable educational experiences, and from news and observations of how education is being produced during the crisis, to assess the likely consequences on educational outcomes both overall and for relatively disadvantaged subgroups.

We reviewed the research on what to expect when children experience a substantial loss of learning time, when schools make a sudden shift to remote learning and home schooling without meeting the conditions for their effectiveness, and when circumstances lead to a massive increase in stress and disruption for children and their families. We also reviewed evidence that has emerged during the crisis on the multiple challenges that children, their teachers, schools, families, and communities face, all of which exacerbate opportunity gaps. Indeed, the evidence points to disparities in opportunities that exacerbate existing inequities and place major stress on low-income students and their teachers, in particular. Due to the digital divide and many other factors, these children are most likely to lose more substantial learning time. And their families are also most likely to experience compounded stresses—such as job loss, the loss of health care, the lack of paid sick leave, the lack of child care, and the need to work on site in “essential” jobs that put them at health risks: all these factors make it much harder for these families to attend to children who are suddenly home schooling and struggling with ad-hoc efforts at remote learning.

Together, the lessons learned point to the need to enact an agenda that lifts up children and reduces educational inequities after the interruption to schooling due to the coronavirus is over. The agenda must also rebuild the system so that lifting up children and reducing inequities in education become the new norm. To accomplish this, we outline a three-stage response. The first stage is immediate relief for students and educators so they can function better in the early 2020–2021 school year as remote learning continues in some form for many children. The second stage is significant short-term investments during the recovery that will enable students whose education was interrupted by the coronavirus crisis to catch up and continue their development. The third stage is longer-term reforms to rebuild the education system so that the challenges documented here are corrected and the system finally delivers an excellent, equitable education to all children.

In the rebuilding phase, it is essential to establish an education system that embraces a whole-child approach, addresses the impacts of poverty and inequality on students’ capacity to learn and on teachers’ abilities to do their jobs, offers a flexible set of wraparound supports to mitigate the impacts of the inequities that are built into the system, values education and educators, and creates viable contingency plans for future crises.

In closing, the ultimate consequences of the pandemic for K–12 education in the United States will indeed be a function of the quality, intensity, and comprehensiveness of our response to counter the pandemic’s negative lasting effects. Indeed, our call for relief, recovery, and reform has a historical precedent. As Darren Walker, president of the Ford Foundation, recently noted:

During the Great Depression, President Franklin Delano Roosevelt affirmed the need for relief, recovery, and reform—in that order. Today, we must follow these same steps—beyond reform to a broader, deeper reimagination of our society. (Darren Walker 2020).

This societal reimagination certainly encompasses a reimagination of our education system. With the right vision, we can actually ensure that public education plays a critical role in restoring the human and social capital in our country and in readying us for the next challenges, big or small, that we may confront in the future. Our children and our future depend on it.

About the authors

Emma García is an education economist at the Economic Policy Institute, where she specializes in the economics of education and education policy. Her research focuses on the production of education (cognitive and noncognitive skills), evaluation of educational interventions (early childhood, K–12, and higher education), equity, returns to education, teacher labor markets, and cost analysis in education. She has held research positions at the Center for Benefit-Cost Studies of Education, the Campaign for Educational Equity, the National Center for the Study of Privatization in Education, and the Community College Research Center; consulted for MDRC, the World Bank, the Inter-American Development Bank, and the National Institute for Early Education Research; and served as an adjunct faculty member at the McCourt School of Public Policy, Georgetown University. She received her Ph.D. in economics and education from Columbia University’s Teachers College.

Elaine Weiss is the lead policy analyst for income security at the National Academy of Social Insurance, where she spearheads projects on Social Security, unemployment insurance, and workers’ compensation. Prior to her work at the academy, Weiss was the national coordinator for the Broader Bolder Approach (BBA) to Education, a campaign launched by the Economic Policy Institute, from 2011 to 2017. BBA promoted a comprehensive, evidence-based set of policies to allow all children to thrive in school and life. Weiss has co-authored and authored EPI and BBA reports on early achievement gaps and the flaws in market-oriented education reforms. She is co-author of Broader, Bolder, Better, a book with former Massachusetts Secretary of Education Paul Reville published by Harvard Education Press . Weiss came to BBA from the Pew Charitable Trusts, where she served as project manager for Pew’s Partnership for America’s Economic Success campaign. She has a Ph.D. in public policy from the George Washington University Trachtenberg School and a J.D. from Harvard Law School.

Acknowledgments

The authors are grateful to EPI Publications Director Lora Engdahl for having edited this report, as well as for co-authoring one of the pieces this report builds on, and for her suggestions on news reports that provide useful context. To the last point, we also acknowledge the extensive work on the repercussions of COVID-19 for education conducted by many of our colleagues, of which we are only able to cite a fraction. We appreciate EPI Vice President John Schmitt’s supervision and support of this project, EPI Research Assistant Melat Kassa for her assistance with the tables and figures, and EPI’s communications staff for their assistance with the production and dissemination of this study.

Endnotes

1. For references on production of education, see Coleman et al. 1966; Hanushek 1979; Todd and Wolpin 2003.

2. Note, too, that we do not offer an in-depth review of these very extensive bodies of work, but rather use them to better understand what it is at play and to frame what we should anticipate the next-phase and post-pandemic outcomes to look like.

3. Students in grades kindergarten and first, for example, experienced larger gains as measured by the ECLS-K assessments in math and reading between the fall and the spring of those years. For example, our descriptive analysis of the ECLS-K 2010–2011 data suggests that students gain an average of 0.7 SD in kindergarten. For a discussion on spring to spring gains by grades (average of 0.45 SD across grades), see Bloom et al. 2008.

4. These kinds of challenges and trade-offs may also be relevant to the decisions schools will need to make for 2020–2021. For example, von Hippel (2020), when discussing school instruction that spans 12 months, explains that although year-round calendars increase summer learning, in most cases they reduce learning at other times of year, so that the total amount learned over a 12-month period is no greater under a year-round calendar than under a nine-month calendar.

5. Assessing a seminal study by Alexander, Entwisle, and Olson (2007), based on a sample of Baltimore students who were tracked from first grade in 1982 to age 22, Kuhfel explains that most of the test-score gap by socioeconomic status (SES) in ninth grade was explained by “differing summer experiences in the early elementary years.”

6. The more recent research also discusses several technical challenges that would require some concern about the findings. For example, there were characteristics of the tests used to assess skills before and after the summer that made them not comparable, or that made the tests more difficult in the fall than in the spring; very small samples in particular contexts; and other caveats. See von Hippel and Hamrock (2019) and von Hippel, Workman, and Downey (2018).

7. Atteberry and McEachin (2020) find that slightly over half of the students lose nearly all their school-year progress but the rest of the students actually maintain their school-year learning. Kuhfeld (2019) similarly finds that the summer loss is not generalized, but points to a larger loss overall, with around 60–80% of students losing ground in the elementary school grades (and an even larger share with respect to math). Kuhfeld (2019) also finds that the slide is larger in higher grades than in lower grades, and that performance gaps between minority and nonminority students did not increase, but gaps between students in high-poverty versus low-poverty schools increased significantly but by a small amount (at most, students in high-poverty schools lost one week of learning). The two studies (Atteberry and McEachin 2020 and Kuhfeld 2019) use the NWEA’s MAP Growth reading and math assessments. von Hippel, Workman, and Downey (2018) estimate that during the summer, performance gaps by socioeconomic status slightly increase for children in their first years in school. Our own exploratory analysis of the ECLS-K 2010–2011 data coincides with finding most students experience gains during the summers (both in math and reading), and that the performance gaps widen between low- and high-income children (using household income as a proxy for socioeconomic status). See also Quinn et al. 2016.

8. Definitions of chronic absenteeism vary by study, school district, etc. They typically are based on the number of days or a share of days missed over an entire school year, and they are only available on a yearly basis. For example, the U.S. Department of Education (2016) defines chronically absent students as those who “miss at least 15 days of school in a year.” Elsewhere, chronic absenteeism is frequently defined as missing 10% or more of the total number of days the student is enrolled in school or missing a month or more of school in the previous year (Ehrlich et al. 2013; Balfanz and Byrnes 2012).

9. Some examples are J-PAL 2017, Jordan 2019, and Balu 2019.

10. This 1.5 million figure is of course not completely illustrative today because overall enrollment numbers are expected to have grown since 2010. As a related reference, the National Center for Education Statistics estimates that there were 656 virtual schools in the U.S. in 2017–2018, enrolling about 279,000 students (0.55 percent of total enrollment) (NCES 2019b).

11. The literature on use of devices for education covers a lot of ground: findings tend to be a function of the type of technology/device used, the intensity, the developmental period/age, etc. (Crone and Konijn 2018; Walsh et al. 2018, see a summary in García 2018). To illustrate a few of these associations, researchers have found that time spent using a mobile phone and watching TV and sending text messages is correlated with lower achievement, slower reading times, and more intuitive but less analytic thinking, and it is also correlated with a faster but less accurate performance in a test of selective attention capacity and skills, as well as in processing-speed ability (Evans-Schmidt and Vandewater 2008; Lepp, Barkley, and Karpinski 2014; Fox, Rosen, and Crawford 2009; Barr et al. 2015; Abramson et al. 2009). Video-gaming can positively influence visual attention and spatial skills (attention capacity, quicker attention deployment, and faster processing,  according to Evans-Schmidt and Vandewater 2008). More frequent use of social media is negatively correlated with grade point averages (GPA), academic performance, and hours per week spent studying (Junco 2012; Karpinski et al. 2012; Kirschner and Karpinski 2010). Texting, using Facebook (and accessing Facebook while studying), and conducting internet searches unrelated to academic activity concurrent with homework completion all negatively correlate with GPA (Junco and Cotten 2012; Rosen, Carrier, and Cheever 2013; Wilmer, Sherman and Chein 2017). Media use (including social media) positively correlates with social and emotional learning (SEL) development, relationships with peers, and engagement, but also with addiction, bullying, mood and self-esteem problems, and time not sleeping/exercising/studying, some due to the trade-offs between time spent on some of these activities (Crone and Konijn 2018; Lemola et al. 2015; American Academy of Pediatrics 2011). The evidence also points out that if the content watched is high-quality educational programming, and does not displace other cognitively enriching experiences, screen time is positively correlated with achievement, engagement, and attitudes toward learning (Evans-Schmidt and Vandewater 2008). Concerns with excessive screen time have been well covered in the media during the months of the pandemic. See for example Kamenetz 2020a; Cheng and Wilkinson 2020.

12. Some information for households with children during the pandemic has been released by the U.S. Census Bureau through the Household Pulse Survey Tables for a target population of adults 18 years and older. See U.S. Census Bureau 2020a.

13. They say: “These students’ learning and persistence outcomes are worse when they take online courses than they would have been had these same students taken in-person courses.” See Zhao 2020 for some discussion of the challenges around online learning. NCES has used this period to build a repository of this research, which is discussed in Soldner 2020.

14. One in three online charter schools reported that all of their courses were self-paced. On average, online charter schools provide less simultaneous learning and teaching in a week than conventional schools would have in a day and less one-on-one instruction, with larger student-to-teacher ratios. Principals in these schools reported that the greatest challenge was student engagement (a challenge cited almost three times as often as any other issue) (Gill et al. 2015). Based on national data, across all tested students in online charters, the typical annual academic losses are -0.25 SD for math and -0.10 SD for reading (Woodworth et al. 2015). See Bueno 2020 for a more updated study of full-time virtual school attendance in Georgia, which shows negative effects ranging from -0.1 to -0.4 SD on performance.

15. This share has been relatively stable since 2007.

16. Subjects tested include reading, language l, mathematics (with computation), science, social studies, core (with computation), and composite (with computation).

17. As researchers note, the evidence is limited by the inability to use experimental or even quasi-experimental methods, precluding them from drawing conclusions as to causality (Belfield 2004; Cheng and Donnelly 2019; Lubienski, Pukett, and Brewer 2013). Belfield (2004) explains the three empirical issues that arise when comparing outcomes from home schooling against public schooling: 1) the common concern over the endogeneity of school choice, that is different types of families choose the type of school that their children attend, and little can be inferred about the impacts of schools for students who do not attend them; 2) the need to distinguish the absolute performance of home-schoolers from the treatment effect of home schooling—“Given the above-median resources of many home-schooling families, academic performance should be high even if home schooling itself is not differentially effective. Full controls for family background are needed, however, to identify a treatment effect”; 3) “home-schoolers can often choose which tests to take and when to take them (and have parents administer them), introducing other biases.”

18. Bacher-Hicks, Goodman, and Mulhern (2020) examine the search for online learning platforms used by schools and supplemental resources on Google. They find that the search intensity had roughly doubled relative to baseline. (They also find that the intensity rose twice as much in areas with above-median SES as in areas with below-median SES, where SES is measures by household income, parental education, and computer and internet access.

19. This lack of time for planning has in a way continued during the summer. As the news reports have broadly shown, many schools were going to reopen but they had to cancel at the last minute, which probably meant that the plans in place were no longer aligned with students’ and teachers’ needs. In other cases, the uncertainty about resources available (as discussed later in the report) led to a squandered opportunity to plan accordingly.

20. The authors point to the nine factors that determine the quality of online teaching and learning, including modality, pacing, student-instructor ratio, pedagogy (type), role of online assessments, students’ online roles, instructors’ online roles, online communication synchrony, and source of feedback. While all may not apply as strongly in K–12 education, the range of considerations highlights the challenges public school teachers will face in attempting to make remote instruction effective.

21. More broadly, these aspects about online instruction also touch upon the relevance of teacher professional development, the importance of establishing learning communities for teachers, and teachers’ access to a sound system of supports (Darling-Hammond et al. 2017; Kraft, Blazar, and Hogan 2018; García and Weiss 2019). Among other advantages, learning communities allow teachers to acquire new skills, update their knowledge, and strengthen their practice and effectiveness in the classroom, all critically important factors for education quality and also for the stability of the teaching workforce (García and Weiss 2019).

22. As we explained in our study, the professional development module that delivered data for the 2011–2012 SASS is rotating and was not included in the most recent data set available when we were conducting our study (2015–2016), but it will be in the next cycle, 2017–2018.

23. Teachers also reported having very little input on which activities to undertake for their professional development. Only 11.1 percent of teachers have a great deal of influence determining the content of in-service professional development programs. As we noted in García and Weiss 2019, this disregard for teachers’ input is quite troubling, given national and international surveys and testimonies showing that teachers want to play a more direct role in selecting the types and content of professional development opportunities offered to them (see Bill & Melinda Gates Foundation 2014; Loewus 2019; OECD 2019; Kirk 2019; Schwartz 2019).

24. For example, in Washington, D.C., the school district has indicated attendance is compulsory for students ages 5–17. Schools will use daily attendance as an indicator of student engagement in learning together with information on completing assignments and participation in live classes (District of Columbia Office of the Mayor 2020).

25. This sharply academic focus narrowed with the 2001 passage of the federal No Child Left Behind Act, which replaced the earlier version of the Elementary and Secondary Education Act (ESEA). The 2015 passage of the Every Student Succeeds Act attempted to dial back that pressure (see CASEL 2020; Kostyo, Cardichon, and Darling-Hammond 2018). Useful references on these issues and some others discussed below are Bloom 1964; Borghans et al. 2008; Duckworth and Yeager 2015; Levin 2012; Jones et al. 2016; Jones et al. 2019; Shonkoff and Phillips 2000; Lippman et al. 2015; Petway, Brenneman, and Kyllonen 2016; UNESCO’s Incheon Declaration for Education 2030 (UNESCO et al. 2016); and our own work on these issues: García 2014; García and Weiss 2016.

26. For those interested in this approach, Tirivayi et al. (2020) offer a comprehensive examination of past public policy responses to emergency crises.

27. Technically, this is known as lack of external validity. This research documents that approximately 50 million primary- and lower-secondary-age children are out of school in conflict-affected countries around the world (Save the Children 2013). Natural disasters, which also displace large numbers of students, are four times as prevalent today as they were in the 1980s, likely due to the growing impacts of climate change, and that number is predicted to increase exponentially in the next 20 years (Oxfam International 2007; Save the Children 2008; USAID 2014).

28. Further, research has explored the effects on the communities to which children and their families migrate (known as spillover effects from emergency migrants on the host communities), as well as some of the factors that explain them. Hurricane Maria in September 2017 caused a large influx of students from Puerto Rico to Florida’s public schools—about 12,000 students between October 2017 and May 2018. Studies found immediate negative effects on the performance outcomes of host students (students in the schools accepting new students from the disaster area) following hurricane Maria. Studies also found immediate negative effects on the performance outcomes of host students following Hurricane Katrina in September 2005, though they found zero effects on Florida’s public schools following the Haitian migrant influx after the earthquake in January 2010 and two years after it (Özek 2020; Imberman, Kugler, and Sacerdote 2012; Figlio and Özek 2019). Özek (2020) found significant adverse effects of hurricane migrants on the educational outcomes of existing students in the first year. Specifically, he found that a 5-percentage-point increase in the share of hurricane migrants reduced test scores in math and in English language arts (ELA) by an amount equivalent to one to two months of instruction, increased the likelihood of being involved in a disciplinary incident by 15–20% (of the dependent variable mean) in middle and high school, and increased the likelihood of existing students leaving their schools before the start of the 2018–2019 school year by roughly 7% (with larger increases among white and African American students). Effects were mainly concentrated among higher-performing students, especially in disadvantaged school settings.

29. Historically, there is strong agreement that in these circumstances, having access to education (versus not having access) leads “to a range of positive outcomes including child protection and well-being, economic development, peace building, and reconstruction” (Burde et al. 2017).

30. Other contingency planning strategies involve providing psychosocial programs or supplemental educational activities that protect children from harm. The strategies avoid unstructured days where traumatizing memories linger, fears thrive, and violence is always possible (Sommers 1999). Some education content, for example in refugee contexts, may be designed to mitigate conflict, and peace education programs show promise in changing attitudes and behaviors toward members of those perceived as the “other” (Burde et al. 2017). As Anderson (2020) indicates, “it is not only the mechanism and approach that is used but also the quality and methods of teaching that are critical to understand.” Different mechanisms for delivering education include radio, podcast, or television broadcasts; online programs or virtual peer learning circles; and even the provision of kits with basic materials (pencils, exercise books, erasers, etc.). Another critical element is to ensure that children have access to the instructional mechanisms used.

31. A recent publication by The Century Foundation notes “the significant variation in both per-pupil spending and student outcomes across the country” and estimates that the U.S. needs to spend an additional $150 billion to ensure that all students “achieve national average outcomes” (TCF 2020). For research about the important role that opportunity gaps and family income play in education performance, see Coleman et al. 1966; Reardon 2011; García and Weiss 2017; Putnam 2015; Rothstein 2004; and Weiss and Reville 2019.

32. Food insecurity is a different measure than poverty. The former, in the Bauer article, refers to the share of households reporting to the U.S. Census Bureau that it was sometimes or often the case that the children in the household “were not eating enough because we just couldn’t afford enough food.” But poverty rates are also an instructive measure during this crisis. Using an unlikely scenario of an unemployment rate of 30% this year due to COVD-19, Parolin and Wimer (2020) estimate that poverty rates in the United States could reach their highest levels in 50 years. Specifically, they estimate that if unemployment rates stay at 30% throughout the year, the supplemental poverty measure (SPM) rate for children would rise by more than 7 percentage points, from 13.6% to 20.9% (the SPM created by the U.S. Census Bureau is a measure of poverty that some researchers consider more accurate than the official poverty measure because it takes into account income from such benefits as food stamps and housing assistance).

33. A total of 1.5 million students surveyed in the 2017–2018 school year had experienced homelessness at some point during the last three school years (USICH 2020).

34. Even if they don’t lose their jobs, some workers and virtually all essential workers don’t have access to work remotely (following the traditional racial/SES inequities). The inability to work remotely means that keeping their jobs and thus their access to health insurance disproportionately exposes them to the virus (Gould and Shierholz 2020; Bivens and Zipperer 2020) and makes it nearly impossible for them to supervise their children and assist them in their education needs.

35. For updated information, nationally and for various subgroups, see the CDC COVID Data Tracker (CDC 2020c).

36. This is a problem both for students in dense urban areas, where normally strong hospital systems have been overwhelmed at times during the pandemic, and in rural areas, where already gutted systems have lacked the capacity to deal with the onslaught of cases. See, for example, the description of New York City’s hospitals when that city was hit hard early in the pandemic in Arnold 2020 as well as Sandoval 2020’s more recent account of a small rural hospital on the Texas–Mexico border.

37. Specifically, in our studies, poor students are students eligible for the federal free or reduced-price lunch programs under federal guidelines that deliver such meals based on family income falling below a certain threshold. Non-poor students are students who are ineligible for those programs. For a recent discussion, see Cookson 2020.

38. While 25% of superintendents reported that almost all of their students (91–100%) had internet access at home and 26% reported that almost all of their students had devices to connect to the internet at home, substantial shares of superintendents reported gaps in that access: 23% estimated that just 81–90% had access to internet and devices; 16–17% estimated that 71–80% had access to internet and devices; 11% estimated that just 61–70% had access to internet and devices; 10% said the share with access to internet and devices was 50% or less; and 14% said the share with access to internet and devices was 50% or less (Rogers and Ellerson Ng 2020).

39. As early as March, Texas waived requirements that students take its standardized state STAAR test due to the closure of schools (Swaby 2020), and Massachusetts did the same in April (Lisinski 2020). See also Brookings Institution 2020; Darling-Hammond and Kini 2020; NEPC 2020; Ravitch 2020.

40. AFT 2020d. Capstone projects are end-of-year term projects that students can complete to bring the school year to a close in lieu of statewide standardized assessments (see Weingarten 2020). For some examples of these projects, see Dickinson 2020.

41. U.S. Census Bureau (2020b). McNichol and Leachman (2020) estimate “$555 billion in shortfalls over state fiscal years 2020–2022.” Bivens (2020) reviews estimates of a revenue shortfall for state and local governments of nearly $1 trillion.

42. See Baker and DiCarlo 2020; Leachman and Figueroa 2019; Partelow, Yin, and Sargrad 2020.

43. Since March 2020, the House of Representatives and the Senate have passed four coronavirus relief packages totaling over $3 trillion. The most current proposed measures are the HEROES and HEALS Acts (Lee 2020a, b; Progressive Caucus Action Fund 2020). For a discussion on the relatively small amounts that public schools and education have received, see Jordan 2020b; Reber and Gordon 2020. See also Snell 2020.

44. An obvious lesson learned from the COVID-19 crisis is that schools and related sectors like early childhood education and child care are undervalued relative to their key contributions to the societal good. Schools are “essential to the operation of the country… It is impossible to restart the economy without the schools, they go together” and are “a critical part of the social safety net for children” (ASI 2020). Education and also health and social services are “forms of investment, not consumption; necessities, not luxuries” (Folbre 2016). Just as we have learned that many formerly invisible workers are “essential” to the daily functioning of our economy, we must treat education as the essential service it is and support it as such.

45. Blad 2020; Broadwater 2020; Calargo 2020; Ferris 2020; Ferguson 2020; Strauss 2020; Valant 2020.

46. See Tinubu Ali and Herrera 2020; Cohodes 2020.

47. One potential silver lining of the coronavirus pandemic is that it brings attention to a longstanding issue in education: the inadequate systems of professional development for teachers (see García and Weiss 2019). As practitioners, researchers, and policymakers collaborate more closely on professional development offerings that will help teachers teach during the pandemic, that model can inform a broader look at the systems of professional supports available to teachers and prompt more research on what constitutes optimal professional development—i.e., what professional development offerings need to cover, how the offerings should be delivered and where and for how long, and how teachers are connected to the opportunities. As we showed in García and Weiss 2019, teachers want these supports but too often are offered one-size-fits-all programs when there is no single optimal combination valid for all teachers at all times and in all settings. Also shown in García and Weiss 2019, enhanced professional development would play a role in keeping teachers in the classroom and attracting new professionals into teaching.

48. See for example U.S. Senate 2020 for an overview of the proposed Coronavirus Child Care and Education Relief Act.

49. See CDC 2020a, 2020b; AASA 2020; UNESCO et al. 2020; NEA 2020; AFT 2020a; National Superintendents Roundtable 2020. There are still many things that scientists and public health experts do not know about the prevalence, transmission, and long-term consequences of contracting COVID-19 among children and adolescents. Likewise, there is no universally agreed on threshold of incidence of the disease under which activities can safely resume. While these questions are beyond the scope of this report and our areas of expertise, they are critical factors weighing on the reopening of our schools. Several studies point to lower prevalence of infection among children than on average but also to the need to assess whether the incidence of the disease among children can be influenced by selective testing, how prevalence of the virus among children compares with prevalence among their parents (i.e., whether the rate of infection of parents is different from their children’s), how these have changed over time (i.e., whether the immunity lasts longer for children or for parents, etc.), etc. (Idele et al. 2020; Pollán et al. 2020; Heald-Sargent et al. 2020). The American Academy of Pediatrics (2020) is requesting that schools reopen. See Goldstein 2020b.

50. While there is no precise estimate of how much following these guidelines would cost, the School Superintendents Association estimates that the average school district will need an additional $1.78 million to meet the COVID-19-related expenses of reopening schools (AASA 2020). The National Academy of Sciences estimates the cost of health-related supplies at $1.8 million for a school district serving 3,200 students (National Academies of Sciences, Engineering, and Medicine 2020). The Council of Chief State School Officers explains that the costs associated with opening schools safely under appropriate health and safety protocols would add up to about $30 billion across all schools (CCSSO 2020). The American Federation of Teachers culls from a number of sources to estimate that a total of $116.5 billion is needed for all measures, $35 billion of which would be needed for additional instructional staff to support adequate social distancing (AFT 2020b, 2020c). See also DiNapoli Jr. 2020 and Berman 2020. The cost of reopening schools is an unsettled issue.

51. See ASI 2020; CPCC, The Education Trust, NEA 2020; Duflo 2020; Brookings Institution 2020.

52. See Gordon 2013; RESEARCHED 2020.

53. AFT 2020d; Weingarten 2020; Dickinson 2020.

54. See García and Weiss 2020; Will 2020; Page 2020; Hamilton, Kaufman, and Diliberti 2020; NIRS 2020. For early retirements of teachers and principals, see Will 2020 and Page 2020. For challenges imposed by remote instruction, see Greif Green and Bettini 2020; Prothero 2020. In terms of recessions, public education job losses following the Great Recession exceeded 316,000 between September 2008 and September 2011 (BLS 2020). The job losses in April 2020 alone were already greater than in all of the Great Recession: 468,800 jobs were lost just a month after the pandemic started (Gould 2020b; see BLS 2020 for a still deeper decrease in May and a slight recovery in June and July). An estimate of the consequences of a 15% reduction in state education funding says that it could lead to the loss of more than 300,000 teaching positions (or 8.4%; see Griffith 2020).

55. See Darling-Hammond et al. 2020; Shonkoff and Phillips 2000; García and Weiss 2016; Walker, Tim 2020; Weiss and Reville 2019; Zhao 2020; Clark et al. 2020; Goldstein 2020a.

56. See Mishel and Rothstein 2003 and Schanzenbach 2020 for a recent review of the influence of class size on achievement. Note that this literature was not reviewed in the literature review section of this report because class size has generally not been a feature of the pandemic. However, in the literature, smaller classes are an implicit recommendation from various subfields. For evidence on summer programs, see McCombs et al. 2019. For evidence on tutoring effectiveness, see Nickow, Oreopoulos, and Quan 2020. On personalized learning, see Kim 2019.

57. Ferguson et al. 2020; García 2020; Hamilton, Kaufman, and Diliberti 2020.

58. Oakes, Maier, and Daniel 2017; Gonzalez 2018; Weiss and Reville 2019; Darling-Hammond et al. 2020; Starr 2020.

59. Darling-Hammond et al. (2020) discuss this framework as informed by evidence from the science of learning and development. See the different principles of practice in their Figure 1.

60. Weiss and Reville 2019; Shonkoff and Williams 2020.

61. EPI’s series of reports on the teacher shortage documents the factors that lead teachers to quit (and likely discourage people from entering the profession). See Economic Policy Institute 2020. See Allegretto and Mishel 2019 for estimates of the teacher pay penalty (how much less teachers earn in wages and benefits than comparable college-educated workers in other professions).

62. See García 2015 and García and Weiss 2017, among others.

References

AASA. 2020. Guidelines for Reopening Schools from the AASA COVID-19 Recovery Task Force.

Abramson, M.J., G.P. Benke, C. Dimitriadis, I.O. Inyang, M.R. Sim, R.S. Wolfe et al. 2009. “Mobile Telephone Use Is Associated with Changes in Cognitive Function in Young Adolescents.” Bioelectromagnetics 30: 678–686. https://doi.org/10.1002/bem.20534.

Albert Shanker Institute (ASI). 2020. Panel on Coronavirus Pandemic and K-12 Education Funding (Expert Presenters: Pedro Noguera, Sylvia Allegretto, Bruce Baker; Randi Weingarten) (webinar). April 24, 2020.

Alexander, Karl L., Doris R. Entwisle, and Linda Steffel Olson. 2007. “Lasting Consequences of the Summer Learning Gap.” American Sociological Review 72, no. 2: 167–180. https://doi.org/10.1177/000312240707200202.

Allegretto, Sylvia, and Lawrence Mishel. 2019. The Teacher Weekly Wage Penalty Hit 21.4 Percent in 2018, a Record High: Trends in the Teacher Wage and Compensation Penalties Through 2018. Economic Policy Institute and the Center on Wage and Employment Dynamics at the University of California, Berkeley, April 2019.

Allensworth, Elaine, and Shayne Evans. 2016. “Tackling Absenteeism in Chicago.”  Phi Delta Kappan 98, no. 2: 16–21. https://doi.org/10.1177/0031721716671900.

American Academy of Pediatrics. 2011. Children, Adolescents, Obesity, and the Media Pediatrics. Policy Statement, 128, no. 1, American Academy of Pediatrics, July 2011. https://doi.org/10.1542/peds.2011-1066.

American Academy of Pediatrics. 2020. COVID-19 Planning Considerations: Guidance for School Re-entry Critical Updates on COVID-19, Clinical Guidance. June 25, 2020.

American Federation of Teachers (AFT). 2020a. A Plan to Safely Reopen America’s Schools and Communities: Guidance for Imagining a New Normal for Public Education, Public Health and Our Economy in the Age of COVID-19.

American Federation of Teachers (AFT). 2020b. A Time to Act: The Importance of Investment in Public Education and Other State and Local Services in the Time of COVID-19.

American Federation of Teachers (AFT). 2020c. Reopening Schools During a Time of Triple Crisis: Financial Implications.

American Federation of Teachers (AFT). 2020d. Student Assessment and the Use of Data (guidance sheet).

American Federation of Teachers, NAACP Legal Defense and Educational Fund, and the Leadership Conference on Civil and Human Rights (AFT, LDF, and Leadership Conference). 2020. “Joint Statement of Education and Civil Rights Organizations Concerning Equitable Education During the COVID-19 Pandemic School Closures and Beyond.” April 29, 2020; updated with additional signers as of May 15, 2020.

Anderson, Allison. 2020. COVID-19 Outbreak Highlights Critical Gaps in School Emergency Preparedness, The Brookings Institution March 11, 2020.

Arnold, Carrie. 2020. “New York City’s Coronavirus Outbreak Is Already Overwhelming Hospitals.” New Scientist, March 31, 2020.

Atteberry, Allison, and Andrew McEachin. 2020. “School’s Out: The Role of Summers in Understanding Achievement Disparities.” American Educational Research Journal. https://doi.org/10.3102/0002831220937285.

Azzi-Hucktigran Kaliope, and Tigran Shmis. 2020. Managing the Impact of COVID-19 on Education Systems Around the World: How Countries Are Preparing, Coping, and Planning for Recovery, The World Bank, March 18, 2020.

Bacher-Hicks, Andrew Joshua Goodman, and Christine Mulhern. 2020. “Inequality in Household Adaptation to Schooling Shocks: Covid-Induced Online Learning Engagement in Real Time.” National Bureau of Economic Research Working Paper no. 27555.

Bakeman, Jessica. 2020. “Taking Attendance During COVID-19 School Closures; It’s Complicated.” April 8. WLRN.

Baker, Bruce, and Matthew DiCarlo. 2020. The Coronavirus Pandemic and K-12 Education Funding, Albert Shanker Institute, April 2020.

Balfanz, Robert. 2017. “Absenteeism Matters to Schools and Students.” Phi, Delta, Kappan. https://kappanonline.org/absenteeism-school-matters/.

Balfanz, Robert, and Vaughan Byrnes. 2012. The Importance of Being in School: A Report on Absenteeism in the Nation’s Public Schools. Center for the Social Organization of Schools (Johns Hopkins University), May 2012.

Balu, Rekha. 2019. “Intervention Design Choices and Evaluation Lessons from Multisite Field Trials on Reducing Absenteeism.” In Absent from School: Understanding and Addressing Student Absenteeism, edited by Michael A. Gottfried and Ethan L. Hutt, 199–212, Cambridge, Mass.: Harvard Education Publishing Group.

Barr, N., G. Pennycook, J.A. Stolz, and J.A. Fugelsang. 2015. “The Brain in your Pocket: Evidence that Smartphones Are Used to Supplant Thinking.” Computers in Human Behavior 48: 473–480. https://doi.org/10.1016/j.chb.2015.02.029.

Bauer, Lauren. 2020. “About 14 Million Children in the US Are Not Getting Enough to Eat.” The Brookings Institution, July 9, 2020.

Belfield, C. 2004. Home-Schooling in the US (Occasional Paper No. 88). New York: National Center for the Study of Privatization in Education, Teachers College, Columbia University.

Berman, Jillian. 2020. Here’s How Much School Closures Will Cost Parents in Lost Wages, Reduce GDP—and Negatively Impact the Nation’s Education System. Market Watch, July 24, 2020.

Bettinger, Eric, and Susanna Loeb. 2017. Promises and Pitfalls of Online Education. The Brookings Institution, June 9, 2017.

Bill & Melinda Gates Foundation. 2014. Teachers Know Best: Teachers’ Views on Professional Development. December 2014.

Bivens, Josh. 2020. As Economic Forecasts Worsen, Up to $1 Trillion in Federal Aid to State and Local Governments Could Be Needed by the End of 2021, Economic Policy Institute, May 11, 2020.

Bivens, Josh, and Ben Zipperer. 2020. Health Insurance and the COVID-19 Shock: What We Know So Far About Health Insurance Losses and What It Means for Policy. Economic Policy Institute, August 2020.

Blad, Evie. 2020. “Trump, Fauci, and Reopening Schools: What You Need to Know” (blog post). Education Week, May 14, 2020.

Bloom, Benjamin S. 1964. Stability and Change in Human Characteristics: New York: Wiley.

Bloom, Howard S., Carolyn J. Hill, Alison Rebeck Black, and Mark W. Lipsey. 2008. “Performance Trajectories and Performance Gaps as Achievement Effect-Size Benchmarks for Educational Interventions.” Journal of Research on Educational Effectiveness 1, no. 4: 289–328. https://doi.org/10.1080/19345740802400072.

Boca News Now. 2020. “Coronavirus: All FL Doctors Offices Must Close for Non-Emergency Care; Definition of “Procedure” Is Unclear, But Intent Is to Reduce COVID-19 Spread In Medical Offices.” Boca News Now, March 20.

Borghans, Lex, Angela L. Duckworth, James J. Heckman, and Bas ter Weel. 2008. “The Economics and Psychology of Personality Traits.” Journal of Human Resources 43, no. 4: 972–1059. https://doi.org/10.1353/jhr.2008.0017.

Broadwater, Luke. 2020. “Congress’s Ideological Divide Has Stymied Aid for Pandemic-Stricken Schools.” New York Times, August 6, 2020.

Brookings Institution. 2020. “Reopening Schools in the Fall Amid the COVID-19 Pandemic” (webinar). Expert panelists: John Allen, Michael Hansen, Daniel A. Domenech, Heather J. Hough, Emiliana Vegas. May 21, 2020.

Bueno, Carycruz. 2020. “Bricks and Mortar vs. Computers and Modems: The Impacts of Enrollment in K-12 Virtual Schools.” EdWorkingPapers no. 20–250. Annenberg Institute at Brown University. https://doi.org/10.26300/kahb-5v62.

Burde, Dana, Amy Kapit, Rachel L. Wahl, Ozen Guven, and Margot Igland Skarpeteig. 2017. “Education in Emergencies: A Review of Theory and Research.” Review of Educational Research 87, no. 3: 619–658. https://doi.org/10.3102/0034654316671594.

Bureau of Labor Statistics (BLS). 2020. “All Employees, Thousands, Local Government Education, Seasonally Adjusted” (online data table), Employment, Hours, and Earnings from the Current Employment Statistics Survey (National), accessed August 19, 2020.

Calargo, Jessica. 2020. “What Is Betsy DeVos Thinking?” New York Times, July 15, 2020.

Carnoy, Martin, and Emma García. 2017. Five Key Trends in U.S. Student Performance: Progress by Blacks and Hispanics, the Takeoff of Asians, the Stall of Non-English Speakers, the Persistence of Socioeconomic Gaps, and the Damaging Effect of Highly Segregated Schools. Economic Policy Institute, January 2017.

Centers for Disease Control (CDC). 2020a. “Preparing K–12 School Administrators for a Safe Return to School in Fall 2020” (web page), last updated August 26, 2020.

Centers for Disease Control (CDC). 2020b. “Operating Schools During COVID-19: CDC’s Considerations” (web page), last updated September 1, 2020.

Centers for Disease Control (CDC). 2020c. CDC COVID Data Tracker, last updated August 25, 2020.

Century Foundation (TCF). 2020. Closing America’s Education Funding Gaps. July 2020.

Cheng, Albert, and Michael Donnelly. 2019. “New Frontiers in Research and Practice on Homeschooling.” Peabody Journal of Education 94, no. 3: 259–262. https://doi.org/10.1080/0161956X.2019.1617576.

Cheng, Erika R., and Tracey A. Wilkinson. 2020. “Agonizing Over Screen Time? Follow the Three C’s.” New York Times, April 13, 2020.

Clark, Helen, et al. 2020. “A Future for the World’s Children? A WHO–UNICEF–Lancet Commission.” The Lancet 395: 605–658. https://doi.org/10.1016/S0140-6736(19)32540-1.

Cluver, Lucie, Jamie M. Lachman, Lorraine Sherr, Inge Wessels et al. 2020. “Parenting in a Time of COVID-19,” The Lancet, March 25, 2020. https://doi.org/10.1016/S0140-6736(20)30736-4.

Cohodes, Sarah. 2020. A Better Fall Is Possible, The Atlantic, July 7, 2020.

Coleman, J.S., E. Campbell, C. Hobson, J. McPartland, A. Mood, F. Weinfeld, and R. York. 1966. Equality of Educational Opportunity. Washington, D.C.: U.S. Office of Education, Department of Health, Education, and Welfare.

Collaborative for Academic, Social, and Emotional Learning (CASEL). 2020. “Federal Legislation to Promote Social and Emotional Learning.”

Congressional Progressive Caucus Center, The Education Trust, National Education Association (CPCC, The Education Trust, NEA). 2020. “Dispatch from the Frontlines—Standing Together to Demand Safe Schools” (webinar). Participants: Congresswoman Suzanne Bonamici, John King, Lily Eskelsen Garcia, Beatriz Beckford. August 5, 2020.

Conlin, Michelle, Lisa Baertlein, and Christopher Walljasper. 2020. “U.S. Food Banks Run Short on Staples as Hunger Soars.” Reuters Business News, April 24.

Cookson Jr., P.W. 2020. Measuring Student Socioeconomic Status: Toward a Comprehensive Approach. Learning Policy Institute, June 25, 2020.

Cooper, David, and Jaimie Worker. 2020. The Coronavirus Pandemic Requires State and Local Policymakers to Act, in Addition to Demanding a Strong Federal Response. Economic Policy Institute, March 17, 2020.

Cooper, Harris, Barbara Nye, Kelly Charlton, James Lindsay, Scott Greathouse. 1996. “The Effects of Summer Vacation on Achievement Test Scores: A Narrative and Meta-analytic Review.” Review of Educational Research 66, no. 3: 227–268. https://doi.org/10.3102/00346543066003227.

Council of Chief State School Officers (CCSSO). 2020. Letter to Honorable Lamar Alexander Chairman, Committee on Health, Education, Labor and Pensions United States Senate. June 24, 2020.

Crone, Eveline A. and Elly A. Konijn. 2018. “Media Use and Brain Development during Adolescence.” Nature Communications 9, no. 588: 1–10. https://doi.org/10.1038/s41467-018-03126-x.

Darling-Hammond, Linda, and Tara Kini. 2020. A New “New Deal” for Education: Top 10 Policy Moves for States in the COVID 2.0 Era. Learning Policy Institute, May 19, 2020.

Darling-Hammond, Linda, Dion Burns, Carol Campbell, A. Lin Goodwin, Karen Hammerness, Ee-Ling Low, Ann McIntyre, Mistilina Sato, and Ken Zeichner. 2017. Empowered Educators: How High-Performing Systems Shape Teaching Quality Around the World, 1st Edition. San Francisco: Jossey-Bass.

Darling-Hammond, Linda, Lisa Flook, Channa Cook-Harvey, Brigid Barron, and David Osher. 2020. “Implications for Educational Practice of the Science of Learning and Development.” Applied Developmental Science 24, no. 2: 97–140. https://doi.org/10.1080/10888691.2018.1537791.

Dickinson, Lisa. 2020. “Culminating Capstone Projects: Another Option for Engaging Students Remotely.” Share My Lesson Blog, May 12, 2020.

DiNapoli Jr., Michael A. 2020. Making School Budgets Whole and Equitable During and After COVID-19. Learning Policy Institute, July 17, 2020.

Dinarski, Susan. 2020. “The School Year Really Ended in March.” New York Times, May 7, 2020.

District of Columbia Office of the Mayor. 2020. DC’s COVID-19 Situational Update: August 26.

Dorn, Emma, Bryan Hancock, Jimmy Sarakatsannis, and Ellen Viruleg. 2020. COVID-19 and Student Learning in the United States: The Hurt Could Last a Lifetime. McKinsey & Company. June 1, 2020.

Duckworth, Angela L., and David S. Yeager. 2015. “Measurement Matters: Assessing Personal Qualities Other Than Cognitive Ability for Educational Purposes.” Educational Researcher 44, no. 4: 237–251. https://doi.org/10.3102/0013189X15584327.

Duflo, Esther. 2020. “Research and Policy Seminar with Esther Duflo: Good Economics for Covid-19 Times” (webinar). Inter-American Development Bank, June 26, 2020.

Economic Policy Institute. 2020. “Teacher Shortages” (Featured work web page). Accessed September 4, 2020.

Ehrlich, Stacy B., Julia A. Gwynne, Amber Stitziel Pareja, and Elaine M. Allensworth with Paul Moore, Sanja Jagesic, and Elizabeth Sorice. 2013. Preschool Attendance in Chicago Public Schools: Relationships with Learning Outcomes and Reasons for Absences. The University of Chicago Consortium on Chicago School Research, September 2013.

Evans-Schmidt, Marie, and Elizabeth Vandewater. 2008. “Media and Attention, Cognition, and School Achievement.” The Future of Children 18, no. 1: 63–85. https://doi.org/10.1353/foc.0.0004.

Ferguson, Katie, Amy Hysick, Rachel Murat, and Alhassan Susso. 2020. “The Reinvention Schools Really Need: Four New York Teachers of the Year Push Back at Gov. Cuomo.” New York Daily News, May 17, 2020.

Ferguson, Maria. 2020. “A Bad Time to Pick a Fight on Public Education.” KappanOnline, July 13, 2020.

Ferris, Sarah. 2020. “War Over Reopening Schools Hits Congress.” Politico, July 10, 2020.

Figlio, David, and Umut Özek. 2019. “Unwelcome Guests? The Effects of Refugees on the Educational Outcomes of Incumbent Students.” Journal of Labor Economics 37, no. 4: 1061–1096.

Figlio, David, Kristian L. Holden, and Umut Özek. 2018. “Do Students Benefit from Longer School Days? Regression Discontinuity Evidence from Florida’s Additional Hour of Literacy Instruction.” Economics of Education Review 67: 171–183. https://doi.org/10.1016/j.econedurev.2018.06.003.

Folbre, Nancy. 2016. “The Economics of Care“ (webinar), Institute for New Economic Thinking, February 24, 2016.

Fox, A.B., J. Rosen, M. Crawford. 2009. “Distractions, Distractions: Does Instant Messaging Affect College Students’ Performance on a Concurrent Reading Comprehension Task?” Cyberpsychol. Behav. 12: 51–53. https://doi.org/10.1089/cpb.2008.0107.

García, Emma. 2014. The Need to Address Noncognitive Skills in the Education Policy Agenda. Economic Policy Institute, December 2014.

García, Emma. 2015. Inequalities at the Starting Gate. Cognitive and Noncognitive Skills Gaps Between 2010–2011 Kindergarten Classmates. Economic Policy Institute, June 2015.

García, Emma. 2018. “Internet, Social Media, Games, Physical Activity and Educational Outcomes.” Unpublished manuscript.

García, Emma. 2020. “The Pandemic Sparked More Appreciation for Teachers, but Will It Give Them a Voice in Education and Their Working Conditions?” Working Economics Blog (Economic Policy Institute), May 7, 2020.

García, Emma, and Elaine Weiss. 2016. Making Whole-Child Education the Norm: How Research and Policy Initiatives Can Make Social and Emotional Skills a Focal Point of Children’s Education. Economic Policy Institute, August 2016.

García, Emma, and Elaine Weiss. 2017. Education Inequalities at the School Starting Gate: Gaps, Trends, and Strategies to Address Them. Economic Policy Institute, September 2017.

García, Emma, and Elaine Weiss. 2018. Student Absenteeism: Who Misses School and How Missing School Matters for Performance. Economic Policy Institute, September 2018.

García, Emma, and Elaine Weiss. 2019. The Role of Early Career Supports, Continuous Professional Development, and Learning Communities in the Teacher Shortage. The Fifth Report in ‘The Perfect Storm in the Teacher Labor Market’ Series. Economic Policy Institute, July 2019.

García, Emma, and Elaine Weiss. 2020. A Policy Agenda to Address the Teacher Shortage in U.S. Public Schools. The Sixth Report in ‘The Perfect Storm in the Teacher Labor Market’ Series. Economic Policy Institute, forthcoming 2020.

García, Emma, Elaine Weiss, and Lora Engdahl. 2020. “Access to Online Learning Amid Coronavirus Is Far from Universal, and Children Who Are Poor Suffer from a Digital Divide.” Working Economics Blog (Economic Policy Institute), April 17, 2020.

Ghandour, Reem M., Laura J. Sherman, Catherine J. Vladutiu, Mir M. Ali, Sean E. Lynch, Rebecca H. Bitsko, and Stephen J. Blumberg. 2018. “Prevalence and Treatment of Depression, Anxiety, and Conduct Problems in US Children.” Journal of Pediatrics 206: 256–267. https://doi.org/10.1016/j.jpeds.2018.09.021.

Gill, Brian, Lucas Walsh, Claire Smither Wulsin, Holly Matulewicz, Veronica Severn, Eric Grau, Amanda Lee, and Tess Kerwin. 2015. Inside Online Charter Schools. Mathematica Policy Research, October 2015.

Goldstein, Dana. 2020a. “Research Shows Students Falling Months Behind During Virus Disruptions.” New York Times, June 10, 2020.

Goldstein, Dana. 2020b. “Why a Pediatric Group Is Pushing to Reopen Schools This Fall.” New York Times, June 30, 2020.

Gonzalez, Jennifer. 2018. “10 Ways Educators Can Take Action in Pursuit of Equity: My interview with Pedro Noguera.” Cultofpedagogy.com, December 2, 2018.

Goodman, Joshua. 2014. “Flaking Out: Student Absences and Snow Days as Disruptions of Instructional Time,” National Bureau of Economic Research Working Paper no. 20221.

Gordon, Edmund W. 2013. “Community View: Standardized Testing Is Ineffective, Immoral Policy.” Journal News, April 15, 2013.

Gottfried, Michael A., and Stacy B. Ehrlich. 2018. “Introduction to the Special Issue: Combating Chronic Absence.” Journal of Education for Students Placed at Risk 23, nos. 1–2: 1–4. https://doi.org/10.1080/10824669.2018.1439753.

Gould, Elise. 2020a. “Six Months Into the Recession and an 11.5 Million Jobs Deficit Remains.” Economic Policy Institute Economic Indicators, September 4, 2020.

Gould, Elise. 2020b. “Public Education Job Losses in April Are Already Greater Than in All of the Great Recession.” Working Economics Blog (Economic Policy Institute), June 3, 2020.

Gould, Elise, and Heidi Shierholz. 2020. Not Everybody Can Work From Home: Black and Hispanic Workers Are Much Less Likely to Be Able to Telework. Economic Policy Institute, March 19, 2020.

Grady, Sarah. 2017. “A Fresh Look at Homeschooling in the U.S.” NCES Blog (National Center for Education Statistics), September 26, 2017.

Greeley, Christopher S. 2020. “Child Maltreatment Prevention in the Era of Coronavirus Disease 2019.” JAMA Pediatrics. Published online August 03, 2020. https://doi.org/10.1001/jamapediatrics.2020.2776.

Greif Green, Jennifer, and Elizabeth Bettini. 2020. “Addressing Teacher Mental Health during the COVID-19 Pandemic.” Teachers College Record, July 31, 2020.

Griffith, Michael. 2020. “The Impact of the COVID-19 Recession on Teaching Positions.” Learning Policy Institute Blog, April 30, 2020.

Hamilton, Laura S., Julia H. Kaufman, and Melissa Diliberti. 2020. Teaching and Leading Through a Pandemic: Key Findings from the American Educator Panels Spring 2020 COVID-19 Surveys, RAND Corporation.

Hanushek, Eric A. 1979. “Conceptual and Empirical Issues in the Estimation of Educational Production Functions.” Journal of Human Resources 14, no. 3: 351–388. https://doi.org/10.2307/145575.

Harper, Averi. 2020. “COVID-19 Exposes Mistrust, Health Care Inequality Going Back Generations for African Americans.” ABC News, April 28, 2020.

Heald-Sargent, T., W.J. Muller, X. Zheng, J. Rippe, A.B. Patel, and L.K. Kociolek. 2020. “Age-Related Differences in Nasopharyngeal Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Levels in Patients With Mild to Moderate Coronavirus Disease 2019 (COVID-19).” JAMA Pediatrics. Published online July 30, 2020. https://doi.org/10.1001/jamapediatrics.2020.3651.

Hodges, Charles, Stephanie Moore, Barb Lockee, Torrey Trust, and Aaron Bond. 2020. “The Difference Between Emergency Remote Teaching and Online Learning.” Educause Review, March 27, 2020.

Idele, Priscilla; David Anthony,. Lynne M. Mofenson, Jennifer Requejo, Danzhen You, Chewe Luo, and Stefan Peterson. 2020. “The Evolving Epidemiologic and Clinical Picture of SARS-CoV-2 and COVID-19 Disease in Children and Young People.” Innocenti Working Paper no. WP-2020-07, UNICEF Office of Research – Innocenti.

Imberman, Scott A., Adriana D. Kugler, and Bruno I. Sacerdote. 2012. “Katrina’s Children: Evidence on the Structure of Peer Effects from Hurricane Evacuees.” American Economic Review 102, no. 5: 2048–2082. https://doi.org/10.1257/aer.102.5.2048.

Institute of Education Sciences (IES), U.S. Department of Education. 2020. Dropout Prevention in the Time of COVID-19, Regional Educational Laboratory Appalachia.

Irons, John. 2009. Economic Scarring: The Long-Term Impacts of the Recession, Economic Policy Institute, September 30, 2009.

Jackson, Kirabo C., Cora Wigger, and Heyu Xiong. 2018. “Do School Spending Cuts Matter? Evidence from The Great Recession.” National Bureau of Economic Research Working Paper no. 24203.

Jin Jez, Su, and Robert W. Wassmer. 2013. “The Impact of Learning Time on Academic Achievement.” Education and Urban Society 47, no. 3: 284–306. https://doi.org/10.1177/0013124513495275.

Jones, Stephanie, et al. 2016. What Is the Same and What Is Different? Making Sense of the “Non-Cognitive” Domain: Helping Educators Translate Research into Practice. Harvard University EASEL Lab.

Jones, Stephanie, et al. 2019. “Explore SEL” (web page). Harvard University EASEL Lab.

Jordan, Phyllis W. 2019. Attendance Playbook: Smart Solutions for Reducing Chronic Absenteeism, Future Ed (McCourt School of Public Policy, Georgetown University) and Attendance Works.

Jordan, Phyllis W. 2020a. “Kids Lose Access to Critical Health Care Source When Schools Shutter Due to COVID-19.” Georgetown University Health Policy Institute Center for Children and Families. April 6, 2020.

Jordan, Phyllis W. 2020b. “What Congressional Covid Funding Means for K-12 Schools.” Future Ed (McCourt School of Public Policy, Georgetown University), August 9, 2020.

J-PAL. 2017. “Roll Call: Getting Children into School.” J-PAL Policy Bulletin, August 2017.

Junco, R. 2012. “Too Much Face and Not Enough Books: The Relationship Between Multiple Indices of Facebook Use and Academic Performance.” Computers in Human Behavior 28: 187–198.  https://doi.org/10.1016/j.chb.2011.08.026.

Junco, R., and S.R. Cotten. 2012. “No A 4 U: The Relationship Between Multitasking and Academic Performance.” Comput. Educ. 59: 505–514. https://doi.org/10.1016/j.compedu.2011.12.023.

Kamenetz, Anya. 2020a. “I Was a Screen–Time Expert: Then the Coronavirus Happened.” New York Times, July 30, 2020.

Kamenetz, Anya. 2020b. “4 In 10 U.S. Teens Say They Haven’t Done Online Learning Since Schools Closed.” NPR, April 8, 2020.

Karpinski A.C., P.A. Kirschner, I. Ozer, J.A. Mellott, and P. Ochwo. 2012. “An Exploration of Social Networking Site Use, Multitasking, and Academic Performance Among United States and European University Students.” Computers in Human Behavior 29: 1182–1192. https://doi.org/10.1016/j.chb.2012.10.011.

Kidronl, Yael, and Jim Lindsay. 2014. The Effects of Increased Learning Time on Student Academic and Nonacademic Outcomes: Findings from a Meta-Analytic Review. U.S. Department of Education Institute of Education Sciences’ National Center for Education Evaluation and Regional Assistance and Regional Educational Laboratory Appalachia.

Kim, E. Tammy. 2019. “The Messy Reality of Personalized Learning: Untangling the Mixed Record of the Latest Big-Fix Educational Trend Promoted by Silicon Valley.” The New Yorker, July 10, 2019.

Kirk, Joy. 2019. “Teachers Are Always There to Help, But Now We’re the Ones Who Need a Boost.” Working Economics Blog (Economic Policy Institute), June 14, 2019.

Kirschner, Paul A., and Aryn C. Karpinski. 2010. “Facebook® and Academic Performance.” Computers in Human Behavior 26: 1237–1245. https://doi.org/10.1016/j.chb.2010.03.024.

Kostyo, Stephen, Jessica Cardichon, and Linda Darling-Hammond. 2018. Making ESSA’s Equity Promise Real: State Strategies to Close the Opportunity Gap. Learning Policy Institute, September 2018.

Kraft, Matthew A., David Blazar, and Dylan Hogan. 2018. “The Effect of Teacher Coaching on Instruction and Achievement: A Meta-Analysis of the Causal Evidence.” Review of Educational Research 88, no. 4: 547–558. https://doi.org/10.3102/0034654318759268.

Kuhfeld, Megan. 2019. “Rethinking Summer Slide: The More You Gain, the More You Lose.” KappanOnline, June 6.

Lambert, Diana, Michael Burke, and Ali Tadayon. “Spike in Coronavirus Cases Means Some Schools Won’t Open at All This Fall,” EdSource, July 10, 2020.

Leachman, Michael, and Eric Figueroa. 2019. K-12 School Funding Up in Most 2018 Teacher-Protest States, But Still Well Below Decade Ago. Center on Budget and Policy Priorities, March 2019.

Lee, Thea. 2020a. “HEROES Act Provides Critical Relief and Recovery Measures to U.S. Workers” (Statement), Economic Policy Institute, May 12, 2020.

Lee, Thea. 2020b. “The GOP HEALS Act Fails to Heal People Harmed by the Coronavirus, Will Cost Millions of Jobs, and Protects Bad Employers” (statement), Economic Policy Institute, July 28, 2020.

Lemola, S., N. Perkinson-Gloor, S. Brand, J.F. Dewald-Kaufmann, A. Grob. 2015. “Adolescents’ Electronic Media Use at Night, Sleep Disturbance, and Depressive Symptoms in the Smartphone Age.” Journal of Youth Adolescence 44, no. 2: 405–418. https://doi.org/10.1007/s10964-014-0176-x.

Lepp A., J.E. Barkley, A.C. Karpinski. 2014. “The Relationship Between Cell Phone Use, Academic Performance, Anxiety, and Satisfaction With Life in College Students.” Computers in Human Behavior 31: 343–350. https://doi.org/10.1016/j.chb.2013.10.049.

Lerner, Michele. 2020. “One Home, a Lifetime of Impact.” Washington Post, July 23, 2020.

Levin, Henry M. 2012. “More Than Just Test Scores.” Prospects 42: 269–284. https://doi.org/10.1007/s11125-012-9240-z.

Lippman, L.H., R. Ryberg, R. Carney, and K.A. Moore. 2015. Key “Soft Skills” That Foster Youth Workforce Success: Toward a Consensus Across Fields, Child Trends, FHI360, and USAID, 2015.

Lisinski, Chris. 2020. “Massachusetts Scraps School Testing, Modifies Graduation Requirements.” WGBH. April 11, 2020.

Loewus, Liana. 2019. “A Clearer Vision for Teacher Professional Learning.” Education Week, May 14, 2019.

Lubienski, Christopher, Tiffany Puckett, and T. Jameson Brewer. 2013. “Does Homeschooling ‘Work’? A Critique of the Empirical Claims and Agenda of Advocacy Organizations.” Peabody Journal of Education 88, no. 3, 378–392. https://doi.org/10.1080/0161956X.2013.798516.

Marcotte, Dave E. 2007. “Schooling and Test Scores: A Mother-Natural Experiment.” Economics of Education Review 26, no. 5: 629–640.

Marcotte, Dave E., and Benjamin Hansen. 2010. “Time for School? When the Snow Falls, Test Scores also Drop.” Education Next, Winter 2010: 52–59.

Marcotte, Dave E., and Steven W. Hemelt. 2008. “Unscheduled School Closings and Student Performance.” Education Finance and Policy 3, no. 3: 316–338. https://doi.org/10.1162/edfp.2008.3.3.316.

McCombs, Jennifer Sloan, Catherine H. Augustine, Fatih Unlu, Kathleen M. Ziol-Guest, Scott Naftel, Celia J. Gomez, Terry Marsh, Goke Akinniranye, and Ivy Todd. 2019. Investing in Successful Summer Programs: A Review of Evidence Under the Every Student Succeeds Act. Santa Monica, CA: RAND Corporation, 2019.

McNichol, Elizabeth, and Michael Leachman. 2020. States Continue to Face Large Shortfalls Due to COVID-19 Effects, Center on Budget and Policy Priorities, July 2019.

Meltzer, Erica, Yesenia Robles, and Gabrielle LaMarr LeMee. 2020. “Why Colorado School Districts Are Serving Fewer Meals During Coronavirus Closures.” Denver Post, May 12.

Menas, Amanda. 2019. “The Widening Mental Health Treatment Gap in Schools.” NEA Today, May 2, 2019.

Mishel, Lawrence, and Richard Rothstein. 2003. The Class Size Debate. Washington, D.C.: Economic Policy Institute.

Morsy, Leila, and Richard Rothstein. 2019. Toxic Stress and Children’s Outcomes: African American Children Growing Up Poor Are at Greater Risk of Disrupted Physiological Functioning and Depressed Academic Achievement, Economic Policy Institute, May 1, 2019.

National Academies of Sciences, Engineering, and Medicine. 2020. Reopening K-12 Schools During the COVID-19 Pandemic: Prioritizing Health, Equity, and Communities. Washington, DC: The National Academies Press. https://doi.org/10.17226/25858.

National Center for Education Statistics (NCES), U.S. Department of Education. 2018. “T. 206.10. Number and Percentage of Homeschooled Students Ages 5 Through 17 with a Grade Equivalent of Kindergarten Through 12th Grade, by Selected Child, Parent, and Household Characteristics: Selected Years, 1999 Through 2016.” https://nces.ed.gov/programs/digest/d18/tables/dt18_206.10.asp.

National Center for Education Statistics (NCES), U.S. Department of Education. 2019a. “Table 216.20. Number and Enrollment of Public Elementary and Secondary Schools, by School Level, Type, and Charter, Magnet, and Virtual Status: Selected Years, 1990–91 Through 2017–18” Digest of Education Statistics: 2019, Institute of Education Sciences, U.S. Department of Education, December 2019. https://nces.ed.gov/programs/digest/d19/tables/dt19_208.20.asp.

National Center for Education Statistics (NCES), U.S. Department of Education. 2019b. “Table 208.20. Public and Private Elementary and Secondary Teachers, Enrollment, Pupil/Teacher Ratios, and New Teacher Hires: Selected Years, Fall 1955 Through Fall 2028.” Digest of Education Statistics: 2019, Institute of Education Sciences, U.S. Department of Education, November 2019. https://nces.ed.gov/programs/digest/d19/tables/dt19_216.20.asp.

National Center for Education Statistics (NCES), U.S. Department of Education. Early Childhood Longitudinal Study, Kindergarten Class of 2010–11 (ECLS-K 2010–2011).

National Education Association (NEA). 2020. All Hands On Deck: Guidance Regarding Reopening School Buildings, July 2020.

National Education Policy Center (NEPC). 2020. “The Misguided Push to Reintroduce Standardized Testing During the Pandemic.” NEPC Newsletter, August 6, 2020.

National Institute on Retirement Security (NIRS). 2020. “Will COVID-19 Trigger Teacher Retirements?” (webinar). Expert speakers: Dan Doonan, David Lamoureux, Paul Angelo, Rocky Joyner. August 14, 2020.

National Superintendents Roundtable. 2020. “10 Takeaways to Safely Reopen Schools. Infectious Disease Specialists Speak.” Roundtable News, July 19, 2020.

Network for Public Education (NPE). 2020. “Diane Ravitch in Conversation with Michael Hynes” (webinar), May 13, 2020.

Neuman, Ari, and Oz Guterman. 2016. “Academic Achievements and Homeschooling—It All Depends on the Goals.” Studies in Educational Evaluation 51: 1–6. https://doi.org/10.1016/j.stueduc.2016.08.005.

New York City (NYC) Department of Education. 2020. “Getting Started With Your iPad” (web page), accessed September 3, 2020.

Nickow, Andre, Philip Oreopoulos, and Vincent Quan. 2020. “The Impressive Effects of Tutoring on PreK-12 Learning: A Systematic Review and Meta-Analysis of the Experimental Evidence.” National Bureau of Economic Research Working Paper no. 27476.

Oakes, Jeannie, Anna Maier, and Julia Daniel. 2017. Community Schools: An Evidence-Based Strategy for Equitable School Improvement, National Education Policy Center and Learning Policy Institute, June 2017.

Organisation for Economic Co-operation and Development (OECD). 2009. “The Usefulness of PISA Data for Policy Makers, Researchers and Experts on Methodology.” In PISA Data Analysis Manual: SPSS, Second Edition. Paris: OECD Publishing. https://doi.org/10.1787/9789264056275-2-en.

Organisation for Economic Co-operation and Development (OECD). 2018. Program for International Student Assessment (PISA) data for 2018.

Organisation for Economic Co-operation and Development (OECD). 2019. TALIS 2018 Results (Volume I): Teachers and School Leaders as Lifelong Learners. Paris: TALIS, OECD Publishing. https://doi.org/10.1787/1d0bc92a-en.

Oxfam International. 2007. Climate Alarm: Disasters Increase as Climate Change Bites.

Özek, Umut. 2020. “Examining the Educational Spillover Effects of Severe Natural Disasters: The Case of Hurricane Maria.” American Institutes for Research/Calder Working Paper no. 233-0320, March 2020.

Page, Susan, 2020. “Back to School? 1 in 5 Teachers Are Unlikely to Return to Reopened Classrooms This Fall, Poll Says.” USA Today, May 26, 2020.

Pane, John F., Daniel F. McCaffrey, Nidhi Kalra, Annie J. Zhou. 2008. “Supplemental Materials For: Effects of Student Displacement in Louisiana During the First Academic Year After the Hurricanes of 2005.” RAND Corporation Working Paper no. WR-570-RC.

Parolin, Zachary, and Christopher Wimer. 2020. “Forecasting Estimates of Poverty During the COVID-19 Crisis.” Poverty & Social Policy Brief 4, no. 6: April 16, 2020. Center on Poverty and Social Policy at Columbia University.

Partelow, Lisette, Jessica Yin, and Scott Sargrad. 2020. Why K-12 Education Needs More Federal Stimulus Funding. Center for American Progress, July 2020.

Petway II, Kevin T., Meghan W. Brenneman, and Patrick C. Kyllonen Kyllonen. 2016. “Connecting Noncognitive Development to the Educational Pipeline.” In Non-cognitive Factors and Educational Attainment, edited by Myint Swe Khine and Shaljan Areepattamannil. Boston: Sense Publishers

Pollán, Marina, et al. 2020. “Prevalence of SARS-CoV-2 in Spain (ENE-COVID): A Nationwide, Population-Based Seroepidemiological Study.” The Lancet, July 6, 2020. https://doi.org/10.1016/S0140-6736(20)31483-5,

Progressive Caucus Action Fund. 2020. “Coronavirus Explainer: Comparison of the Heroes Act and HEALS Act.” Last updated August 4, 2020.

Prothero, Arianna. 2020. “Social Emotional Learning and School Reopenings: A Guide for Schools.” EdWeek, July 14, 2020.

Putnam, Robert. 2015. Our Kids: The American Dream in Crisis. New York: Simon and Schuster.

Quinn, D.M., N. Cooc,, J. McIntyre, and C.J. Gomez. 2016. “Seasonal Dynamics of Academic Achievement Inequality by Socioeconomic Status and Race/Ethnicity: Updating and Extending Past Research With New National Data.” Educational Researcher 45, no. 8: 443–453. https://doi.org/10.3102/0013189X16677965.

Ravitch, Diane. 2011. The Death and Life of the Great American School System: How Testing and Choice Are Undermining Education. New York: Basic Books.

Ravitch, Diane. 2020. Slaying Goliath: The Passionate Resistance to Privatization and the Fight to Save America’s Public Schools. New York: Alfred A. Knopf.

Ray, Brian D. 2017a. “A Systematic Review of the Empirical Research on Selected Aspects of Homeschooling as a School Choice.” Journal of School Choice 11, no. 4: 604–621. https://doi.org/10.1080/15582159.2017.1395638.

Ray, Brian. 2017b. “A Review of Research on Homeschooling and What Might Educators Learn?.” Dossiê: Homeschooling e o Direito à Educação. http://dx.doi.org/10.1590/1980-6248-2016-0009.

Reardon, Sean F. 2011. “The Widening Academic Achievement Gap Between the Rich and the Poor: New Evidence and Possible Explanations.” In Whither Opportunity?: Rising Inequality, Schools, and Children’s Life Chances, edited by Greg J. Duncan and Richard Murnane. New York: Russell Sage Foundation.

Reber, Sarah, and Nora Gordon. 2020. “Schools Need Flexible Federal Aid Now: What Should Congress Do Next?” Brown Center Chalkboard (The Brookings Institution), July 27, 2020.

RESEARCHED. 2020. “Formative Assessment in the Age of Covid: How Can Practitioners Apply Evidence-based Assessment Practices to Today’s Challenges?” (panel discussion video: moderator: Joshua Starr; panelists: Paul Zinni, Melissa Spadin, Kristin Huff, Lauren Merkley, Utah Teacher of the Year 2020). August 10, 2020.

Rogers, Chris, and Noelle Ellerson Ng. 2020. Memorandum: Report of Initial Findings: COVID-19 Impact on Public Schools. AASA: The School Superintendents Association. March 27, 2020.

Rosen, L.D., L. Mark Carrier, and N.A. Cheever. 2013. “Facebook and Texting Made Me Do It: Media-Induced Task-Switching While Studying.” Computers in Human Behavior 29: 948–958. https://doi.org/10.1016/j.chb.2012.12.001.

Rothstein, Richard. 2004. Class and Schools: Using Social, Economic, and Educational Reform to Close the Achievement Gap. Washington, D.C.: Economic Policy Institute.

Rothstein, Richard. 2020. “The Coronavirus Will Explode Achievement Gaps in Education.” Shelterforce, April 13, 2020.

Rothstein, Richard, Rebecca Jacobsen, and Tamara Wilder. 2008. Grading Education: Getting Accountability Right. New York: Teachers College Press and Economic Policy Institute.

Sacerdote, Bruce. 2012. “When the Saints Go Marching Out: Long-Term Outcomes for Student Evacuees from Hurricanes Katrina and Rita.” American Economic Journal: Applied Economics 4, no. 1: 109-135. https://doi.org/10.1257/app.4.1.109.

Sandoval, Edgar. 2020. “Like a Horror Movie: A Small Border Hospital Battles the Coronavirus.” New York Times, August 4, 2020.

Save the Children. 2008. In the Face of Disaster: Children and Climate Change.

Save the Children. 2013. Attacks on Education: The Impact of Conflict and Grave Violations on Children’s Futures.

Schanzenbach, Diane Whitmore. 2020. “The Economics of Class Size.” In The Economics of Education (Second Edition), A Comprehensive Overview, edited by Steve Bradley and Colin Green, 321–331. Chicago: University of Chicago Press. https://doi.org/10.1016/B978-0-12-815391-8.00023-9.

Schanzenbach, Diane Whitmore, and Natalie Tomeh. 2020. “Visualizing Food Insecurity: App Offers Snapshot of Weekly National and State-by-State Averages.” Institute For Policy Research, Northwestern University, July 14, 2020.

Schwartz, Sarah. 2019. “What Do Teachers Really Want from Professional Development? Respect.” Education Week, May 15, 2019.

Shafiq, M. Najeeb. 2010. “The Effect of an Economic Crisis on Educational Outcomes: An Economic Framework and Review of the Evidence.” Current Issues in Comparative Education 12, no. 2: 1523–1615.

Shierholz, Heidi. 2020. “Total Initial UI Claims Have Risen in Each of the Last Four Weeks: Congress Must Act.” Working Economics Blog (Economic Policy Institute), September 3, 2020.

Shonkoff, J.P., and D.A. Phillips. 2000. From Neurons to Neighborhoods: The Science of Early Childhood Development. Washington, D.C.: National Academies Press.

Shonkoff, Jack P., and David R. Williams. 2020. Thinking About Racial Disparities in COVID-19 Impacts Through a Science-Informed, Early Childhood Lens. The Center on the Developing Child at Harvard University, April 2020.

Snell, Kelsey. 2020. “Here’s How Much Congress Has Approved For Coronavirus Relief So Far And What It’s For.” NPR, May 15, 2020.

Soldner, Matthew. 2020. “Seeking Your Help in Learning More About What Works in Distance Education: A Rapid Evidence Synthesis.” NCEE Blog, March 26, 2020.

Sommers, Marc. 1999. Emergency Education for Children.

Southern Education Foundation. 2015. A New Majority Research Bulletin: Low-Income Students Now a Majority in the Nation’s Public Schools. January 2015.

Stancati, Margherita, Leslie Brody, and Xavier Fontdeglòria. 2020. “The Pandemic Sent 1.5 Billion Children Home From School: Many Might Not Return.” Wall Street Journal, June 3, 2020.

Starr, Joshua. 2020. Responding to COVID-19: Short- and Long-Term Challenges, KappanOnline, March 23, 2020.

St. George, Donna. 2020. “Montgomery County Schools Launch Remote Learning: ‘It Will be Chaotic.’” Washington Post, March 30, 2020.

Stratford, Brandon. 2020. As Schools Reopen, Addressing COVID-19–Related Trauma and Mental Health Issues Will Take More Than Mental Health Services. Child Trends, July 2020.

Strauss, Valerie. 2020. “How to Stop Magical Thinking in School Reopening Plans.” Washington Post, July 30, 2020.

Streicher, Stan. 2020. “Anne Arundel County Public Schools Has Handed Out 1M Meals During Coronavirus Pandemic.” CBS Baltimore, May 12, 2020.

Swaby, Aliyya. 2020. “Texas STAAR Test Requirements Waived Due To Coronavirus Outbreak.” Texas Public Radio. March 17, 2020.

Thompson, Paul N. 2019. Effects of Four-Day School Weeks on Student Achievement: Evidence from Oregon, IZA DP No. 12204, March 2019.

Tinubu Ali, Titlayo, and Mirel Herrera. 2020. Distance Learning During COVID-19: 7 Equity Considerations for Schools and Districts. Southern Education Foundation, April 2020.

Tirivayi, Nyasha, Dominic Richardson, Maja Gavrilovic, Valeria Groppo, Lusajo Kajula, Elsa Valli, and Francesca Viola. 2020. “A Rapid Review of Economic Policy and Social Protection Responses to Health and Economic Crises and Their Effects on Children: Lessons for the COVID-19 Pandemic Response.” Innocenti Working Paper no. 2020-02, UNICEF Office of Research – Innocenti, Florence.

Todd, Petra E., and Kenneth I. Wolpin. 2003. “On the Specification and Estimation of the Production Function for Cognitive Achievement.” Economic Journal 113, 485: F3–F33. https://doi.org/10.1111/1468-0297.00097.

Tolerance Trauma. 2020. A Trauma-Informed Approach to Teaching Through Coronavirus. March 23, 2020.

Torres, Stacy. 2020. “After Coronavirus, Expect High School Dropout Wave: 9/11 Was the Trigger for My Sisters.” USA Today, April 30, 2020.

Tough, Paul. 2012. How Children Succeed: Grit, Curiosity, and the Hidden Power of Character. New York: Mariner Books.

UNESCO, UNICEF, World Bank, World Food Programme, and UNHCR. 2020. Framework for Reopening Schools, April 2020.

UNESCO, the World Bank, UNFPA, UNDP, UN Women, and UNHCR. 2016. Incheon Declaration and Framework for Action for the Implementation of Sustainable Development Goal 4.

United Nations. 2020. Education During COVID-19 and Beyond. Policy Brief. August 2020.

United Nations Inter-Agency Standing Committee (UN IASC). 2007. Guideline: Inter-Agency Contingency Planning for Humanitarian Assistance IASC Sub-Working Group on Preparedness, November 2007.

United Nations Inter-Agency Standing Committee (UN IASC). 2015. Guideline: Emergency Response Preparedness July 2015 IASC Task Team on Preparedness and Resilience, July 2015.

United States Agency for International Development (USAID). 2014. Guide to Education in Natural Disasters: How USAID Supports Education in Crises. January 2014.

United States Interagency Council on Homelessness (USICH). 2020. “Supporting Children and Youth Experiencing Homelessness During the COVID-19 Outbreak: Questions to Consider” (web page) March 16, 2020.

U.S. Census Bureau. 2019. “CPS Historical Time Series Tables on School Enrollment Table A-1. School Enrollment of the Population 3 Years Old and Over, by Level and Control of School, Race, and Hispanic Origin: October 1955 to 2018,” December 2019.

U.S. Census Bureau. 2020a. Household Pulse Survey Data Tables, accessed July 2020.

U.S. Census Bureau. 2020b. 2018 Public Elementary-Secondary Education Finance Data, last revised April 14, 2020

U.S. Department of Education. 2016. “Chronic Absenteeism in the Nation’s Schools: An Unprecedented Look at a Hidden Educational Crisis” (online fact sheet), accessed August 2018.

U.S. Senate. 2020. “Coronavirus Child Care and Education Relief Act” (fact sheet), accessed September 4, 2020.

Valant, Jon. 2020. “School Reopening Plans Linked to Politics Rather than Public Health.” Brown Center Chalkboard (Brookings Institution blog), July 29, 2020.

von Hippel, Paul T. 2020. “Year-Round School Calendars: Effects on Summer Learning, Achievement, Parents, Teachers, and Property Values,” EdWorkingPaper no. 20-209. Annenberg Institute at Brown University.

von Hippel, Paul T., and Caitlin Hamrock. 2019. “Do Test Score Gaps Grow Before, During, or Between the School Years? Measurement Artifacts and What We Can Know in Spite of Them.” Sociological Science. http://dx.doi.org/10.15195/v6.a3.

von Hippel, Paul T., J. Workman, and D.B.  Downey. 2018. “Inequality in Reading and Math Skills Forms Mainly Before Kindergarten: A Replication, and Partial Correction, of ‘Are Schools the Great Equalizer?’” Sociology of Education 91, no. 4: 323–357. https://doi.org/10.1177/0038040718801760.

Wadhera, Rishi K., Priya Wadhera, and Prakriti Gaba et al. 2020. “Variation in COVID-19 Hospitalizations and Deaths Across New York City Boroughs.” JAMA Network Research Letter, April 29, 2020. https://jamanetwork.com/journals/jama/fullarticle/2765524.

Walker, Darren. 2020. “Extraordinary Times, Extraordinary Measures.” Equal Change Blog (Ford Foundation), June 11, 2020.

Walker, Tim. 2020. “Social-Emotional Learning Should Be Priority During COVID-19 Crisis.” NEA Today, April 15, 2020.

Walsh, Jeremy J., Joel D. Barnes, Jameason D. Cameron, Gary S. Goldfield, Jean-Philippe Chaput, Katie E. Gunnell, Andrée-Anne Ledoux, Roger L. Zemek, and Mark S. Tremblay. 2018. “Associations Between 24 Hour Movement Behaviours and Global Cognition in US Children: A Cross-sectional Observational Study.” The Lancet Child & Adolescent Health,  https://doi.org/10.1016/S2352-4642(18)30278-5.

Waxman, Elaine, Poonam Gupta, and Michael Karpman. 2020. More Than One in Six Adults Were Food Insecure Two Months Into the COVID-19 Recession: Findings from the May 14–27 Coronavirus Tracking Survey. Urban Institute, July 2020.

Weiner, Rachel. 2020. “More Than 80 Percent of Hospitalized COVID-19 Patients in Georgia Were African American, Study Finds.” Washington Post, April 29.

Weingarten, Randi. 2020. “How to Cap This Unprecedented School Year.” AFT Voices, March 20, 2020.

Weiss, Elaine, and Paul Reville. 2019. Broader, Bolder, Better: How Schools and Communities Help Students Overcome the Disadvantages of Poverty. Cambridge, Mass.: Harvard Education Publishing Group.

The White House. 2006. The Federal Response to Hurricane Katrina: Lessons Learned.

Wicks, Matthew. 2010. A National Primer on K-12 Online Learning. Version 2. International Association for K-12 Online Learning, October 2010.

Will, Madeline. 2020. “Teachers Say They’re More Likely to Leave the Classroom Because of Coronavirus.” Education Week, June 3, 2020.

Wilmer, Henry H., Lauren E. Sherman, and Jason M. Chein. 2017. “Smartphones and Cognition: A Review of Research Exploring the Links Between Mobile Technology Habits and Cognitive Functioning.” Frontiers in Psychology 8: 1–16. http://dx.doi.org/10.3389/fpsyg.2017.00605.

Woodworth, J.L., M.E. Raymond, K. Chirbas, M. Gonzalez, Y. Negassi, W. Snow, and C. Van Donge. 2015. Online Charter School Study 2015. Center for Research on Education Outcomes, Stanford University, 2015.

Zhao, Yong. 2020. “Does It Work? The Most Meaningless Question to Ask about Online Education.” Zhaolearning.com (blog), March 28, 2020.

What are the consequences of this lack of progress in integrating schools for black children?

• It depresses education outcomes for black students; as shown in this report, it lowers their standardized test scores.
• It widens performance gaps between white and black students.
• It reflects and bolsters segregation by economic status, with black students being more likely than white students to attend high-poverty schools.
• It means that the promise of integration and equal opportunities for all black students remains an ideal rather than a reality.

In contrast, when black students have the opportunity  to attend schools with lower concentrations of poverty and larger shares of white students they perform better, on average, on standardized tests.

Black children are still relegated to separate and unequal schools

Findings on school segregation and student performance come from the National Center for Education Statistics’ National Assessment of Educational Progress (NAEP), the most comprehensive study of education performance in the country. We use the most recently released data to describe school segregation and its consequences for math performance of eighth-graders. These data show that only about one in eight white students (12.9%) attends a school where a majority of students are black, Hispanic, Asian, or American Indian. (We refer to this group collectively as students of color hereafter.) In contrast, nearly seven in 10 black children (69.2%) attend such schools (see Figure A).

As shown in Figure B, black students are also in economically segregated schools. Less than one in three white students (31.3%) attend a high-poverty school, compared with more than seven in 10 black students (72.4%).

Figure A

Figure B

In America, race and poverty are intertwined, doubly disadvantaging black students

The known connection between race/ethnicity and poverty in the United States appears in data on the composition of schools attended by for black children. Figure C shows that a black child faces a very high probability of ending up in a school where a majority of her peers are both poor and students of color. While less than 1 in 10 white students (8.4%) attend high-poverty schools with a high share of students of color, six in 10 black students (60.0%) do.

In contrast, about a fourth of white students (23.5%) attend schools where most of their peers are white and not poor, while only 3.1 percent of black children attend such schools.

Figure C

When black children have the opportunity to attend the same schools that white children routinely attend, black children perform markedly better on standardized math tests, which we use here as a measure of education performance.

Figure D shows math scores of black eighth-graders in low-poverty, mostly white schools and in high-poverty schools with a high share of students of color. In high-poverty schools with a high share of students of color, black students scored on average 20 points less on standardized math tests than their counterparts in low-poverty, mostly white schools (255.4 vs. 275.3). In other words, scores are much lower in the type of school that black children are overwhelmingly more likely to attend (high-poverty, mostly students of color) than in the type of school (low-poverty, mostly white) that only 3.1% of black children have a chance of attending.

Though not shown in the figure, the gap between black and white student test scores is larger in high-poverty schools with a high share of students of color than in low-poverty, mostly white schools. By promoting policies that facilitate a shift away from our current pattern of heavily segregated schools, we would thus help close the gap between black and white students overall.

Figure D

Unaddressed school segregation is a major longstanding policy failure. It consigns most black children to schools that put them behind academically. The persistent performance gaps between white and black children that challenge the education and career prospects of black children from early on demonstrate that school segregation continues to cast a very long shadow—from well before Brown v. Board of Education to today, and into the future.

This brief, published by EPI to highlight education issues for Black History Month, shows data that are part of ongoing EPI research on student performance and education inequalities. Information using earlier data on segregation and the consequences for performance for other groups, and technical details, are available in Martin Carnoy and Emma García, Five Key Trends in U.S. Student Performance: Progress by Blacks and Hispanics, the Takeoff of Asians, the Stall of non-English Speakers, the Persistence of Socioeconomic Gaps, and the Damaging Effect of Highly Segregated Schools, Economic Policy Institute, 2017.

This report is the fifth in a series examining the magnitude of the teacher shortage and the working conditions and other factors that contribute to the shortage.

What this series finds: The teacher shortage is real, large and growing, and worse than we thought. When indicators of teacher quality (certification, relevant training, experience, etc.) are taken into account, the shortage is even more acute than currently estimated, with high-poverty schools suffering the most from the shortage of credentialed teachers.

What this report finds: Our review of the early career supports, ongoing professional development, and opportunities for cooperation and influence offered to public school teachers reveals a mixed picture, with clear room to improve the system of professional supports that play a role in teacher retention and expand the knowledge base of the teaching workforce.

On the positive side, the set of supports already broadly offered in the schools is a strong foundation to build upon. Large shares of first-year teachers work with a mentor (79.9 percent) or participate in teacher induction programs (72.7 percent). And large shares of teachers generally are accessing certain types of professional development, including workshops or training sessions (91.9 percent), activities focused on the subjects that teachers teach (85.1 percent), regularly scheduled collaboration with other teachers on issues of instruction (80.8 percent), and opportunities to observe or be observed by other teachers in their classrooms (67.0 percent).

On the negative side, there are multiple weaknesses to address if we want to help teachers do their jobs better and advance in their careers:

• First, there is limited access to some of the types of professional development that are highly valued and more effective. Small shares of teachers attend university courses related to teaching (26.6 percent), present at workshops (23.1 percent), or make observational visits to other schools (21.6 percent).
• Second, novice and veteran teachers largely don’t get the time and resources they need to study, reflect, and prepare their practice. Small shares of first-year teachers are released from classroom instruction to participate in support activities for new or beginning teachers (37.1 percent); small shares receive teachers’ aides to enhance classroom management and one-on-one attention for students (26.9 percent); and small shares get a reduced teaching schedule (10.7 percent). For all teachers, only half have released time from teaching to participate in professional development (50.9 percent), less than a third are reimbursed for conferences or workshop fees (28.2 percent) or receive a stipend for professional development accessed outside of regular work hours (27.3 percent), and only one in 10 teachers (9.4 percent) receives full or partial reimbursement of college tuition.
• Third, teachers are not highly satisfied with their professional development experiences. Less than a third of teachers found any of the activities they accessed “very useful,” and over a third of novice teachers felt that working with a mentor was only a little or not at all helpful.
• Fourth, teachers are not by and large immersed in the kind of learning community that can support their teaching and career growth. In a learning community, teachers have opportunities to cooperate and coordinate and have a say in school policy and classroom instruction and management. We find that more than two-thirds of teachers report that they have less than a great deal of influence over what they teach in the classroom (71.3 percent) or what instructional materials they use (74.5 percent), which suggests low consideration for their knowledge and judgment. Just 11.1 percent of teachers report having a great deal of influence in determining the content of professional development programs.
• Fifth, some key resources and professional development opportunities are particularly lacking in high-poverty schools, where, if anything, stronger supports for teachers are needed. In high-poverty schools, compared with low-poverty schools, smaller shares of first-year teachers work with a mentor (78.3 percent vs. 83.7 percent) and say that working with a mentor helps a lot (32.1 percent vs. 34.5 percent). Compared with teachers in low-poverty schools, larger shares of teachers in high-poverty schools participate in professional development activities that they consider less useful (such as workshops, 92.7 percent vs. 90.8 percent) and smaller shares participate in activities that they find more useful (such as observational visits to other schools, 20.9 percent vs. 22.5 percent, and teacher-led research, 42.7 percent vs. 49.5 percent). High-poverty schools also score lower on most indicators that a school has a strong learning community.

Our data suggest a relationship between these systems of professional supports and teacher retention. When we compare teachers who stayed in teaching with those who quit teaching, we observe that larger shares of staying teachers had received early support in the form of an assigned mentor (77.0 percent vs. 69.2 percent) or induction programs (85.9 percent vs. 80.0 percent), had found their subject-specific professional development activities to be very useful (27.4 percent vs. 19.5 percent), had worked in highly cooperative environments (38.7 percent vs. 33.9 percent), and had felt they had more influence over the content taught in their classrooms (28.6 percent vs. 25.4 percent).

Why professional supports matter: The demands of teaching are constantly changing and teachers need to continually adapt their knowledge and practice. By failing to provide teachers with broad access to effective training and professional development, as well as to learning communities where their professional judgment is considered, we hurt teachers’ effectiveness, sense of purpose, and career advancement opportunities. This likely plays a role in the teacher shortage. And the teacher shortage—which is more acute in high-poverty schools—harms students and teachers and challenges the U.S. education system’s goal of providing a sound education equitably to all children.

What we can do to support teachers: We must improve both the types and the usefulness of the professional supports offered and ensure that teachers have the resources needed to access those opportunities. Strengthening the system of supports includes increasing teachers’ influence over their day-to-day work and developing cultures of learning. High-poverty schools and their teachers, in particular, require additional funding to close gaps in these resources and supports.

Introduction

The teacher shortage—the gap between the number of qualified teachers needed in the nation’s K–12 schools and the number available for hire in a given year—is an increasingly recognized but still poorly understood crisis. The shortage is discussed by the media and policymakers, and researchers have estimated its size (about 110,000 teachers short in the 2017–2018 school year, according to Sutcher, Darling-Hammond, and Carver-Thomas [2016]) and even quantified part of its cost.1 The shortage constitutes a crisis because of its negative effects on students, teachers, and the education system at large.2 But the shortage is poorly understood because it has multiple complex and interdependent causes. The first report in this series, The Teacher Shortage Is Real, Large and Growing, and Worse Than We Thought (García and Weiss 2019a), establishes that current national estimates of the teacher shortage likely understate the magnitude of the problem: When issues such as teacher qualifications and the unequal distribution of highly credentialed teachers across high- and low-poverty schools are taken into consideration, the teacher shortage problem is much more severe than previously identified.

The second report in this series, U.S. Schools Struggle to Hire and Retain Teachers (García and Weiss 2019b), builds on the research in the first report, employing the same quality and equity angles to show that schools are having difficulties filling teacher vacancies and are, in some cases, leaving vacancies unfilled despite actively trying to hire teachers to fill them. High-poverty schools are hit hardest: They find it more difficult to fill vacancies than do low-poverty schools and schools overall, and they experience higher turnover and attrition rates than do low-poverty schools. One factor behind staffing difficulties is the high share of public school teachers leaving their posts: 13.8 percent were either leaving their school or leaving teaching altogether in a given year, according to the most recent data. Another factor is the dwindling pool of applicants to fill vacancies: From the 2008–2009 to the 2015–2016 school year, the annual number of education degrees awarded fell by 15.4 percent, and the annual number of people who completed a teacher preparation program fell by 27.4 percent. Schools are also having a harder time retaining credentialed teachers, as is evident in the small but growing share of all teachers who are both newly hired and in their first year of teaching (4.7 percent) and in the substantial shares of teachers who are quitting who are certified and experienced. It is even more difficult for high-poverty schools to retain credentialed teachers.

The third report in the series focuses on one likely reason teachers are leaving the profession and fewer people are becoming teachers: low teacher pay. Specifically, Low Relative Pay and High Incidence of Moonlighting Play a Role in the Teacher Shortage, Particularly in High-Poverty Schools (García and Weiss 2019c) describes how teacher compensation compares with compensation in nonteaching occupations, and calls attention to the high share of teachers who supplement their earnings by moonlighting during the school year. The report shows a correlation between measures of teacher compensation and teachers leaving the profession. Specifically, it finds that teachers who ended up quitting teaching reported receiving, on average, lower salaries than those who stayed at their schools. They also reported participating less in the kinds of paid extracurricular activities that might complement their professional development (activities like coaching students or mentoring other teachers) than did teachers who stayed, and they reported participating more in working options outside the school system than did teachers who stayed. In high-poverty schools, teachers face compounded challenges. Relative to their peers in low-poverty schools, teachers in high-poverty schools are paid less and receive smaller amounts of income from moonlighting, and the moonlighting that they do is less likely to involve paid extracurricular or additional activities for the school system that also help them grow professionally.

The fourth report, Challenging Working Environments (‘School Climates’), Especially in High-Poverty Schools, Play a Role in the Teacher Shortage (García and Weiss 2019d), explores another likely factor behind the exodus of teachers from the profession and the shrinking supply of future teachers: the working environments, or school climates, in which teachers do their work. We show that school climate is challenging because of the presence of widespread barriers to teaching and learning, threats to teachers’ emotional well-being and physical safety, and a troubling lack of teacher influence over school policy and over what and how they teach in their classrooms. We observe that poor school climate affects teacher satisfaction, morale, and expectations about staying in the profession, and that there is a correlation between these indicators of difficult working environments and teachers leaving the profession a year later. Consistent with other gaps being more acute in high-poverty schools, we also document that school climates are more challenging in high-poverty schools than in low-poverty schools.

This report, the fifth in the series, examines the early career supports available to novice teachers in the first year of their careers, as well as the continued learning opportunities available to teachers throughout their careers. We also explore the extent to which certain aspects of the working environment—the presence or absence of supportive and collaborative relationships; cooperation among teachers, colleagues, and principals; and teachers’ influence over policy and day-to-day classroom decisions—establish a culture of learning in which teachers’ knowledge and professionalism are recognized and cultivated.

Unlike with some of the stressors identified in our earlier reports, it is not easy to trace a direct link between suboptimal professional supports and the teacher staffing crisis that is the focus of our series. (We discuss the less clear-cut nature of the role of professional supports in the next section of this report.) However, all the professional development and continuous training components we examine have the potential to help teachers do their jobs better, progress in their profession, and gain satisfaction with and a sense of ownership of their careers. These supports are essential to guaranteeing the quality of the teaching workforce and to professionalizing teaching. And, as we show, there is evidence, albeit less direct, that these supports play a role in the teacher shortage. Professional development and continuous training opportunities and the learning climate—or the lack thereof—can directly influence teacher retention and recruitment. In some cases, these factors can also indirectly influence retention and recruitment when they mitigate—or compound—problems already identified in previous reports in “The Perfect Storm in the Teacher Labor Market” series. The findings here suggest that efforts to address teacher shortages must include providing teachers with strengthened continuous training opportunities that professionalize teaching and support teachers, especially in high-poverty schools where the teacher shortage and the lack of meaningful professional development opportunities are most concerning.

Why professional development and continuous training are needed in teaching and how they are implicated in the teacher shortage

It is important to begin this installment of the “Perfect Storm” series with some clarifications about the limitations we face in this study. In the prior reports, we are able to identify shares of teachers experiencing factors that are objectively negative stressors. For example, we show that teachers are paid less than comparable workers and that their safety is not guaranteed. In those cases, their experiences are objectively bad, and the correlations between those experiences and the teachers’ choices to quit or stay in the profession could be seen as decisive.

With regard to professional development and continuous training opportunities, there is no universally accepted set of supports that constitutes a good, supportive early training and ongoing professional development system as opposed to a bad, unsupportive system. We lack research, policy, or practice recommendations that say, “This specific set of supports offered in this mode and style for this duration and on these contents is unequivocally essential to helping teachers succeed and keeping them in the classrooms.” Little is known regarding how teachers get assigned to training and professional development opportunities; whether teachers have any say in the opportunities presented to them; who offers the opportunities; the optimal duration, location, and timing of professional development opportunities; and who oversees the quality—let alone what funding is available. We also lack knowledge of how professional development needs may change depending on the teacher’s profile (including teacher credentials, his or her experience, the field of teaching, etc.) and circumstances (including whether he or she teaches in a high- or in a low-poverty school, etc.). And we lack confirmation of how these components of the system affect teachers’ effectiveness and transitions in and out of the teacher labor markets.

Nevertheless, this report analyzes the currently offered set of supports based on common sense—what should be broadly accepted assumptions about the importance of professional development and a helpful, supportive environment—and on evidence from research on professional development and school climate. With regard to our assumptions, we take as given that good professional development is critical in education just as it is in medicine, law, engineering, and other professions where continuous learning and professionalization are expected or mandated. Continuous learning via professional development helps teachers do their jobs more effectively and efficiently and advance in their careers, increasing their sense of dedication, purpose, satisfaction, and professionalism and, significantly, helping their students’ learning and performance as well. A strong learning community and a positive working environment also help teachers in these ways. It should thus be evident that professional supports can play an important role in increasing the availability of teachers and that they are critical to improving the quality of the teaching workforce—both key aspects of the teacher shortage problem.

With regard to the evidence, existing research presents many reasons why, in general, good professional development and continuous training opportunities (those that result in improvements to teacher practices and student outcomes) matter greatly (see Bill & Melinda Gates Foundation 2014; Darling-Hammond, Burns, et al. 2017; Darling-Hammond, Hyler, et al. 2017; ESSA 2015; Hill 2009; Ingersoll and Collins 2018; Jensen et al. 2016; Kraft, Blazar, and Hogan 2018; Learning Forward 2019; Mizell 2010; Robinson 2019).

First, teachers pursue professional development opportunities to earn a master’s degree, credit toward recertification, or other credential, or to gain additional qualifications to prepare for a leadership position.3

Second, continuous training and professional development help teachers develop new knowledge and skills to better serve their students. This includes helping teachers update their instructional techniques in response to new research on learning and teaching processes and to adjust to the needs of a more diverse student body.

Third, evidence-driven public policies have identified professional development as a key component in building systems of professional learning. For example, the Every Student Succeeds Act of 2015 (ESSA) made professional development an important cornerstone of schools’ improvement plans.4 ESSA’s recommendations and guidance are in part based on a body of research showing that solid early career supports and continuous training can strengthen teachers’ practices and effectiveness—i.e., that when teachers use what they learned in professional development, their practice improves in ways that benefit children.

Fourth, early career supports help new teachers transition successfully from teacher training programs to actually being in a classroom, and continuous training helps veteran teachers adapt to changes in what they need to teach and test—and in how they need to teach and test—to accommodate changes in state and federal laws and standards.

Fifth, when teachers have these training and professional development opportunities, it nurtures a culture of learning schoolwide: Teachers and staff routinely develop their own knowledge and skills; they model, for students, the belief that learning is important and useful; they feel more respected; and they see ways to progress in their careers.

And sixth, intense early supports, continuous training, and professional development are actually recommended, and the norm, in the most highly regarded systems—systems in which teaching is also a more prestigious and sought-after profession.

In this report we draw upon national public school teacher data to describe early career supports and ongoing professional development opportunities in detail: the kinds of opportunities available to teachers, how many teachers access them, whether teachers get fee reimbursements or other help accessing them, and how useful they are to teachers. We also examine teachers’ assessments of their relationships with school administrators and peer teachers and their involvement in setting schoolwide and classroom policies. All three of these components—early career supports, ongoing professional development, and relationships that capitalize on teachers’ professional judgment—are necessary to establish a culture of learning in the school and to provide teachers with meaningful career advancement pathways.5

Overview of findings

We find reasons for both optimism and concern. On the positive side, a great majority of teachers participate in some form of training, which could be seen as an asset of the system to be further cultivated.6 However, some of the most common types of professional development and training activities are also those that receive lower teacher satisfaction ratings. Teachers also report that they don’t get to choose or help design the professional development options offered to them. Together, these findings suggest critical weaknesses in the menu of options available, in how opportunities are assigned to teachers, or in the quality of the offerings (or all of the above). This disconnect between what teachers are receiving and what they find useful suggests that there is significant room for improvement in the supports and career advancement opportunities that schools and the profession offer. Teachers may be participating in the types of activities that are less useful to them because they are required to, either by law or to maintain certification, suggesting that these mandates are not informed by the type of support or the quality needed—but we can’t test any of these hypotheses with the available data.

With regard to the culture of learning, we present evidence that teachers have low levels of autonomy and influence in general, and that teachers’ relationships with one another, with administrators, and with parents have clear room for improvement. And, as past research shows, despite substantial shares of teachers moonlighting, they are not always taking on second jobs that foster collaboration and allow them to learn from or support one another (García and Weiss 2019c; García and Weiss 2019d; Mizell 2010; Ingersoll and Collins 2018).7 In short, there are many features of schools today that are not conducive to building a culture of learning, a culture in which obtaining useful and needed training is the norm and in which training benefits teachers and their students.

Finally, we demonstrate that there is a relationship between systems of professional supports—professional development, career-building supports, and teachers’ autonomy and influence—and whether or not teachers stay in the profession. This finding contributes to our efforts in this series of reports to identify factors that can help explain the most troubling trends in the teacher labor market: decreased interest in becoming a teacher and in staying in teaching (Ingersoll 2004, 2014; Sutcher, Darling-Hammond, and Carver-Thomas 2016; Darling-Hammond, Burns, et al. 2017; García and Weiss 2019b). Strengthening professional development and career-building supports would help professionalize teaching and provide teachers with opportunities for career advancement, which would make teaching a more appealing profession8 (and may mitigate some of the other factors driving the teacher shortage, such as tough working environments).

It is important to note that suboptimal professional supports not only likely play a role in the teacher shortage, but also directly affect the knowledge base of the teacher workforce. Thus while we are discussing professional development in the context of the teacher shortage, these broader repercussions add to the urgency of identifying whether ineffective and insufficient professional development is a problem in U.S. schools, i.e., affecting not only teacher recruitment and retention but also teacher effectiveness, student learning, school performance, and the health of the educational system overall.9

Novice teachers are not getting the support they need to translate their training into effective teaching

The first few years that a teacher spends in the classroom tend to be among the most difficult of his or her career. As is true of every job, getting to know and adjusting to the workplace—in this case, the school, district, colleagues, students, students’ parents, and surrounding communities—poses challenges. In the education context, these natural difficulties are compounded by the steep learning curve to balancing engaging instruction with effective classroom control. Regardless of how solid teachers’ preparation and innate ability may be, and despite new-job energy and motivation, all young teachers need to acclimate to the job and practice to strengthen their teaching. New teachers must translate what they learned in their teacher preparation programs into real classroom practices, and research shows that this process can be sped up if they receive specific supports to help them with the transition and make their teaching practice more effective. It is especially in teachers’ early years, then, that appropriate supports (including induction programs, mentors, and other classroom-based resources) are critical, both to helping teachers succeed and to retaining them—but these supports are not universally available to teachers (Darling-Hammond, Burns, et al. 2017; Liston, Whitcomb, and Borko 2006; NCEE 2016; Smith and Ingersoll 2004; Ingersoll and Collins 2018).10

Our analysis confirms that more supports for teachers are needed early in their careers. Table 1 shows that it is actually very common for novice teachers to feel less than very well prepared to handle all the tasks required in their classrooms in their first year on the job. The table shows the results of our analysis of questions posed to early-career teachers (teachers in their first five years of teaching) about how prepared they were, in their first year of teaching, to handle a range of classroom tasks. There is only one task out of the 10 listed here that even one in every three teachers felt “very well prepared” to perform: teaching their subject matter. For the other nine tasks listed in Table 1—including using a variety of instructional methods, assessing students, and differentiating instruction—the shares of teachers who reported that they felt very well prepared are much smaller. To look at this data another way, the shares of novice teachers who felt less than very well prepared (“not at all prepared,” “somewhat prepared,” or “well prepared”) to handle all the tasks in their classrooms and, thus, who could significantly benefit from early career supports, are large: For example, fully two-thirds (66.8 percent) of teachers reported feeling less than “very well prepared” to teach their subject matter and nine in 10 teachers (91.6 percent) reported not feeling “very well prepared” to teach English language learners (ELLs).11

Table 1

As noted earlier in this report, research to date on the topic of early career supports has produced no set menu of the exact supports schools need to offer to ensure that teachers are ready to teach from the start of their careers.12 However, it is obvious that a lack of these supports leaves novice teachers struggling to adjust by themselves, hurting their ability to use their time wisely and effectively, and precluding opportunities for novice teachers to learn from one another. Unfortunately, as other researchers have noted, novice teachers suffer from inadequate support for teacher learning, including inadequate peer support, challenging emotional experiences, and lack of development programs for teachers when they are on the job (i.e., “in-service education”) (Liston, Whitcomb, and Borko 2006; Dias-Lacy and Guirguis 2017). In our data, we also find that while most teachers do receive some forms of support and do participate in preparation programs in the early years, these supports and programs are not universally available, i.e., not all teachers get them. Some of the most important supports are available to only a minority of new teachers. We present the evidence for general supports in Table 2 and for specific early preparation programs in Table 3.

Table 2

As Table 2 shows, three resources that would increase the time that young teachers have to study, reflect, and prepare their practice are unavailable to large shares of new teachers: a reduced teaching schedule, teachers’ aides, and time away from the classroom to receive new-teacher supports. Only a little more than one-third (37.1 percent) of all teachers are released from classroom instruction to have time to participate in support activities for new or beginner teachers. Only about one-fourth (26.9 percent) receive aides to enhance classroom management and one-on-one attention for students, and only about one in ten (10.7 percent) get a reduced teaching schedule.

Four other classroom-based supports were more widely available, though still unevenly so, with less than two-thirds of first-year teachers enjoying common planning time with their fellow subject-matter teachers (61.2 percent), two-thirds participating in new-teacher classes (66.4 percent), and just over two-thirds receiving constructive feedback based on classroom observations (69.0 percent). About three in four (74.5 percent), however, did report “regular supportive communication” with the principal and others.

Differences between teachers based on the concentration of low-income students in their classrooms were small and mixed, with a split pattern of advantage and disadvantage for those in high-poverty schools. For example, teachers in low-poverty schools are slightly less likely than their peers in high-poverty schools to benefit from a reduced teaching schedule and from having teachers’ aides. Teachers in high-poverty schools are less likely than their peers in low-poverty schools to report having regular supportive communication with the principal and others, being observed and receiving feedback, and participating in seminars or classes for beginning teachers.

Other interventions that effectively facilitate young teachers’ adaptation to the profession, and that foster cooperation and collegiality, include teacher induction and mentoring programs. These programs are designed to support inexperienced teachers in their early years through role modeling, feedback, and support, and to keep those who have strong potential from leaving the school or the profession before they have a chance to master the art of teaching (Sorensen and Ladd 2018; Ingersoll and Strong 2011; Goldhaber, Krieg, and Theobald 2018b). They can thus somewhat reduce the inadequacies described above and improve new teachers’ skills and confidence.

However, our analysis shows that though some of these programs are fairly widely offered, they are not always rated as useful, which suggests room for improvement in their quality and how they are offered. As Table 3 shows, most—though not all—first-year teachers had access to induction programs and to mentors. Overall, 72.7 percent of teachers participated in a teacher induction program, and 79.9 percent were assigned a master or mentor teacher.

Table 3

Figure A provides a more in-depth look at first-year teachers’ experience with mentors. More than half (53.5 percent) of teachers reported having met with the mentor frequently when they were in their first year (top panel). Somewhat surprisingly, this relatively high access to mentors was not accompanied by an equally high share of teachers reporting benefiting from working with mentors (bottom panel). Only a third of teachers (33.2 percent) said that working with a mentor teacher improved their teaching a lot, whereas a slightly larger share (35.7 percent) said that working with a mentor improved their teaching only a little or not at all. In essence, the data in Table 3 and Figure A indicate that while mentoring programs are extensively available, almost half of teachers in such programs only rarely or occasionally work with their mentors, and over a third felt that working with their mentors helped them only a little or not at all.

Figure A

Going back to Table 3, we see that teachers in low-poverty schools are about 5 percentage points more likely to have had access to induction programs and mentor programs than were teachers in high-poverty schools. The disparity across types of school is consistent with data showing that high-poverty schools have higher shares of novice teachers (and lower shares of experienced teachers; see García and Weiss 2019a), so there are simply relatively fewer veteran teachers available to serve as effective mentors.13 This disparity in access across low- and high-poverty schools is amplified when we explore the characteristics of interactions between mentors and the teachers they are mentoring, as shown in Figure A. Although teachers in high-poverty schools are just as likely to work with mentors at least once a week, they are more likely to work with them only rarely (top panel). Also, they are less likely than their peers in low-poverty schools to find these mentoring relationships very effective (bottom panel): Only 32.1 percent of teachers receiving mentoring in high-poverty schools felt that working with mentors improved their teaching a lot, compared with 34.5 percent in low-poverty schools. And 37.3 percent of teachers receiving mentoring in high-poverty schools thought that working with their mentors improved their teaching only a little or not at all, compared with 32.3 percent of teachers receiving mentoring in low-poverty schools. The implication that mentoring programs are less helpful in high-poverty schools could be attributable to a number of factors, including fewer highly credentialed teachers being available for mentoring, as well as to the presence of many other stressors that affect teachers’ early readiness.14

Teachers do not receive resources for accessing meaningful professional development activities and do not find those experiences particularly useful

Teaching is a profession in which continuous training is needed to complement the know-how acquired with experience. It is therefore important to examine patterns of professional development supports and activities that teachers receive throughout their careers. The highest-performing systems in the world provide professional development as part of the regular daily and weekly experience of teaching and continuous training, which are “inextricably linked together,” as the authors of Empowered Educators note (Darling-Hammond, Burns, et al. 2017). In the book, they explain that continuous learning is provided through “incentives and infrastructure for [teacher] learning; time and opportunity for collaboration; curriculum development and lesson study; teacher research; teacher-led PD; appraisal and feedback” (Darling-Hammond, Burns, et al. 2017).15

Our analyses of teacher survey data allow us to assess the prevalence or absence of some of these ingredients of a high-performing professional development regime for our public school teachers, as well as whether teachers have the resources to access professional development activities.16

As the tables and figures in this section show, access to some form of professional development is widespread overall, but some supports and resources needed for participation (time and economic support, among others) are not. It is also troubling that, despite generalized participation in professional development, it is not perceived as being very useful to the majority of teachers who accessed it.17

Table 4 shows that large shares of teachers lack important resources needed to access professional development—mainly time and reimbursements. Although four in five teachers have scheduled time in their contracts for professional development, only half (50.9 percent) of teachers have released time from teaching to participate in professional development, less than a third are reimbursed for conferences or workshop fees (28.2 percent) or receive a stipend for activities that take place outside regular work hours (27.3 percent), and only one in 10 teachers (9.4 percent) receives full or partial reimbursement of college tuition.

Table 4

Teachers in high-poverty schools are at a disadvantage when it comes to accessing several of these resources, but have an advantage in some others. We cannot determine why access is not uniform across high- and low-poverty schools and how consequential the gaps may be.18 A smaller share of teachers in high-poverty schools have scheduled time in their contracts for professional development compared with teachers in low-poverty schools (77.7 percent vs. 80.1 percent), though the very small advantage teachers in high-poverty schools have in released time from teaching (51.2 percent versus 50.1 percent in low-poverty schools) could counter that a bit. Larger shares of teachers in high-poverty schools also get a stipend for activities outside of regular hours, have their travel expenses reimbursed, and receive credits toward recertification or advanced certification, but the shares of these teachers who get reimbursement for college tuition, workshops, or conferences are smaller than the shares for their peers in low-poverty schools.

Just as important as resources being available to facilitate professional development are the types of activities that teachers can access to advance their skills. Curriculum development and lesson study, teacher research, teacher-led professional development (i.e., professional development that is more self-directed by teachers and more actively informed and overseen by them), and appraisal and feedback are key components of solid systems of professional supports (Darling-Hammond, Burns, et al. 2017). Researchers note that professional development programs that are effective are content-focused; they support collaboration and job-embedded practice;19 they are of intense and sustained duration; they focus on discrete skill sets; they offer opportunities for feedback and reflection; and they are characterized by active learning and collaboration (Darling-Hammond, Hyler, et al. 2017; Kraft, Blazar, and Hogan 2018; OECD 2019).20

The survey data presented in Table 5 show that while large shares of teachers are participating in some form of professional development, the most prevalent activities—the ones serving the largest shares of teachers—are not always the most effective types of professional development. They tend to be much more passive than the types of professional development described above.

Table 5

The top panel of Table 5 shows the shares of teachers who participated in four standard types of professional development activities. More than nine in 10 teachers attended workshops, conferences, or training sessions, by far the most common category of activity as well as the least effective and least highly regarded (see Kraft, Blazar, and Hogan 2018; ESSA 2015; Bill & Melinda Gates Foundation 2014; Hirsh et al. 2016; Quint 2011).21 In contrast, only between one-fifth and one-fourth of the teachers participated in what these studies argue are the more effective and highly regarded of these four activities: attending university courses related to teaching; presenting at workshops, conferences, or training sessions; and making observational visits to other schools.

The middle panel of Table 5 shows the shares of teachers engaged in any type of professional development that is focused on a specific area. As we note above, professional development that is content-specific is considered more effective than general professional development, and thus the high shares of teachers participating in some sort of content-specific professional development is encouraging. More than four-fifths of teachers have accessed professional development that is specific to the subject or subjects they teach, about two-thirds have participated in activities focused on the use of computers for instruction, and well over half have participated in professional development for reading instruction.

But some critical subject areas or classroom management practices appear to be neglected. Only about four in 10 teachers received instruction in student discipline and classroom management, about a third got training in teaching students with disabilities, and just over a quarter got training in teaching English language learners (ELLs).

Nevertheless, the hours that teachers report spending in content-focused professional development activities (the activities in the middle panel) add up to about 44 hours—more than a full week of work hours—over the course of the school year. And those 44 average hours spent on professional development do not even include the activities presented in the top and bottom panels. (Note that the hours estimate is not provided directly in the survey, but is reported through a categorical variable; we use standard interpolation to calculate the midpoint between the hours-intervals, use that as an estimate of the hours for each activity, and add these individual estimates to get the total hours.)

The activities presented in the table’s bottom panel reflect teachers’ access to teacher-led research and to opportunities for feedback and appraisal, which are important components of teacher training and professional development systems in some of the world’s highest-performing school systems, such as those in Finland, Singapore, Canada (specifically Alberta, Canada), and Shanghai (Darling-Hammond, Burns, et al. 2017). The shares of teachers who have participated in these activities are in general notable. More than four-fifths of teachers have participated in regularly scheduled collaboration with other teachers on issues of instruction (80.8 percent), and two-thirds have been observed or have observed other teachers in their classrooms (67.0 percent). The exception is the more modest, though still substantial, share of teachers who have engaged in research on a topic of interest for them, with less than half of the teachers (45.2 percent) having done so.

A key lesson from Table 5 is that, while professional development is widespread and teachers are devoting time to it, some of the more effective types of professional development, and some of the critical professional development subject areas, are not being accessed by the majority of teachers. In Figure B, we explore this issue in more depth by analyzing data that indicate how useful (or not useful) specific professional development activities are to teachers.

Figure B

Figure B shows that teachers generally are not highly satisfied with their professional development experiences. Across the range of professional development opportunities listed, only 19.5 to 26.8 percent of teachers found any of the activities very useful, with larger shares (29.6 to 38.1 percent) finding the activities either not useful or just somewhat useful. (Note that this information is only available for the professional development activities by subject—the middle panel of Table 5—not for other activities listed in the table.)

Moreover, not all teachers have access to those activities that are reported to be most useful. A larger share of teachers (26.8 percent) say they are very satisfied with professional development received on the subject they teach—but while this activity is available to most teachers (85.1 percent have access to it), about one in seven (14.9 percent) still lack access to such training. Smaller shares (less than 20 percent) of teachers found training very useful when they were trained on how to teach ELL students (26.7 percent of teachers received training in this area) or on student discipline and management in the classroom (42.3 percent of teachers received training in this area).

Now we turn to disparities between teachers in high- and low-poverty schools with respect to accessing different types of professional development offerings (shown in the last three columns in Table 5) and disparities in professional development usefulness ratings (shown in Appendix Table 1). The data here raise further questions about the responsiveness of professional development systems to teachers’—and students’—needs. Relative to their peers in low-poverty schools, teachers in high-poverty schools were more likely to attend workshops, conferences, or training sessions—the professional development category deemed least meaningful, as mentioned above. The findings (in the bottom panel of Table 5) also point to gaps in teachers’ access to other activities that are highly regarded in education, although teachers in high-poverty schools are not always at a disadvantage. For example, while the share of teachers engaging in teacher-led research is close to 7 percentage points lower in high-poverty schools than in low-poverty schools, the share of teachers engaged in peer observation is about 3 percentage points higher in high-poverty schools than in low-poverty schools.

There are also some gaps between teachers in high- and low-poverty schools with regard to how useful they find specific professional development activities. Indeed, we find an inverse relationship between level of participation in a type of activity shown in Table 5 and its utility for teachers (shown in Appendix Table 1), by school type. We identify two areas of concern with regard to high-poverty schools: Teachers in high-poverty schools are participating more often in professional development activities that they find less useful, and less often in activities that they find more useful, relative to teachers in low-poverty schools. Specifically, while larger shares of teachers in high-poverty schools participate in subject-specific professional development or in classroom management and discipline programs than teachers in low-poverty schools do, smaller shares find those activities very useful. Conversely, while smaller shares of teachers in high-poverty schools participate in computer-for-instruction programs, larger shares find such programs useful.22

While these differences in the shares of teachers finding these activities very useful are small, it is important to note that even small differences tend to become cumulatively large problems in high-poverty schools, as the challenges compound one another (García and Weiss 2019a, 2019b, 2019c, 2019d). The lower levels of preparation and first-year support for new teachers in high-poverty schools described in this report, as well as the gaps in meaningful professional development opportunities, can not only widen the gaps in the qualifications and credentials of the teaching workforce in high- versus low-poverty schools, but may also further demoralize teachers in high-poverty schools and erode their sense of purpose.

Teachers’ lack of support and limited say in school and classroom policies impedes the development of learning communities

As we discuss in our previous report on school climate (García and Weiss 2019d), teacher satisfaction and retention are affected by teachers’ working environments, including their relationships with other teachers and with administrators in their schools. Having nurturing and supportive relationships with colleagues and with administrators, being listened to as professionals, and having a say over the policies of their schools and practices in their classrooms are important components of teachers’ overall satisfaction and sense of purpose. These attributes of a supportive school climate also correlate with their retention in the profession (García and Weiss 2019d; Ladd 2011).

In this report on professional development, we return to these indicators because they also shed important light on how positive learning communities for teachers can support their teaching and career growth. Different sources point out that collegial relationships, opportunities to cooperate and coordinate, and consideration for teachers’ say in school policy and classroom practices are just as important to creating learning communities in schools as formal training and other, more standard forms of professional development (Quint 2011; Warner-Griffin, Cunningham, and Noel 2018; Schwartz 2019; Ingersoll and Collins 2018; OECD 2016, 2019).23 As shown in Tables 6 and 7 (reproduced from Tables 3 and 4 in García and Weiss 2019d), relationship-related indicators of a learning community are far from universal. The text below borrows heavily from our language in García and Weiss 2019d, but discusses what the findings mean for teacher professionalism and for building learning communities in our nation’s schools.

Table 6 shows that, across the board, there is a troubling lack of support for teachers from administrators and colleagues. This means that schools are not providing teachers with strong learning communities characterized by solid administrative supports and leadership, time for peer collaboration, and a shared sense of purpose among school staff.

Table 6

In six of the seven categories reviewed in the table, less than half of the teachers report feeling fully supported by the school administration, their colleagues, or the community in general. The one exception is a proxy for leadership: whether “the principal knows what kind of school he or she wants and has communicated it to the staff.” Just over half (51.6 percent) of teachers surveyed said that their principals exhibit that attribute. And about half (49.6 percent) of teachers report that they see “supportive and encouraging behavior” by school administrators (a proxy for a positive working environment set by the administration). Slightly less than half (47.9 percent), however, strongly agree with the statement, “I make a conscious effort to coordinate the content of my courses with that of other teachers” (a proxy for the community environment created by teachers to facilitate coordination). Only slightly more than a third strongly agree that “there is a great deal of cooperative effort among the staff members” (38.4 percent) or that their colleagues share their views of what the school’s mission should be (36.0 percent). Fewer than one in three teachers affirm that they are recognized for a job well done (32.4 percent), and only 13.3 percent of teachers, or about one in 10, affirm that they receive a great deal of support from the parents of their students for the work they do. Put another way, the survey responses indicate that high shares of teachers experience some level of conflict or disagreement in their schools.24

Table 7 shows that there is also a troubling lack of consideration for teachers’ say in school policy and in their classrooms, which impedes efforts to build strong learning communities (not to mention that it demonstrates disrespect for teachers’ professional knowledge and judgment). Schools are missing out when it comes to learning and benefiting from the contributions of teachers, including when determining the content of in-service professional development programs. As the table shows, meager shares of teachers report having a great deal of influence or control over school policy, suggesting a lack of control over key aspects of their working environments. (And, as noted in our May 2019 report, this generalized disrespect for teachers’ knowledge of their jobs and professional judgment also hurts morale and satisfaction and even affects teachers’ plans as to whether they will or will not stay in teaching indefinitely.) Just 11.1 percent of teachers have a great deal of influence determining the content of in-service professional development programs; this is quite troubling, given national and international surveys and testimonies showing that teachers want to play a more direct role in selecting the types and content of professional development opportunities offered to them (see Bill & Melinda Gates Foundation 2014; Loewus 2019; OECD 2019; Kirk 2019; Schwartz 2019). Only a tiny share (3.2 percent) of teachers report having a great deal of influence over how they are evaluated. The other school policy categories with shares under 10 percent are “setting discipline policy” and “hiring new teachers.” The category with the highest share of teachers reporting a great deal of influence is “establishing curriculum,” but even that is true for just one in five teachers (20.4 percent).

Table 7

Although teachers report much more influence in their classrooms than on school policies, they still indicate a surprisingly small level of control over their daily activities. This indicates that they see little room for contributing to and self-guiding their professional growth and exercising their own judgment and expertise. The shares of teachers who report a great deal of influence or control range from 60 to 70 percent for the most basic actions, such as evaluating and grading students or assigning the amount of homework, but fall to much lower sub-30-percent shares when the actions involve selecting textbooks and other instructional materials and controlling topics and skills to be taught. To put it another way, a large majority of teachers lack authority with respect to how they teach and how their classrooms operate.25

Strengthening professional development and the culture of learning could help attract and retain teachers

Proper professional development should not be seen as an extraneous luxury good in education. As mentioned previously, it is the norm in other countries, where it is embedded as part of the regular daily and weekly experience of teaching and learning (Darling-Hammond, Burns, et al. 2017). And in the United States it has gained growing visibility since the passage of ESSA in 2015. Proper professional development not only validates teachers’ professional standing and strengthens the teaching workforce, but it also correlates with teacher retention and thus could contribute to ameliorating the national teacher shortage. In our two most recent reports (García and Weiss 2019c, 2019d), we argue that low salaries and difficult school climates make teaching less attractive for both potential teachers and highly credentialed teachers already in the profession; here, we likewise argue that lack of early career supports and lack of meaningful professional development opportunities diminish the attractiveness of the teaching profession. Conversely, early career supports, meaningful professional development opportunities, and a supportive climate and culture of learning can help mitigate some of the factors that make it harder for many schools to attract and retain teachers and to strengthen the quality of the teaching workforce overall.

Figure C lists a subset of the early career, continuous training, and influence indicators we have examined so far and shows the shares of “staying” and “quitting” teachers who reported, in their responses to the 2011–2012 Schools and Staffing Survey, that they received the support or experienced the indicator. “Staying” teachers are those who, in the 2012–2013 Teacher Follow-Up Survey, were still at the same school, while “quitting” teachers are those who had left the school and were not in the teaching profession at the time of the follow-up survey. (Those teachers who left to teach at another school are not included in the figure.) Across the board, larger shares of teachers who stayed in teaching had reported the year before that they felt well prepared, received early supports, had more useful professional development opportunities, worked in more cooperative environments, and felt they had more influence over the school and in their classrooms. More than three-quarters of teachers who stayed at their school had participated in teacher mentoring programs, versus just over two-thirds among teachers who quit. Larger shares of staying teachers reported that the professional development that was specific to their subject of main assignment was useful. And relative to quitting teachers, larger shares of teachers who stayed felt that they had had real influence over policy or classroom decisions or worked in cooperative environments.

Figure C

Conclusion

In a knowledge-based profession such as teaching, continuous learning—through professional development, career ladder systems that enable teachers to progress in their profession, and collaborative relationships—is critical for novice and experienced teachers alike. Continuous learning helps teachers keep up with advances in research on effective teaching and learning and with the changing demands of the profession. In addition, early supports and continuous training can make teaching a more attractive occupation, and thus help maintain a stable workforce of highly credentialed teachers. Finally, continuous learning helps professionalize teaching, enhancing respect for the profession.

This report portrays a mosaic of indicators that describe teachers’ continuous training and the degree to which their schools function as learning communities. Our analyses uncover a few areas where opportunities and supports are strong and can be expanded on; but we also uncover other, more numerous areas where there is substantial room to improve the professional supports offered. With regard to the latter, schools should work to ensure that more teachers can access the types of training and development that they find most helpful and most effective, and allow teachers to exercise their judgement and autonomy. As we note at the beginning of the report, there is no established benchmark for what a universal, optimal set of professional development activities looks like. We are thus understandably in gray areas in terms of how continuous learning indicators act as stressors or facilitators that affect the teacher labor market. Yet, even with the limitations of our analyses, it is clear that these supports play a potentially important role in teacher recruitment and retention.

For one, the fairly broad access to certain types of professional development—workshops and conferences or training sessions, activities focused on the subjects that teachers teach, and, for novice teachers, the opportunity to work with a mentor in the first year of teaching—suggests that there already is, in most schools or for most teachers, a foundation for providing professional development that could be expanded to incorporate other types of opportunities. However, neither access to resources for professional development (such as reimbursement or released time for teaching) nor participation in other reportedly more useful forms of early and continuous training (such as teachers leading training sessions or participating in observational visits to other schools) are nearly as widespread. Teachers’ reported satisfaction with the training and professional development they are offered is also limited, suggesting that there may be issues with both the quality and quantity, broadly speaking, of the options offered. The low shares of teachers participating in the types of professional development shown by research to be most effective, and the low shares of teachers reporting that the activities they do access are “very useful,” suggest that a significant portion of what is being offered is suboptimal. The high shares of teachers reporting that they have less than a great deal of influence over what they teach in the classroom (71.3 percent), and over what instructional materials they use (74.5 percent), show a clear need to amplify teachers’ say in their schools.

Put another way, there is an opportunity to greatly improve our understanding of the use of early supports and continuous training opportunities and to make sure these are helping our teachers do their jobs well, feel more valued, and perceive possibilities to advance in their careers. There is also room to further professionalize teaching by giving teachers a greater say in their day-to-day actions and over the policies and rules in place in their schools. In short, there is an opportunity to effect a total shift toward establishing a real culture of learning in our schools that seems ripe for exploration.

Given the associations between optimal professional development (early career supports, ongoing professional development, and relationships that capitalize on teachers’ professional judgment) and teacher retention and recruitment, efforts to establish a system of supports and a real learning community would also help address the teacher shortage. To ensure that both early supports and ongoing professional development fulfill their intended missions, they need to be adequate, sustained, and meaningful to teachers. In order to improve the system’s quality as a whole and elevate the teaching profession, it is also essential that we improve these conditions across the board, so that the needs of teachers in high-poverty schools are not overlooked. As suggested in our companion pieces in this series, only if policymakers think holistically about how to address the teacher shortage will they find the necessary resources to adequately fund our schools, to eliminate the barriers to teaching and learning, and to elevate the level of respect for teachers’ knowledge, experience, and judgment.

About the authors

Emma García is an education economist at the Economic Policy Institute, where she specializes in the economics of education and education policy. García’s research focuses on the production of education (cognitive and noncognitive skills); evaluation of educational interventions (early childhood, K–12, and higher education); equity; returns to education; teacher labor markets; and cost-effectiveness and cost–benefit analysis in education. She has held research positions at the Center for Benefit-Cost Studies of Education, the Campaign for Educational Equity, the National Center for the Study of Privatization in Education, and the Community College Research Center; she has consulted for MDRC, the World Bank, the Inter-American Development Bank, and the National Institute for Early Education Research; and she has served as an adjunct faculty member at the McCourt School of Public Policy, Georgetown University. García received her Ph.D. in economics and education from Columbia University Teachers College.

Elaine Weiss is the lead policy analyst for income security at the National Academy of Social Insurance, where she spearheads projects on Social Security, unemployment insurance, and workers’ compensation. Prior to her work at the academy, Weiss was the national coordinator for the Broader, Bolder Approach to Education (BBA), a campaign launched by the Economic Policy Institute, from 2011–2017. BBA promoted a comprehensive, evidence-based set of policies to allow all children to thrive in school and life. Weiss has authored and co-authored EPI and BBA reports on early achievement gaps and the flaws in market-oriented education reforms. She is co-author, with former Massachusetts Secretary of Education Paul Reville, of Broader, Bolder, Better, published by Harvard Education Press in 2019. Weiss came to BBA from the Pew Charitable Trusts, where she served as project manager for Pew’s Partnership for America’s Economic Success campaign. She has a Ph.D. in Public Policy from the George Washington University Trachtenberg School and a J.D. from Harvard Law School.

Acknowledgments

The authors are grateful to Lora Engdahl for her edits and suggested additions to this piece and for her extraordinary contributions to structuring the contents of this series of papers. We also appreciate John Schmitt’s supervision of this project and Lawrence Mishel’s guidance in earlier stages of this research. We acknowledge Julia Wolfe for her assistance with the tables and figures in this report, Kayla Blado for her work disseminating the report and her assistance with the media, Krista Faries for her work proofreading the text and wordsmithing sentences in several sections, John Carlo Mandapat for the infographics that accompany this report, and the rest of the communications staff at EPI for their contributions to the different components of this report and the teacher shortage series. We appreciate EPI Communications Director Pedro da Costa’s coordination of all the steps required for the publication of this report and of the series.

Data sources used in this report

The analyses presented in this report rely mainly on the 2011–2012 Schools and Staffing Survey (SASS), the 2012–2013 Teacher Follow-Up Survey (TFS), and the 2015–2016 National Teacher and Principal Survey (NTPS). The surveys collect data on and from teachers, principals, and schools in the 50 states and the District of Columbia.26 All three surveys were conducted by the U.S. Census Bureau for the U.S. Department of Education. The survey results are housed at the National Center for Education Statistics (NCES), which is part of the Department of Education’s Institute of Education Sciences (IES).

The NTPS is the redesigned SASS, with a focus on “flexibility, timeliness, and integration with other Department of Education data” (NCES 2019). Both the NTPS and SASS include very detailed questionnaires at the teacher level, school level, and principal level, and the SASS also includes very detailed questionnaires at the school district level (NCES 2017). The TFS survey, which is the source of data on teachers who stay or quit, was conducted a year after the 2011–2012 SASS survey to collect information on the employment and teaching status, plans, and opinions of teachers in the SASS. Following the first administration of the NTPS, no follow-up study was done, preventing us from conducting an updated analysis of teachers by teaching status the year after the NTPS. NCES plans to conduct a TFS again in the 2020–2021 school year, following the 2019–2020 NTPS.

The 2015–2016 NTPS includes public and charter schools only, while the SASS and TFS include all schools (public, private, and charter schools).27 We restrict our analyses to public noncharter schools and to teachers in public noncharter schools.

Appendix Table 1

Appendix Table 2

Endnotes

1. For a more detailed review of media coverage on the shortage, see García and Weiss 2019a. Research on costs comes from Carver-Thomas and Darling-Hammond (2017) and the Learning Policy Institute (2017), who estimate that filling a vacancy costs $21,000 on average, and from Carroll (2007), who estimates the total annual cost of turnover at $7.3 billion per year. According to Strauss (2017), that estimated annual cost of turnover would exceed $8 billion at present.

2. A lack of sufficient, qualified teachers threatens students’ ability to learn (Darling-Hammond 1999; Ladd and Sorensen 2016). Instability in a school’s teacher workforce (i.e., high turnover and/or high attrition) negatively affects student achievement and diminishes teacher effectiveness and quality (Ronfeldt et al. 2013; Jackson and Bruegmann 2009; Kraft and Papay 2014; Sorensen and Ladd 2018). And high teacher turnover consumes economic resources (i.e., through costs of recruiting and training new teachers) that could be better deployed elsewhere.

3. The two main surveys we reviewed for this report did not ask teachers directly why they need professional development, and we are unaware of other surveys that include such a question.

4. ESSA 2015 provides opportunities for professional development under all titles in the Act and focuses on how to allow teachers to grow (Education Week 2018). Under the term “professional development,” ESSA includes the following: activities that “are an integral part of school and local educational agency strategies for providing educators with the knowledge and skills necessary to enable students to succeed in a well-rounded education and to meet the challenging state academic standards”; activities that are “sustained (i.e., not stand-alone, 1-day, or short term workshops), intensive, collaborative, job-embedded, data-driven, and classroom-focused”; and activities that “improve…teachers’ knowledge of the subjects they teach, their understanding of how students learn…; or are aligned with…academic goals of the school or local education agency” (ESSA 2015; Learning Forward 2019; Education Week 2018). ESSA expands the reach of professional development activities to encompass activities offered to other educators who work with students—including principals and paraprofessionals—and suggests building systems of professional learning (educator development, retention, and advancement) (Hirsh et al. 2016).

5. Our analyses of professional development activities and resources rely on the 2011–2012 Schools and Staffing Survey (SASS). Our analyses of early career supports, the influence and autonomy teachers have, and the culture of learning rely on the 2015–2016 National Teacher and Principal Survey (NTPS). Note that some figures in this report also appear in our May 2019 report on working environments in schools (García and Weiss 2019d). The professional development module that delivered data for the 2011–2012 SASS is rotating and was not included in the most recent data set available when we were conducting our study (2015–2016), but will be in the next cycle, 2017–2018, as noted in the questionnaire here: https://nces.ed.gov/surveys/ntps/question1718.asp. Most questions in this module have been modified, which will prevent comparative analyses over time in any event. We also remind readers that NTPS 2017–2018 will not include the Teacher Follow-Up Survey (TFS), which means it will not be possible to examine the correlation between professional development opportunities and retention with the next release of this study.

6. Other sources confirm widespread access to some sort of professional development for teachers in the U.S. (Bill & Melinda Gates Foundation 2014) and internationally (OECD 2019).

7. See Appendix Table 2 for shares of teachers receiving additional compensation from working for the school district.

8. Croft, Guffy, and Vitale (2018) show that when a sample of students taking the ACT were asked to say why they were not interested in teaching, the lack of opportunities for career development was the second most cited reason, behind low salary. It is reasonable to assume that the presence of such opportunities would play some role in attracting more students into teaching.

9. See Darling-Hammond 1999; Ladd and Sorensen 2016; Ronfeldt et al. 2013; Jackson and Bruegmann 2009; Kraft and Papay 2014; Moore-Johnson, Kraft, and Papay 2012; Ladd 2011; Loeb, Darling-Hammond, and Luczak 2005; and Warner-Griffin, Cunningham, and Noel 2018.

10. In addition to preparation, induction programs are critical to making sure that teachers are ready to teach from the start of their careers (Darling-Hammond, Burns, et al. 2017; NCEE 2016; Ingersoll and Collins 2018). (An induction program is defined in the 2011–2012 Schools and Staffing Survey (SASS) questionnaire as “a program for beginning teachers that may include teacher orientation, mentoring, coaching, demonstrations, and/or assessments aimed at enhancing teacher effectiveness.” An induction program is defined in the 2015–2016 NTPS as “a program for beginning teachers aimed to enhance teachers’ effectiveness by providing systematic support.”) Kraft, Blazar, and Hogan (2018) conducted a meta-analysis to show the significant effect of teacher coaching on both instruction and student performance (pooled effect sizes were 0.60 of a standard deviation (SD) and 0.18 SD, respectively). (Coaching programs include “all in-service PD programs where coaches or peers observe teachers’ instruction and provide feedback to help them improve,” according to the authors.) Other evaluations have demonstrated a positive influence of mentoring programs on both the teachers receiving mentoring and the mentors, as measured by the performance of their students, especially later in teachers’ careers and especially in math (Goldhaber, Krieg, and Theobald 2018a, 2018b; Papay et al. 2016).

11. In keeping with the patterns identified in the previous reports in this series, teachers’ perceived lack of preparedness is greater in high-poverty schools than in low-poverty schools. While our main focus in this report is to document patterns of career-building supports and gaps in them, it is important to acknowledge that, if anything, we need stronger, not weaker, supports for early-career teachers in high-poverty schools. That is because high-poverty schools already suffer from lower shares of highly credentialed teachers (García and Weiss 2019a), and thus teachers in high-poverty schools are especially in need of these early supports. Similarly, high-poverty schools also have higher shares of teachers who came into the profession through alternative certification programs (19 percent of teachers in high-poverty schools entered teaching this way, versus just over 13 percent in low-poverty schools). While research says that the route into teaching is not consistently associated with any significant differences in teacher effectiveness, our data show that, in practice, teachers who entered teaching from alternative programs feel less than “very well” prepared to do well in class. But we cannot disentangle how much of that difference is due to effective versus ineffective preparation, how much is due to concentration of these teachers in low-poverty schools, and how much is due to nongeneralized access to supports or to the quality of those supports: As mentioned in the text explaining Table 2, six of the nine types of supports that could be offered to teachers are being provided to at least 60 percent of teachers in high-poverty schools, so access is relatively broad—although not universal—in those schools. Therefore, we offer this information as complementary evidence that there is a need for strong preparation and early supports to ensure strong preparation of novice teachers, and that professional preparation and early career supports are worthy topics for future research.

12. Preparation received in teacher preparation programs and standards for becoming a teacher obviously play an important role in whether teachers are ready to teach in their first year, but teacher preparation programs and certification standards are beyond the scope of this report. With regard to teacher preparation programs, see note 11. With regard to certification standards, earlier reports in this series identify certain trends in our exploration of Title II data from the U.S. Department of Education (2017a, 2017b) that could negatively affect teacher qualifications at the beginning of their careers. For example, we find that the number of states requiring content-specific bachelor’s degrees for initial teaching credentials decreased between 2008–2009 and 2015–2016. Examining the requirements across all initial certificates available nationwide, we also note a large decrease in the share of initial teaching certificates requiring a content-specific bachelor’s degree for middle school, which fell from 38.6 percent of all initial certificates in 2008–2009 to 22.8 percent of all initial certificates in 2015–2016, a 15.8 percentage-point decrease. Over the same period there were also drops in the share of initial certificates requiring performance assessments (down 16.2 percentage points), supervised clinical experience (down 10.8 percentage points), or a police record examination (down 17.2 percentage points). However, there was an increase in the share of initial certificates requiring “prescribed coursework” (up 10.8 percentage points) (see García and Weiss 2019b).

13. Indeed, the ratio of mentors to teachers was 1.01 in low-poverty schools—i.e., there were just enough mentors to have one per teacher—versus just 0.86 in high-poverty schools—i.e., there were more teachers who needed a mentor than there were mentors available.

14. As mentioned earlier, the share of novice teachers who felt they were very well prepared to teach in their classrooms was lower in high-poverty schools than in low-poverty schools (Table 1). Earlier reports in this series discuss the other stressors heightened in high-poverty schools, which we summarize here. On average, the credentials of teachers in high-poverty schools—overall and for staying teachers (versus those who quit)—are worse than in low-poverty schools. “Lower credentials” here means higher shares of inexperienced teachers, teachers who entered into teaching via alternative routes and who are not fully certified or do not have a background in the subject of main assignment (García and Weiss 2019a). There is a larger churn rate and therefore more staff instability in high-poverty schools (García and Weiss 2019b). Teacher pay is lower in high-poverty schools, on average, and teachers in high-poverty schools not only receive a smaller amount of income from moonlighting (if they moonlight), but also the moonlighting that do is less likely to involve paid extracurricular or additional activities for the school system that would not only generate extra pay but also help them grow professionally (García and Weiss 2019c; see Appendix Figure B). Finally, school climates are more challenging in high-poverty schools, where, relative to low-poverty schools, larger shares of teachers report facing barriers to teaching, experiencing threats to physical and mental safety, being dissatisfied, and not planning to stay in teaching indefinitely (García and Weiss 2019d).

15. See also Learning Forward 2019; ESSA 2015; Bill & Melinda Gates Foundation 2014; Kraft, Blazar, and Hogan 2018; and Darling-Hammond, Hyler, et al. 2017.

16. Unfortunately, we cannot assess when professional development is offered and whether all teachers can access it, nor what assignment mechanisms to professional development are in place. Information on time and effectiveness is also limited.

17. Although U.S. teachers share this broad access to professional development with their rich-country peers, U.S. teachers may be less satisfied than their peers with the usefulness of these activities. New data from the Teaching and Learning International Survey (TALIS) administered by the Organization for Economic Cooperation and Development in 48 countries and economies show that more than 90 percent of teachers and principals in OECD countries attended “at least one professional development activity” in the year prior to the survey and 82 percent of teachers reported that the training had a positive impact on their teaching practice (OECD 2019).

18. For instance, in some cases, resources to access professional development may be limited due to insufficient funding or for other reasons. There is also not sufficient information in our data to discern how consequential the gaps across schools may be, but the important point is to emphasize that there is a lack of these important resources for accessing professional development, and more so in high-poverty schools (with the exception of scheduled time in the contract, which is provided for a majority of teachers).

19. According to Croft et al. 2010, job-embedded professional development “refers to teacher learning that is grounded in day-to-day teaching practice and is designed to enhance teachers’ content-specific instructional practices with the intent of improving student learning” (Croft et al. 2010, citing work by Darling-Hammond and McLaughlin 1995 and Hirsh 2009).

20. The research by Kraft, Blazar, and Hogan (2018) to identify these features of effective professional development builds on Darling-Hammond et al. 2009; Desimone 2009; Desimone and Garet 2015; Garet at al. 2001; and Hill 2007 (see p. 548).

21. With regard to workshops, Kraft, Blazar, and Hogan (2018) explain that they “are often viewed as insufficient to address the inherently multifaceted nature of teachers’ practice” (see p. 551, citing Kennedy 2016; Opfer and Pedder 2011; Schachter 2015). Quint (2011) acknowledges that professional development is criticized for being offered in the form of a “one-shot” (low-cost) lecture or workshop, rather than in the forms that teachers prefer, including “intensive summer institutes, follow-up group sessions, and coaching of individual teachers.”

22. The exception to this trend is that larger shares of teachers in high-poverty schools participated in ELL teacher training than in low-poverty schools (33.5 percent vs. 19.3 percent), and larger shares of teachers in high-poverty schools also found such training “very useful” compared with teachers in low-poverty schools (20.9 percent vs. 19.1 percent). This differential may be partly attributable to the fact that teachers in high-poverty schools serve larger shares of students in ELL programs.

23. As some of these sources argue, teachers’ relationships and influence are important components of teacher professionalism. For example, Ingersoll and Collins (2018) assess whether teaching meets the model attributes of a professional career—including workplace authority and high prestige, among other attributes. The OECD reports on the Teaching and Learning International Survey (TALIS) (OECD 2016, 2019) argue that there are several pillars in teaching professionalism, including perceived prestige, career opportunities, collaborative culture, and level of professional autonomy and responsibility. Other sources discuss teachers’ relationships and influence on a practical level, including as components of a learning community for teachers. For example, Quint (2011) emphasizes the idea of “a broader conception of teaching learning that involves all teachers in a school in a professional learning community that is engaged in a continuous and collegial cycle of learning, practice, reflection, and improvement.”

24. As discussed in García and Weiss 2019d, indicators of a poor learning community are more pronounced in high-poverty schools. A larger share of teachers in high-poverty schools indicate some level of conflict or disagreement in attitudes or beliefs with the administration or colleagues than do teachers in low-poverty schools. By far the biggest gap between high- and low-poverty schools is in support teachers receive from their students’ parents: Nine out of 10 teachers in high-poverty schools do not feel fully supported by parents for the work they do compared with a still-very-high eight in 10 teachers in low-poverty schools. See research on comprehensive supports that include engaging parents in Weiss and Reville 2019.

25. The usual gaps between shares of teachers reporting positive working conditions in high- and low-poverty schools also appear here. In the areas of school policy, the gaps between teachers’ autonomy or influence in high- and low-poverty schools are small, in general (under 2 percentage points for all categories except “establishing curriculum”), and in fact a slightly greater share of teachers in high-poverty schools report having a great deal of control over setting discipline policy and evaluating teachers. In terms of autonomy in their classrooms: The bottom half of the table shows that in all tasks listed except assigning homework, teachers in high-poverty schools have less of a say than their counterparts in low-poverty schools and that the gaps range from 2.2 to 4.3 percentage points.

26. The 2015–2016 NTPS does not produce state-representative estimates. The forthcoming 2017–2018 NTPS will support state-level estimates.

27. The forthcoming 2017–2018 NTPS will include private schools.

References

Bill & Melinda Gates Foundation. 2014. Teachers Know Best: Teachers’ Views on Professional Development. December 2014.

Carroll, Thomas G. 2007. Policy Brief: The High Cost of Teacher Turnover. National Commission on Teaching and America’s Future.

Carver-Thomas, Desiree, and Linda Darling-Hammond. 2017. Teacher Turnover: Why It Matters and What We Can Do About It. Learning Policy Institute, August 2017.

Croft, Michelle, Gretchen Guffy, and Dan Vitale. 2018. Encouraging More High School Students to Consider Teaching. ACT Policy Research, June 2018.

Croft, Andrew, Jane G. Coggshall, Megan Dolan, and Elizabeth Powers (with Joellen Killion). 2010. Job-Embedded Professional Development: What It Is, Who Is Responsible, and How to Get It Done Well. National Comprehensive Center for Teacher Quality, Mid-Atlantic Comprehensive Center, and National Staff Development Council.

Darling-Hammond, Linda. 1999. Teacher Quality and Student Achievement: A Review of State Policy Evidence. Center for the Study of Teaching and Policy, University of Washington, Seattle.

Darling-Hammond, Linda, Dion Burns, Carol Campbell, A. Lin Goodwin, Karen Hammerness, Ee Ling Low, Ann McIntyre, Mistilina Sato, and Kenneth Zeichner. 2017. Empowered Educators: How High-Performing Systems Shape Teaching Quality Around the World. San Francisco: Jossey-Bass.

Darling-Hammond, Linda, Maria E. Hyler, and Madelyn Gardner, with assistance from Danny Espinoza. 2017. Effective Teacher Professional Development. Learning Policy Institute, May 2017.

Darling-Hammond, Linda, and Milbrey W. McLaughlin. 1995. “Policies That Support Professional Development in an Era of Reform.” Phi Delta Kappan 76, no. 8: 597–604.

Darling-Hammond, Linda, Ruth Chung Wei, Alethea Andree, Nikole Richardson, and Stelios Orphanos. 2009. Professional Learning in the Learning Profession: A Status Report on Teacher Development in the United States and Abroad. National Staff Development Council and School Redesign Network at Stanford University, February 2009.

Desimone, Laura M. 2009. “Improving Impact Studies of Teachers’ Professional Development: Toward Better Conceptualizations and Measures.” Educational Researcher 38, no. 3: 181–199. https://doi.org/10.3102/0013189X08331140.

Desimone, Laura M., and Michael S. Garet. 2015. “Best Practices in Teachers’ Professional Development in the United States.” Psychology, Society and Education 7, no. 3: 252–263.

Dias-Lacy, Samantha L., and Ruth V. Guirguis. 2017. “Challenges for New Teachers and Ways of Coping with Them.” Journal of Education and Learning 6, no. 3. https://doi.org/10.5539/jel.v6n3p265.

Education Week. 2018. “How ESSA Affects You: Shifting Focus to Support Today’s Educators” (Expert Presenters: Francie Alexander and Sue Gendron)” (webinar).

ESSA. 2015. Every Student Succeeds Act of 2015, Pub. L. No. 114-95 § 114 Stat. 1177 (2015–2016).

García, Emma, and Elaine Weiss. 2019a. The Teacher Shortage Is Real, Large and Growing, and Worse Than We Thought: The First Report in ‘The Perfect Storm in the Teacher Labor Market’ Series. Economic Policy Institute, March 2019.

García, Emma, and Elaine Weiss. 2019b. U.S. Schools Struggle to Hire and Retain Teachers: The Second Report in ‘The Perfect Storm in the Teacher Labor Market’ Series. Economic Policy Institute, April 2019.

García, Emma, and Elaine Weiss. 2019c. Low Relative Pay and High Incidence of Moonlighting Play a Role in the Teacher Shortage, Particularly in High-Poverty Schools: The Third Report in ‘The Perfect Storm in the Teacher Labor Market’ Series. Economic Policy Institute, May 2019.

García, Emma, and Elaine Weiss. 2019d. Challenging Working Environments (‘School Climates’), Especially in High-Poverty Schools, Play a Role in the Teacher Shortage: The Fourth Report in ‘The Perfect Storm in the Teacher Labor Market’ Series. Economic Policy Institute, May 2019.

Garet, Michael S., Andrew C. Porter, Laura Desimone, Beatrice F. Birman, and Kwang Suk Yoon. 2001. “What Makes Professional Development Effective? Results from a National Sample of Teachers.” American Educational Research Journal 38, no. 4: 915–945. https://doi.org/10.3102/00028312038004915.

Goldhaber, Dan, John Krieg, and Roddy Theobald. 2018a. “Exploring the Impact of Student Teaching Apprenticeships on Student Achievement and Mentor Teachers.” National Center for Analysis of Longitudinal Data in Education Research (CALDER) Working Paper no. 207-1118-1, November 2018.

Goldhaber, Dan, John Krieg, and Roddy Theobald. 2018b. “Effective Like Me? Does Having a More Productive Mentor Improve the Productivity of Mentees?” National Center for Analysis of Longitudinal Data in Education Research (CALDER) Working Paper no. 208-1118-1, November 2018.

Hill, Heather C. 2007. “Learning in the Teacher Workforce.” Future of Children 17, no. 1: 111–127. https://doi.org/10.1353/foc.2007.0004.

Hill, Heather C. 2009. “Fixing Teacher Professional Development.” Phi Delta Kappan 90, no. 7: 470–477.

Hirsh, Stephanie. 2009. “A New Definition.” Journal of Staff Development 30, no. 4: 10–16.

Hirsh, Stephanie, et al. 2016. “Re: Proposed Regulations for Consolidated Plans Under the Every Student Succeeds Act (ESSA).” Comments submitted on behalf of Learning Forward to U.S. Department of Education Secretary John King, July 29, 2016.

Ingersoll, Richard M. 2004. “Revolving Doors and Leaky Buckets.” In Letters to the Next President: What We Can Do About the Real Crisis in Public Education, edited by Carl D. Glickman, 141–150. New York: Teachers College Press.

Ingersoll, Richard M. 2014. “Why Do High-Poverty Schools Have Difficulty Staffing Their Classrooms with Qualified Teachers?” Presentation for panel discussion How Do We Get Experienced, Accomplished Teachers into High-Need Schools?, Albert Shanker Institute, October 8, 2014.

Ingersoll, Richard M., and Gregory J. Collins. 2018. “The Status of Teaching as a Profession.” In Schools and Society: A Sociological Approach to Education, 6th ed., edited by Jeanne H. Ballantine, Joan Z. Spade, and Jenny M. Stuber, 199–213. Los Angeles: SAGE.

Ingersoll, Richard M., and Michael Strong. 2011. “The Impact of Induction and Mentoring for Beginning Teachers: A Critical Review of the Research.” Review of Educational Research 81, no. 2: 201–233. https://doi.org/10.3102/0034654311403323.

Jackson, Kirabo, and Elias Bruegmann. 2009. “Teaching Students and Teaching Each Other: The Importance of Peer Learning for Teachers.” American Economic Journal: Applied Economics 1, no. 4: 85–108.

Jensen, Ben, Julie Sonnemann, Katie Roberts-Hull, and Amélie Hunter. 2016. Beyond PD: Teacher Professional Learning in High-Performing Systems. National Center on Education and the Economy.

Kennedy, M. Mary. 2016. “How Does Professional Development Improve Teaching?” Review of Educational Research 86, no. 4: 945–980. https://doi.org/10.3102/0034654315626800.

Kirk, Joy. 2019. “Teachers Are Always There to Help, But Now We’re the Ones Who Need a Boost.” Working Economics Blog (Economic Policy Institute), June 14, 2019.

Kraft, Matthew A., and John P. Papay. 2014. “Can Professional Environments in Schools Promote Teacher Development? Explaining Heterogeneity in Returns to Teaching Experience.” Educational Evaluation and Policy Analysis 36, no. 4: 476–500.

Kraft, Matthew A., David Blazar, and Dylan Hogan. 2018. “The Effect of Teacher Coaching on Instruction and Achievement: A Meta-Analysis of the Causal Evidence.” Review of Educational Research 88, no. 4: 547–558. https://doi.org/10.3102/0034654318759268.

Ladd, Helen. 2011. “Teachers’ Perceptions of Their Working Conditions: How Predictive of Planned and Actual Teacher Movement?” Educational Evaluation and Policy Analysis 33, no. 2: 235–261.

Ladd, Helen F., and Lucy C. Sorensen. 2016. “Returns to Teacher Experience: Student Achievement and Motivation in Middle School.” Education Finance and Policy 12, no. 2: 241–279.

Learning Forward. 2019. “Definition of Professional Development” (web page). Accessed June 18, 2019.

Learning Policy Institute. 2017. What’s the Cost of Teacher Turnover? (calculator). September 2017.

Liston, Dan, Jennie Whitcomb, and Hilda Borko. 2006. “Too Little or Too Much: Teacher Preparation and the First Years of Teaching.” Journal of Teacher Education 57, no. 4: 351–358.

Loeb, Susanna, Linda Darling-Hammond, and John Luczak. 2005. “How Teaching Conditions Predict Teacher Turnover in California Schools.” Peabody Journal of Education 80, no. 3: 44–70.

Loewus, Liana. 2019. “A Clearer Vision for Teacher Professional Learning.” Education Week, May 14, 2019.

Mizell, Hayes. 2010. Why Professional Development Matters. Learning Forward.

Moore-Johnson, Susan, Matthew A. Kraft, and John P. Papay. 2012. “How Context Matters in High-Need Schools: The Effects of Teachers’ Working Conditions on Their Professional Satisfaction and Their Students’ Achievement.” Teachers College Record 114, no. 10: 1–39.

National Center on Education and the Economy (NCEE). 2016. Preparing Profession-Ready Teachers. Policy brief from the Empowered Educators Project.

National Center for Education Statistics (NCES) (U.S. Department of Education). 2011–2012. Licensed microdata from the 2011–2012 Schools and Staffing Survey (SASS).

National Center for Education Statistics (NCES) (U.S. Department of Education). 2012–2013. Licensed microdata from the 2012–2013 Teacher Follow-Up Survey (TFS).

National Center for Education Statistics (NCES) (U.S. Department of Education). 2015–2016. Licensed microdata from the 2015–2016 National Teacher and Principal Survey (NTPS).

National Center for Education Statistics (NCES) (U.S. Department of Education). 2017. Documentation for the 2011–12 Schools and Staffing Survey. March 2017.

National Center for Education Statistics (NCES) (U.S. Department of Education). 2019. “NTPS Overview” (web page). Accessed March 2019.

OECD. 2016. Supporting Teacher Professionalism: Insights from TALIS 2013. Paris: TALIS, OECD Publishing. https://dx.doi.org/10.1787/9789264248601-en.

OECD. 2019. TALIS 2018 Results (Volume I): Teachers and School Leaders as Lifelong Learners. Paris: TALIS, OECD Publishing. https://doi.org/10.1787/1d0bc92a-en.

Opfer, V. Darleen, and David Pedder. 2011. “Conceptualizing Teacher Professional Learning.” Review of Educational Research 81, no. 3: 376–407. https://doi.org/10.3102/0034654311413609.

Papay, John P., Eric S. Taylor, John H. Tyler, and Mary Laski. 2016. “Learning Job Skills from Colleagues at Work: Evidence from a Field Experiment Using Teacher Performance Data.” National Bureau of Economic Research Working Paper no. 21986. https://doi.org/10.3386/w21986.

Quint, Janet. 2011. Professional Development for Teachers: What Two Rigorous Studies Tell Us. MDRC.

Robinson, Java. 2019. “Why Professional Development Matters.” NEA Today, February 11, 2019.

Ronfeldt, Matthew, Hamilton Lankford, Susanna Loeb, and James Wyckoff. 2013. “How Teacher Turnover Harms Student Achievement.” American Educational Research Journal 50, no. 1: 4–36.

Schachter, Rachel E. 2015. “An Analytic Study of the Professional Development Research in Early Childhood Education.” Early Education and Development 26, no. 8: 1057–1085. https://doi.org/10.1080/10409289.2015.1009335.

Schwartz, Sarah 2019. “What Do Teachers Really Want from Professional Development? Respect.” Education Week, May 15, 2019.

Smith, Thomas M., and Richard Ingersoll. 2004. “What Are the Effects of Induction and Mentoring on Beginning Teacher Turnover?” American Educational Research Journal 41, no. 3: 681–714. https://doi.org/10.3102/00028312041003681.

Sorensen, Lucy C., and Helen Ladd. 2018. “The Hidden Costs of Teacher Turnover.” National Center for Analysis of Longitudinal Data in Education Research (CALDER) Working Paper no. 203-0918-1, September 2018.

Strauss, Valerie. 2017. “Why It’s a Big Problem That So Many Teachers Quit—and What to Do About It.” Washington Post, November 27, 2017.

Sutcher, Leib, Linda Darling-Hammond, and Desiree Carver-Thomas. 2016. A Coming Crisis in Teaching? Teacher Supply, Demand, and Shortages in the U.S. Learning Policy Institute, September 2016.

U.S. Department of Education. 2017a. “Requirements for an Initial Teaching Credential, by State” [data table]. Data from the Higher Education Act Title II State Report Card (SRC) Reporting System. Spreadsheet downloadable at https://title2.ed.gov/Public/DataTools/NewExcels/Requirements.aspx.

U.S. Department of Education. 2017b. “States Requiring Content Specific Bachelor’s Degrees for All Initial Teaching Credentials” [data table]. Data from the Higher Education Act Title II State Report Card (SRC) Reporting System. Spreadsheet downloadable at https://title2.ed.gov/Public/DataTools/NewExcels/ContentDegrees.aspx.

Warner-Griffin, Catharine, Brittany C. Cunningham, and Amber Noel. 2018. Public School Teacher Autonomy, Satisfaction, Job Security, and Commitment: 1999–2000 and 2011–12. Institute of Education Statistics, National Center for Education Statistics, U.S. Department of Education, Stats in Brief no. 2018-103, March 2018.

Weiss, Elaine, and Paul Reville. 2019. Broader, Bolder, Better: How Schools and Communities Help Students Overcome the Disadvantages of Poverty. Cambridge, Mass.: Harvard Education Press.

This report is the fourth in a series examining the magnitude of the teacher shortage and the working conditions and other factors that contribute to the shortage.

What this series finds: The teacher shortage is real, large and growing, and worse than we thought. When indicators of teacher quality (certification, relevant training, experience, etc.) are taken into account, the shortage is even more acute than currently estimated, with high-poverty schools suffering the most from the shortage of credentialed teachers.

What this report finds: The working environment for teachers—broadly referred to here as “school climate”—is tough. Students are coming to school unprepared to learn (as reported by 27.3 percent of teachers), parents are struggling to be involved (as reported by 21.5 percent of teachers), and other conditions impede teaching. These conditions are largely byproducts of larger societal forces such as rising poverty, segregation, and insufficient public investments. In addition to barriers to teaching, teachers face threats to their safety. More than one in five teachers (21.8 percent) report that they have been threatened and one in eight (12.4 percent) say they have been physically attacked by a student at their current school. Compounding the stress, teachers report a level of conflict with—and lack of support from—administrators and fellow teachers, and little say in their work. More than two-thirds of teachers report that they have less than a great deal of influence over what they teach in the classroom (71.3 percent) and what instructional materials they use (74.5 percent), which suggests low respect for their knowledge and judgment.

Not surprisingly, one in 20 teachers (4.9 percent) say that the stress and disappointments involved in teaching are not worth it. Considerably larger shares of teachers express some level of dissatisfaction with being a teacher in their school (48.7 percent), say they think about leaving teaching at some point (27.4 percent), or are not certain that they would still become teachers if they could go back to college and make a decision again (57.5 percent). (All these data on school climate indicators are for the 2015–2016 school year except for the share of teachers who in 2011–2012 said they are not sure they would become teachers if they could start over again.)

And, paralleling the finding in the series’ previous reports, teachers in high-poverty schools have it worse: relative to their peers in low-poverty schools, larger shares of teachers in high-poverty schools report barriers to teaching, threats to physical safety and attacks, a lack of supportive relationships, and little autonomy in the classroom.

Our data suggest a relationship between tough climates and quitting. When we compare teachers who end up quitting with those who stay, we find that larger shares of quitting teachers had reported, prior to leaving, that they were teaching unprepared students (39.0 percent vs. 29.4 percent), experiencing demoralizing stress (12.5 percent vs. 3.6 percent), lacking strong influence over what they teach in class (74.6 percent vs. 71.4 percent), and not being fully satisfied with teaching at their school (60.5 percent vs. 43.3 percent). Indeed, the share of teachers who felt that the stress and disappointments involved in teaching weren’t really worth it was 3.5 times as large among those who ended up quitting than among those who stayed.

Why it matters: Working environments clearly play a role in the teacher shortage, along with low pay (as shown in our last report) and weak professional development opportunities (as will be shown in our next report). The teacher shortage harms students, teachers, and public education as a whole. In addition, the fact that the shortage is more acute in high-poverty schools challenges the U.S. education system’s goal of providing a sound education equitably to all children.

What we can do about it: Tackle the poor school climate, low relative pay, and other factors that are prompting teachers to quit and dissuading people from entering the teaching profession. With regard to working environments, we need policy interventions and institutional decisions that channel assistance and resources to teachers who press on despite barriers to teaching, stress and physical threats, a lack of support by administrators, little influence over their day-to-day work, and low satisfaction. High-poverty schools and their teachers require extra resources and funding to support students directly and to reduce the teacher shortage.

Introduction

The teacher shortage—the gap between the number of qualified teachers needed and available for hire in a given year—in the nation’s K–12 schools is an increasingly recognized but still poorly understood crisis. The shortage is discussed by the media and policymakers, and researchers have estimated its size (about 110,000 teachers in the 2017–2018 school year, up from no shortage before 2013, according to Sutcher, Darling-Hammond, and Carver-Thomas 2016) and even quantified part of its cost.1 The shortage constitutes a crisis because of its negative effects on students, teachers, and the education system at large. But it is poorly understood because the reasons for it are complex and interdependent. The first report in this series, The Teacher Shortage Is Real, Large and Growing, and Worse Than We Thought (García and Weiss 2019a), established that current national estimates of the teacher shortage likely understate the magnitude of the problem: When issues such as teacher qualifications and the unequal distribution of highly credentialed teachers across high- and low-poverty schools are taken into consideration, the teacher shortage problem is much more severe than previously identified.

The second report in this series, U.S. Schools Struggle to Hire and Retain Teachers (García and Weiss 2019b), built on the research in the first report and used the same quality and equity angles. It showed that schools are having difficulties filling teacher vacancies and are leaving vacancies unfilled despite actively trying to hire teachers to fill them. High-poverty schools are hit hardest: They find it more difficult to fill vacancies than do low-poverty schools and schools overall, and they experience higher turnover and attrition than low-poverty schools. One factor behind staffing difficulties is the high share of public school teachers leaving their posts: 13.8 percent are either leaving their school or leaving teaching altogether, according to most recent data. Another factor is the dwindling pool of applicants to fill vacancies: from the 2008–2009 to 2015–2016 school years, the number of education degrees awarded fell by 15.4 percent and the number of people who completed a teacher preparation program fell by 27.4 percent. Schools are also having a harder time retaining credentialed teachers, as evident in the small but growing share of all teachers who are newly hired and in their first year of teaching (4.7 percent) and in the substantial shares of teachers who quit who are certified and experienced. Retaining credentialed teachers is also more difficult for high-poverty schools.

The third report in the series focused on a likely factor behind why teachers are leaving the profession and fewer people are becoming teachers: teacher pay. Specifically, Low Relative Pay and High Incidence of Moonlighting Play a Role in the Teacher Shortage, Particularly in High-Poverty Schools (García and Weiss 2019c) described how teacher compensation compares with compensation in nonteaching occupations, and called attention to the high share of teachers who supplement their earnings by moonlighting. The report found a correlation between measures of teacher compensation and teachers leaving the profession: specifically, it found that teachers who ended up quitting teaching reported, in the year before they quit, receiving on average, lower salaries; participating less in the kinds of paid extracurricular activities that complement their professional development (activities like coaching students or mentoring teachers); and participating more in working options outside the school system than did teachers who stayed at their schools. In high-poverty schools, teachers face compounded challenges. Relative to teachers in low-poverty schools, teachers in high-poverty schools are paid less, receive a smaller amount from moonlighting, and the moonlighting that they do is less likely to involve paid extracurricular or additional activities for the school system that generate extra pay but also help teachers grow professionally.

This report, the fourth in the series, explores another likely factor behind the exodus of teachers from the profession and the shrinking supply of future teachers: the working environment for teachers, broadly referred to here as the “school climate.” We show that school climate affects teacher satisfaction, morale, and expectations about staying in the profession. We show that school climate is challenging for a number of reasons: Teachers confront widespread barriers to teaching and learning, face threats to their emotional and physical safety, lack influence over school policy and what and how they teach in their classrooms, and suffer from dissatisfaction and low motivation. We also demonstrate that there is a significant relationship between these indicators of difficult working conditions and teachers leaving the profession. And finally, as in previous reports, we provide evidence that working conditions are more challenging in high-poverty schools than in low-poverty schools, which compounds the problems already identified in this series. The findings suggest that efforts to address teacher shortages must be holistic and include initiatives to improve school climates, especially in high-poverty schools where teacher shortages and school climate problems are most serious. The next paper in this series follows up with a discussion about teacher training and supports, which also have the potential to alter the availability of teachers and therefore interact with these other factors to drive shortages.

School climate is an issue across the board and it is implicated in the teacher shortage

The environment in which an employee works has a major impact on not just job satisfaction but also on the ability to do the job well and the desire to continue to remain in the job and the profession. This is certainly true for teachers, who spend much of their time interacting with students, fellow teachers, and other school staff and thus are immersed in their workplace climate to a high degree. Teachers in the vast majority of contexts are prepared—able and trained—to deal with the challenges of their vocation. However, there are certain challenges related to the working environment that teachers should not have to deal with or that they are ill-equipped to handle and still do their jobs well.

This report addresses the challenges that arise from poor school climate, and resulting low motivation and satisfaction. A school’s climate is “the quality and character of school life” (The National School Climate Center 2019). It is composed of several areas, including relationships between teachers and administrators and students, school safety, the institutional environment, and the school improvement process (Thapa et al. 2013).2 In this paper, we look at the shares of teachers who face barriers that impede teaching (such as student poverty and poor student health), threats to their safety, a lack of voice and influence over school policy decisions, and a lack of autonomy in the classroom. We then explore the level of morale and satisfaction among teachers, which could be a result of school climate and of other influences (such as pay) described in our series of reports.

Importantly, most of the factors that together create a school’s climate are themselves shaped by larger societal forces such as rising poverty, ongoing racial and economic segregation of schools, and insufficient public investments. Because these larger societal forces contribute to deteriorating working conditions in schools, they cannot be blamed on students, parents, and teachers. (As just one example, students who come to school in poor health because they do not have access to medical care or to assistance programs that provide them with nutritional foods aren’t as prepared to learn as they could be.3) Thus, addressing the poverty- and inequality-related factors that help create a challenging school climate requires investing not only in excellent educators, but also in social workers, physicians, counselors, nurses, and other professionals operating outside the traditional education policy domains (García 2015; García and Weiss 2017a). Research shows that policies to improve school climate could improve the odds that teachers stay in the profession.4 But a poor school climate is not just a factor in the teacher shortage; it can also impede student learning and school performance, lessen teacher effectiveness and morale, and damage the health of the profession overall.5

School climate is shaped by barriers to teaching and learning

Teacher surveys point to a number of conditions among the student body that impede teaching and negatively influence student performance. These conditions include behaviors and factors such as student tardiness and absenteeism, parents’ struggles to be involved in their children’s schools, student disengagement, poor student health, and insufficient student preparation for instruction.6 Across the board, the National Teacher and Principal Survey (NTPS) data we analyzed show that large shares of teachers see these factors as serious problems in their schools, and increasingly so, since a number of these problems worsened between the 2011–2012 and 2015–2016 school years.

As shown in Table 1, in 2015, the share of teachers reporting that various factors were “serious problems” in their schools ranged from around 5 percent for poor student health and class-cutting to nearly 30 percent for poverty specifically and students’ unreadiness to learn. A third set of barriers whose degree of severity falls between these two extremes includes tardiness, cited as a serious problem by 12.1 percent of teachers; absenteeism, cited by 14.9 percent of teachers; apathy among students, reported by 18.4 of teachers; and lack of parental involvement, which more than one in five teachers (21.5 percent) sees as a serious problem in their school.

Table 1

These data affirm our previous reports’ assertions that it is harder to attract and retain teachers in high-poverty schools (see García and Weiss 2019a; 2019b; 2019c). Across all aspects described in Table 1, the share of teachers viewing a given factor as a “serious problem” is between two and nearly five times as high in high-poverty schools relative to low-poverty schools.7

The degree to which poor student health is a problem is one particularly troubling example of greater challenges in high-poverty schools. In high-poverty schools, 8.1 percent of teachers pointed to this issue as a challenge, compared with just 2.0 percent in low-poverty schools. This illustrates the striking disparity in the conditions in which those two groups of teachers are trying to do their jobs—students who are ill are not only more likely to miss a lot of school, but to struggle to focus when they are in class and thus to learn more slowly.8 The data also provide confirmation of well-documented opportunity gaps by socioeconomic status that are associated with achievement gaps: an alarming 38.6 percent of teachers in high-poverty schools report students coming to their classrooms underprepared to learn, versus 12.1 percent of teachers in low-poverty schools.

In our earlier reports we showed the high-poverty schools can also be characterized as harder-to-staff schools (García and Weiss 2019a; 2019b; 2019c). Table 1 here shows that teachers in hard-to-staff/high-poverty schools face additional challenges relative to teachers in low-poverty schools, including higher rates of class-cutting (6.5 percent vs. 2.5 percent, tardiness (16.6 vs. 6.1 percent), student apathy (22.3 percent vs. 11.2 percent) and parents who struggle to engage with the school (31.2 vs. 9.1 percent).9 Finally, teachers weighed in on a topic that has gained national policy attention: student absenteeism. The shares of teachers who report student absenteeism being a problem vary widely across schools.10 Only 8.0 percent of teachers in low-poverty schools reported absenteeism as a serious problem, versus more than double that, 19.7 percent, of teachers in high-poverty schools reporting absenteeism as a serious problem.

Teachers report stress and a lack of safety

The set of school climate indicators in Table 2 speaks to the emotional and mental health, and physical safety, of teachers in the workplace. Across all teachers, one in 20 teachers reports that the stress and disappointments of teaching “aren’t really worth it”: 4.9 percent of teachers strongly agree with that assertion, and, when we look at the 2011–2012 Schools and Staffing (SASS) survey, we see that the share is up slightly from 4.4 percent in 2011. A much larger share of teachers (13.1 percent) strongly agree that student misbehavior interferes with their ability to teach. Most concerning, more than one in five teachers (21.8 percent) report that they have been threatened by a student at the school where they currently teach, and one in eight (12.4 percent) report that they were physically attacked by a student at their current school. Without discussion, these indicators shape the work environment and conditions, and can contribute to shortages by making the profession less attractive.

Table 2

Like the previous indicators of a difficult school climate, stress and lack of safety are more acute problems in high-poverty schools than in low-poverty schools. For example, the share of teachers strongly agreeing that the stress and disappointments of teaching are not worth it is 2.2 percentage points higher in high-poverty schools than in low-poverty schools (5.9 percent versus 3.8 percent, or 1.6 times as large). The gap in the share of teachers frustrated with disruption due to student misbehavior reaches 10 percentage points. There is also a 10 percentage-point gap between the share of teachers who have been threatened by students in high-poverty schools and teachers threatened in low-poverty schools: more than one in four teachers in high-poverty schools has been threatened, compared with about one in six teachers in low-poverty schools. Finally, the shares of teachers who had been physically attacked (14.8 percent in high-poverty schools and 9.5 percent in low-poverty schools) greatly compounds the stress that makes today’s school climate tougher in high-poverty schools.

School climate is shaped by the relationships between teachers and administrators, colleagues, and parents

The relationships between teachers, a school’s administration, and the community more broadly shape a school’s working environment and climate, with repercussions for teachers and also for students (Bryk et al. 2010). This climate affects how well the school provides a learning community in which administrative supports and leadership are strong, there is time for peer collaboration, and employees share a strong sense of purpose. (In the next report, some of these indicators will be examined from the perspective of career supports and professional development.)

Our analysis shows unsatisfactory relationship patterns across the board. Table 3 presents seven attributes of a collegial and supportive school environment. In six of the seven categories reviewed in the table, less than half of teachers report strongly agreeing that the school has that attribute; in other words, less than half report being fully supported by the school administration, their colleagues, or the community in general. The one exception is a proxy of leadership: whether “the principal knows what kind of school he or she wants and has communicated it to the staff.” More than half (51.6 percent) of teachers surveyed said that principals exhibit that attribute. About half (49.6 percent) of teachers report that they see “supportive and encouraging behavior” by school administrators (a proxy for a positive working environment set by the administration). And slightly less than half (47.9 percent) strongly agrees with the statement, “I make a conscious effort to coordinate the content of my courses with that of other teachers” (a proxy of the community environment created by teachers to facilitate coordination, as will be explored further in the next report in the teacher shortage series). Only slightly more than a third of teachers strongly agree that “there is a great deal of cooperative effort among the staff members” (38.4 percent do) or that their colleagues share their views of what the school’s mission should be (36.0 percent). Fewer than one in three teachers affirm that they are recognized for a job well done (32.4 percent), and only 13.3 percent of teachers affirm that they receive a great deal of support from parents for the work they do. Put another way, the survey responses indicate that high shares of teachers experience some level of conflict or disagreement in their schools. Given this level of conflict, it is not surprising that teaching is an unattractive career option, both for people making decisions about their careers and for veteran teachers who are leaving the profession (García and Weiss 2019b).

Table 3

This scenario of a working environment with a degree of conflict or disagreement (which could be described as a poor learning community) is worse in high-poverty schools. A larger share of teachers in high-poverty schools indicate some level of conflict or disagreement in attitudes or beliefs from the administration or colleagues than do teachers in low-poverty schools. By far the biggest gap between high- and low-poverty schools is in support teachers receive from their students’ parents: almost nine out of 10 teachers in high-poverty schools do not feel fully supported by parents for the work they do compared with a still very high eight in 10 teachers in low-poverty schools. The gap in parental support affirms our previous comment in the discussion of Table 1 that schools’ and teachers’ struggles to engage with parents are especially difficult in high-poverty schools.

School climate is shaped by the voice and influence teachers have in their schools and day-to-day work

For teachers, having a sense of purpose and a say over the working conditions and policies of their school is an essential component of a positive school climate, and enhances teaching professionalism. But as shown in Table 4, meager shares of teachers report having a great deal of influence or control over school policy, suggesting a generalized disrespect for teachers’ knowledge of their jobs and professional judgment. A scant 3 percent of teachers report having a great deal of influence over how teachers are evaluated. Other school policy categories with shares under 10 percent are setting discipline policy and hiring new teachers. The category with the highest share of teachers reporting a great deal of influence is establishing the curriculum, but even then just one in five (20.4 percent) teachers have influence over the curriculum. To put it another way, 80 percent or more of teachers do not have a great deal of influence or control over the policies at their schools.

Table 4

Although teachers report much more influence in their classrooms than on school policies, they still indicate a surprisingly small level of control over their daily activities. The shares of teachers who report a great deal of influence or control range from 60 to 70 percent when the action is evaluating and grading students or assigning the amount of homework, but falls to a much lower sub-30 percent share when the actions involve selecting textbooks and other instructional materials and controlling topics and skills to be taught. To put it another way, more than seven in 10 teachers do not control the textbooks they use and the topics and skills they teach.

The usual gaps between working conditions in high- and low-poverty schools also appear here. In the areas of school policy, the gaps between teachers’ autonomy or influence in high- and low-poverty schools are small, in general (under 2 percentage points for all categories except establishing the curriculum), and in fact a slightly greater share of teachers in high-poverty schools report having a great deal of control over setting discipline policy and evaluating teachers. In terms of autonomy in their classrooms, the bottom half of the table shows that in all tasks listed except assigning homework, teachers in high-poverty schools have less of a say than their counterparts in low-poverty schools and that the gaps range from 2 to more than 4 percentage points.

Poor school climate depresses teacher satisfaction and motivation, and teachers’ plans to stay in teaching

Figure A summarizes the findings presented in this report—that school climate indicators are tough across the board. Given the challenging school climate for many teachers, it is little surprise that teachers’ satisfaction, motivation, and desire to stay in teaching is low and has even dwindled slightly in the past few years.11 Figure B shows the shares of teachers who say they are satisfied and who say their peers are satisfied. Dissatisfaction is not only the result of a poor school climate but also a factor leading to a poor school climate: when teachers are not as motivated and engaged as they could be it affects the school climate.

Figure A

Figure B

Almost half of all teachers (48.7 percent) express some level of dissatisfaction with being a teacher at their current school. Just over a quarter would definitely describe their schools’ teachers as a satisfied group (28.7 percent) and affirmatively say that they like the way things are run at the school (26.9 percent). All of the “strongly agree” shares in the figure are lower than they were in the 2011–2012 school year, pointing to lowered satisfaction and motivation across the board.

Table 5 shows that, as with positive school climate factors, teacher satisfaction is lower in high-poverty schools than in low-poverty schools. The gaps between shares of teachers in high- and low-poverty schools who “strongly agree” with the three statements of satisfaction in the table range from 6.3 to 7.9 percentage points (see the last column of table 5). The data also demonstrate that teachers’ motivation (as represented by the share of teachers who are or are not certain they would choose teaching today if given the opportunity to start over) is quite weak. Only about four in 10 teachers say that if they could go back to college and start over, they would certainly go into teaching, and the share is slightly lower in high-poverty schools than in low-poverty schools (41.4 percent vs. 45.5 percent, this share is from 2011–2012).

Table 5

With respect to plans about staying in the profession, a large share of teachers express their expectation of leaving teaching at some point, as opposed to staying as long as possible or until retirement (Table 6). More than one in four teachers plans to quit teaching at some point, i.e., does not plan to stay in teaching for the rest of his or her career. Here, the gaps between high- and low-poverty schools are small.

When we look at data from the 2011–2012 SASS we see that the drop in teachers’ satisfaction, motivation, and expectations are paralleled by a drop in the share of teachers who plan to continue in teaching for the remainder of their careers: this share decreased from 76.0 percent in 2011–2012 to 72.6 percent in 2015–2016.

Table 6

School climate and the struggle to attract and retain teachers

In the previous reports in this series, we saw that low salaries and excessive moonlighting to complement wages with profession-building activities have made teaching particularly unattractive for both current and potential teachers; in addition, those impediments could explain, in part, the gaps between credentials of the teaching workforce in high- and low-poverty schools because all indicators are worse in high-poverty schools (García and Weiss, 2019a, 2019b, 2019c). The findings in this paper provide parallel evidence regarding school climate: the challenging working environments can make teaching an unattractive profession overall; in addition, the comparatively more difficult working environment for teachers in high-poverty schools can contribute the the fact that high-poverty schools have a harder time attracting and retaining highly credentialed teachers than do low-poverty schools.

In this next section, we explore how aspects of school climate—barriers to teaching, stress and physical threats, satisfaction and motivation—are correlated with the supply of available teachers, and thus, implicated in the teacher shortage. We would expect that, across the board, teachers who quit the profession were more likely to have reported, in the year before they quit, feeling stressed, unsatisfied, unsupported, and not involved in setting school or classroom policies. Figure C lists a subset of the negative school climate indicators and reports the share of “staying” and “quitting” teachers who reported, in their responses to the 2011–2012 Schools and Staffing Survey, they they experienced the indicator. Staying teachers are those who, were still at the same school while quitting teachers are those who had quit by the in the 2012–2013 Teacher Follow-up Survey.

Figure C

As the figure shows, larger shares of quitting teachers had reported in the year before they left teaching that key aspects of their school’s climate were problematic than was true among teachers who stayed at their schools. For example, 39.0 percent of teachers who quit felt their students were unprepared to learn, versus 29.4 percent of teachers who stayed at their school. There was a big difference in the stress levels of quitting and staying teachers. The share of quitting teachers who reported being very stressed the year before they quit was 3.5 times as large as the share of staying teachers who reported feeling stressed the previous year. Larger shares of quitting teachers reported that collegiality among teachers was lacking and that they had little influence over school policy or over what they teach in class. Most predictably, the shares of teachers who said they weren’t fully satisfied with teaching and who planned to quit teaching at some point were much higher among those who quit than those who stayed.

The numbers above do not paint a pretty picture about the morale of the current teaching workforce. Among those who stayed, more than one-fifth had reported planning to leave at some point, and over 40 percent had reported some level of dissatisfaction with their jobs. Nearly a third had students who were not prepared to learn, and nearly a fourth were frustrated by the challenge of engaging their students’ parents.12

Conclusion

The various components of a negative working environment—barriers to teaching, stress, physical threats, a lack of say in how to run the classroom, and low levels of satisfaction—interact with one another and make it harder for teachers to do their work, and affect students’ ability to learn. And tough school climates definitely play a role in the teacher shortage: despite their substantial training and ability to deal with the challenges of their job, the negative aspects of the school climate can dissuade young people from becoming teachers and driving some teachers out of our classrooms.

In this report, we show that school climate indicators correlate with teachers’ statuses the year after. Across the board, we note that school climate indicators of teachers who quit were worse than of teachers who stayed the school year before teachers made the decision to quit or stay in the profession.

We also show that, aside from those correlations, certain features that make teaching challenging are so concerning that we would not expect teachers or any professionals to have to handle them without being provided with further supports. When students are unprepared, teachers must spend more time reviewing material and potentially neglecting other students. Students who are in poor health, or who miss school frequently, are not just disruptive on a practical level, but cause concern and emotional distress for teachers who watch them struggle and for other students. These disruptions and distractions lead teachers to feel stressed and disappointed, as does the challenge of engaging parents who, for a variety of reasons, have trouble connecting with teachers in ways that boost their children’s ability to learn. And when the working conditions raise safety concerns—students misbehave, or threaten or attack teachers in the school—it adds to the level of distress.

Our research points to a different source of distress as well: the lack of cooperation and support from the administration and other colleagues, and the limited influence and autonomy teachers have over their daily activities or their schools’ needs, further add to a problematic working environment. Significantly large shares of teachers indicate that their voices go unheard—schools are not fully benefiting from their knowledge, preparation, and expertise.

All of this, of course, depresses satisfaction and drives teachers to consider leaving their schools or the professional altogether. Dissatisfaction increases when poor working conditions are accompanied by weak compensation, lack of professional development opportunities, and the deteriorated prestige of teaching. Clearly, the challenging conditions confronting a growing share of teachers are helping to drive teacher shortages across schools and especially in high-poverty schools.

In sum, the evidence presented in this paper shows that it is imperative that we improve working conditions across the board to stop the teacher exodus and address the substantial pent-up frustration among the existing workforce. As suggested in our companion pieces, only if policymakers think holistically about how to address the teacher shortage will they find the necessary resources to adequately fund our schools, to eliminate the barriers to teaching and learning, and to elevate the respect for teachers’ knowledge, experience, and judgment.

About the authors

Emma García is an education economist at the Economic Policy Institute, where she specializes in the economics of education and education policy. Her research focuses on the production of education (cognitive and noncognitive skills); evaluation of educational interventions (early childhood, K–12, and higher education); equity; returns to education; teacher labor markets; and cost-effectiveness and cost–benefit analysis in education. She has held research positions at the Center for Benefit-Cost Studies of Education, the Campaign for Educational Equity, the National Center for the Study of Privatization in Education, and the Community College Research Center; consulted for MDRC, the World Bank, the Inter-American Development Bank, and the National Institute for Early Education Research; and served as an adjunct faculty member at the McCourt School of Public Policy, Georgetown University. She received her Ph.D. in Economics and Education from Columbia University’s Teachers College.

Elaine Weiss is the lead policy analyst for income security at the National Academy of Social Insurance, where she spearheads projects on Social Security, unemployment insurance, and workers’ compensation. Prior to her work at the academy, Weiss was the national coordinator for the Broader, Bolder Approach to Education, a campaign launched by the Economic Policy Institute, from 2011–2017. BBA promoted a comprehensive, evidence-based set of policies to allow all children to thrive in school and life. Weiss has co-authored and authored EPI and BBA reports on early achievement gaps and the flaws in market-oriented education reforms. She is co-author, with former Massachusetts Secretary of Education Paul Reville, of Broader, Bolder, Better, published by Harvard Education Press in 2019. Weiss came to BBA from the Pew Charitable Trusts, where she served as project manager for Pew’s Partnership for America’s Economic Success campaign. She has a Ph.D. in Public Policy from the George Washington University Trachtenberg School and a J.D. from Harvard Law School.

Acknowledgments

The authors are grateful to Lora Engdahl for her edits and suggested additions to this piece and for her extraordinary contributions to structuring the contents of this series of papers. We also appreciate John Schmitt’s supervision of this project and Lawrence Mishel’s guidance in earlier stages of this research. We acknowledge Julia Wolfe for her assistance with the tables and figures in this report, Kayla Blado for her work disseminating the report and her assistance with the media, John Carlo Mandapat for the infographic that accompanies this report, and the rest of the communications staff at EPI for their contributions to the different components of this report and the teacher shortage series. We appreciate EPI Communications Director Pedro da Costa’s coordination of all the steps required for the publication of this report and of the series.

Data sources used in this report

The analyses presented in this report mainly rely on the Schools and Staffing Survey (SASS) 2011–2012, the Teacher Follow-up Survey (TFS) 2012–2013, and the National Teacher and Principal Survey (NTPS) 2015–2016. The surveys collect data on and from teachers, principals, and schools in the 50 states and the District of Columbia.13 All three surveys were conducted by the U.S. Census Bureau for the U.S. Department of Education. The survey results are housed in the National Center for Education Statistics (NCES), which is part of the Department of Education’s Institute of Education Sciences (IES).

The NTPS is the redesigned SASS, with a focus on “flexibility, timeliness, and integration with other Department of Education data” (NCES 2019). Both the NTPS and SASS include very detailed questionnaires at the teacher level, school level, and principal level, and the SASS also includes very detailed questionnaires at the school district level (NCES 2017). The TFS survey, which is the source of data on teachers who stay or quit, was conducted a year after the 2011–2012 SASS survey to collect information on the employment and teaching status, plans, and opinions of teachers in the SASS. Following the first administration of the NTPS, no follow-up study was done, preventing us from conducting an updated analysis of teachers by teaching status the year after the NTPS. NCES plans to conduct a TFS again in the 2020–2021 school year, following the 2019–2020 NTPS.

The 2015–2016 NTPS includes public and charter schools only, while the SASS and TFS include all schools (public, private, and charter schools).14 We restrict our analyses to public schools and teachers in public noncharter schools.

The 2015–2016 NTPS includes public and charter schools only, while the SASS and TFS include all schools (public, private, and charter schools).15 We restrict our analyses to public schools and teachers in public noncharter schools.

Endnotes

1. For a more detailed review of the media coverage of the shortage, see García and Weiss 2019a. Research on costs come from Carver-Thomas and Darling-Hammond 2017 and the Learning Policy Institute 2017 reports, which estimated that filling a vacancy costs $21,000 on average; and from Carroll 2007, which estimated that the total annual cost of turnover was $7.3 billion per year. According to Strauss 2017, that estimated annual cost of turnover would exceed $8 billion at present.

2. School climate is based on patterns of students’, parents’, and school personnel’s experience of school life; it also reflects norms, goals, values, interpersonal relationships, teaching and learning practices, and organizational structures (The National School Climate Center 2019). In a recent post, Kautz and Ross (2019) explain that “school climate covers both tangible and intangible attributes, including relationships among students and staff, school discipline, student engagement, and safety.” A book on supports for school improvement (Bryk et al. 2010) identifies school climate as one of the essential supports (together with school leadership, parent and community ties, professional capacity of the staff, a student-centered learning climate, and instructional guidance system).

3. A recent book covering the range of supports and the role they play in having children prepared to learn is Weiss and Reville 2019. A seminal book on the opportunity gaps created by poverty and inequality is Rothstein 2004.

4. See summary in Katz 2018.

5. See Darling-Hammond 1999; Ladd and Sorensen 2016; Ronfeldt et al. 2013; Jackson and Bruegmann 2009; Kraft and Papay 2014; Moore-Johnson, Kraft, and Papay 2012; Ladd 2011; Loeb, Darling-Hammond, and Luczak 2005; and Warner-Griffin, Cunningham, and Noel 2018.

6. All of these school climate factors are related to student poverty, some more directly than others. Teachers are also explicitly asked about the extent to which “poverty” is a problem at the school.

7. As poverty is the variable (or proxy of it) that we use to classify our schools into low- and high-poverty schools, we do not stress the gap between the shares of teachers who acknowledge “poverty” as a “serious problem” in the two types of schools (45.1 versus 9.5 percent respectively). This perception of poverty as a problem is, by construction, an expected gap, and thus, we focus on the remaining and worrisome evidence in the table.

8. See Weiss and Reville 2019; Rothstein 2011; and Rothstein 2004.

9. Researchers have explored a number of factors that prevent low-income parents from connecting and partnering with their children’s schools. The reasons range from irregular working schedules, having had bad experiences as students, or even having fewer options of supervised care if they need to participate in school activities (Morsy and Rothstein 2015; Weiss and Reville 2019). García and Weiss 2017b, a study examining early education and parenting practices for the kindergarten classes of 1998 and 2010, found that all families but especially low-income families have, over time, become more involved in their young children’s early education and development. Parents were more likely in 2010 than in 1998 to read regularly to their children; to sing to them; to play games with them; and to enroll them in center-based pre-K programs. Parents in 2010 also had significantly higher expectations for their children’s educational attainment, and mothers themselves were more highly educated—both factors that are associated with higher achievement for those children.

10. See our own analysis on the prevalence of absenteeism and its influence on student performance (see García and Weiss 2018).

11. No one of the overarching factors (pay, as described in García and Weiss 2019c, school climate, as described in this report, or career supports, as described in the next report in this series works in isolation. Rather the factors, along with the broader underinvestments in education, jointly influence voluntary attrition, turnover, and lack of incoming teachers.

12. The gaps between the shares of staying and leaving teachers who report these problems are, in general, wider in high-poverty schools than in low-poverty schools (data not shown in this report). This implies that barriers to teaching, safety, and satisfaction play a different role in schools depending on poverty level. These problems likely played a more important part in driving or keeping teachers away in high-poverty schools (but this hypothesis would need a regression analysis to confirm).

13. The 2015–2016 NTPS does not produce state-representative estimates. The forthcoming 2017–2018 NTPS will support state-level estimates.

14. The forthcoming 2017–2018 NTPS additionally includes the private sector.

15. The forthcoming 2017–2018 NTPS additionally includes the private sector.

References

Bryk, Anthony S., Penny Bender Sebring, Elaine Allensworth, John Q. Easton, and Stuart Luppescu. 2010. Organizing Schools for Improvement: Lessons from Chicago. Chicago: Univ. of Chicago Press.

Carroll, Thomas G. 2007. Policy Brief: The High Cost of Teacher Turnover. National Commission on Teaching and America’s Future.

Carver-Thomas, Desiree, and Linda Darling-Hammond. 2017. Teacher Turnover: Why It Matters and What We Can Do About It. Learning Policy Institute, August 2017.

Darling-Hammond, Linda. 1999. Teacher Quality and Student Achievement: A Review of State Policy Evidence. Seattle: Center for the Study of Teaching and Policy, University of Washington.

García, Emma. 2015. Inequalities at the Starting Gate: Cognitive and Noncognitive Skills Gaps Between 2010–2011 Kindergarten Classmates. Economic Policy Institute, June 2015.

García, Emma, and Elaine Weiss. 2017a. Reducing and Averting Achievement Gaps. Key Findings from the Report ‘Education Inequalities at the School Starting Gate’ and Comprehensive Strategies to Mitigate Early Skills Gaps. Economic Policy Institute, September 2017.

García, Emma, and Elaine Weiss. 2017b. Education Inequalities at the School Starting Gate: Gaps, Trends, and Strategies to Address Them. Economic Policy Institute, September 2017.

García, Emma, and Elaine Weiss. 2018. Student Absenteeism: Who Misses School and How Missing School Matters for Performance. Economic Policy Institute, September 2018.

García, Emma, and Elaine Weiss. 2019a. The Teacher Shortage Is Real, Large and Growing, and Worse Than We Thought: The First Report in the “Perfect Storm in the Teacher Labor Market” Series. Economic Policy Institute, March 2019.

García, Emma, and Elaine Weiss. 2019b. U.S. Schools Struggle to Hire and Retain Teachers: The Second Report in “The Perfect Storm in the Teacher Labor Market” Series. Economic Policy Institute, April 2019.

García, Emma, and Elaine Weiss. 2019c. Low Relative Pay and High Incidence of Moonlighting Play a Role in the Teacher Shortage, Particularly in High-poverty Schools. The Third Report in “The Perfect Storm in the Teacher Labor Market” Series. Economic Policy Institute, May 2019.

Jackson, Kirabo, and Elias Bruegmann. 2009. “Teaching Students and Teaching Each Other: The Importance of Peer Learning for Teachers.” American Economic Journal: Applied Economics 1, no. 4: 85–108.

Katz, Veronica. 2018. Teacher Retention: Evidence to Inform Policy. EdPolicyWorks Policy Brief, October 2018.

Kautz, Tim, and Christine Ross. 2019. “Developing School Climate Surveys for Statewide Accountability in Maryland.” RELevant (Regional Educational Laboratory Mid-Atlantic blog), January 24, 2019.

Kraft, Matthew A., and John P. Papay. 2014. “Can Professional Environments in Schools Promote Teacher Development? Explaining Heterogeneity in Returns to Teaching Experience.” Educational Evaluation and Policy Analysis 36, no. 4: 476–500.

Ladd, Helen. 2011. “Teachers’ Perceptions of Their Working Conditions: How Predictive of Planned and Actual Teacher Movement?” Educational Evaluation and Policy Analysis 33, no. 2: 235–261.

Ladd, Helen F., and Lucy C. Sorensen. 2016. “Returns to Teacher Experience: Student Achievement and Motivation in Middle School.” Education Finance and Policy 12, no. 2: 241–279.

Learning Policy Institute. 2017. What’s the Cost of Teacher Turnover? (calculator). September 2017.

Loeb, Susanna, Linda Darling-Hammond, and John Luczak. 2005. “How Teaching Conditions Predict Teacher Turnover in California Schools.” Peabody Journal of Education 80, no. 3: 44–70.

Moore-Johnson, Susan, Matthew A. Kraft, and John P. Papay. 2012. “How Context Matters in High-Need Schools: The Effects of Teachers’ Working Conditions on Their Professional Satisfaction and Their Students’ Achievement.” Teachers College Record 114, no. 10: 1–39.

Morsy, Leila, and Richard Rothstein. 2015. Five Social Disadvantages That Depress Student Performance: Why Schools Alone Can’t Close Achievement Gaps. Economic Policy Institute, June 2015.

National Center for Education Statistics (NCES) (U.S. Department of Education). 2011–2012. Licensed microdata from the 2011–2012 Schools and Staffing Survey (SASS).

National Center for Education Statistics (NCES) (U.S. Department of Education). 2012–2013. Licensed microdata from the 2012–2013 Teacher Follow-up Survey (TFS).

National Center for Education Statistics (NCES) (U.S. Department of Education). 2015–2016. Licensed microdata from the 2015-2016 National Teacher and Principal Survey (NTPS).

National Center for Education Statistics (NCES) (U.S. Department of Education). 2017. Documentation for the 2011–12 Schools and Staffing Survey. March 2017.

National Center for Education Statistics (NCES) (U.S. Department of Education). 2019. “NTPS Overview” (web page), accessed March 2019.

National School Climate Center. 2019. “What Is School Climate and Why Is It Important?” (web page). National School Climate website. Accessed May 2019.

Ronfeldt, Matthew, Hamilton Lankford, Susanna Loeb, and James Wyckoff. 2013. “How Teacher Turnover Harms Student Achievement.” American Educational Research Journal 50, no. 1: 4–36.

Rothstein, Richard. 2004. Class and Schools: Using Social, Economic, and Educational Reform to Close the Achievement Gap. Washington, D.C., and New York: Economic Policy Institute and Columbia University Teachers College.

Rothstein, Richard. 2011. A Look at the Health-related Causes of Low Student Achievement. Economic Policy Institute, March 2011.

Strauss, Valerie. 2017. “Why It’s a Big Problem That So Many Teachers Quit—and What to Do About It,” Washington Post, November 27, 2017.

Sutcher, Leib, Linda Darling-Hammond, and Desiree Carver-Thomas. 2016. A Coming Crisis in Teaching? Teacher Supply, Demand, and Shortages in the U.S. Learning Policy Institute, September 2016.

Thapa, Amrit, Jonathan Cohen, Shawn Guffey, and Ann Higgins-D’Alessandro. 2013. “A Review of School Climate Research.” Review of Educational Research 83, no. 3: 357–385. https:// doi.org/10.3102/0034654313483907.

Warner-Griffin, Catharine, Brittany C. Cunningham, and Amber Noel. 2018. Teachers’ Perceptions of Autonomy, Satisfaction, Job Security, and Commitment: 1999–2000 and 2011–12. Institute of Education Statistics, National Center for Education Statistics, U.S. Department of Education, Stats in Brief no. 2018-103, March 2018.

Weiss, Elaine, and Paul Reville. 2019. Broader, Bolder, Better. How Schools and Communities Help Students Overcome the Disadvantages of Poverty. Cambridge, Mass.: Harvard Education Publishing Group.

This report is the third in a series examining the magnitude of the teacher shortage and the working conditions and other factors that contribute to the shortage.

What this series finds: The teacher shortage is real, large and growing, and worse than we thought. When indicators of teacher quality (certification, relevant training, experience, etc.) are taken into account, the shortage is even more acute than currently estimated, with high-poverty schools suffering the most from the shortage of credentialed teachers.

What this report finds: The perceived financial hardships in teaching are real. This report adds to the compelling evidence in Sylvia Allegretto and Larry Mishel’s recent research showing that teachers are paid a lot less than other comparable college graduates. After accounting for education, experience, and other factors known to affect earnings, teachers’ weekly wages in 2018 were 21.4 percent lower than their nonteaching peers. In 1996 that weekly wage penalty was 6.3 percent. Our report identifies other indicators that teacher pay is too low and declining. For example, in the 2015–2016 school year, 59.0 percent of teachers took on additional paid work either in the school system or outside of it—up from 55.6 percent in the 2011–2012 school year. A majority of moonlighters (44.1 percent) were taking on second jobs within the school system, such as coaching, student activity sponsorship, mentoring other teachers, or teaching evening classes; 18.2 percent were working outside of the school system; and 5.7 percent were receiving compensation based on student performance. For these teachers, moonlighting makes up a substantial 7.0 percent share of their combined base salary and extra income. Financial stress is greater for teachers in high-poverty schools. Relative to teachers in low-poverty schools, teachers in high-poverty schools are paid less ($53,300 vs. $58,900), receive a smaller amount from moonlighting ($4,000 vs. $4,300), and the moonlighting that they do is less likely to involve paid extracurricular or additional activities for the school system that generate extra pay but also help them grow professionally as teachers (data are for 2015–2016). Data suggest a relationship between low salaries and quitting. Teachers who ended up quitting before the 2012–2013 school year had lower base salaries ($50,800 vs. $53,300) and were more likely to be supplementing their base pay with work outside the school system in the year before they quit (18.4 percent vs. 16.3 percent).

Why it matters: The data show that fewer people are willing to make the choice to be in a profession that puts them at a financial disadvantage. And not only is pay unattractive, as the next two reports in this series will show, but also teachers aren’t being provided with the working conditions and professional development opportunities that would help them do their jobs and also build their careers. The existing shortage of teachers harms students, teachers, and the public education system as a whole. In addition, the fact that the shortage—and the factors that drive it—is more acute in high-poverty schools challenges the U.S. education system’s goal of providing a sound education equitably to all children.

What we can do about it: Tackle the pay and other factors that are prompting teachers to quit and dissuading people from entering the teaching profession. In addition to policy interventions and institutional decisions to invest more in teacher pay and to widen teachers’ access to paid profession-building activities, we must provide extra supports and funding to high-poverty schools and their teachers, not only to support students directly, but also to reduce teacher shortages.

Introduction

The teacher shortage in the nation’s K–12 schools is an increasingly recognized but still poorly understood crisis: The shortage is discussed by the media and policymakers, and researchers have estimated its size (about 110,000 teachers in the 2017–2018 school year, up from no shortage before 2013, according to Sutcher, Darling-Hammond, and Carver-Thomas 2016), and even quantified part of its cost.1 The shortage constitutes a crisis because of its negative effects on students, teachers, and the education system at large. But it is poorly understood because the reasons for it are complex and interdependent. The first report in this series, The Teacher Shortage Is Real, Large and Growing, and Worse Than We Thought (García and Weiss 2019a), established that current national estimates of the teacher shortage likely understate the magnitude of the problem: When issues such as teacher qualifications and the unequal distribution of highly credentialed teachers across high- and low-poverty schools are taken into consideration, the teacher shortage problem is much more severe than previously identified.

Building on that research, and using the same quality and equity angles, the second report in this series, U.S. Schools Struggle to Hire and Retain Teachers (García and Weiss 2019b), showed that schools are having difficulties in staffing their schools and are leaving vacancies unfilled despite actively trying to hire for them. High-poverty schools are hit hardest: They find it more difficult to fill vacancies than low-poverty schools and schools overall, and they experience higher turnover and attrition than low-poverty schools. One factor behind staffing difficulties is the high share of public school teachers leaving their posts: 13.8 percent are either leaving their school or leaving teaching altogether, according to most recent data. Another factor is the dwindling pool of applicants to fill vacancies. From the 2008–2009 to 2015–2016 school years, there was a 15.4 percent drop in the number of education degrees awarded and a 27.4 percent drop in the number of people who completed a teacher preparation program. Schools are also having a harder time retaining credentialed teachers: this can be seen in the small but growing share of all teachers who are newly hired and in their first year of teaching (4.7 percent) and in the substantial shares of teachers who quit who are certified and experienced. This challenge is also more acute for high-poverty schools.

This third report in the series focuses on a likely factor behind why teachers are leaving the profession and fewer people are becoming teachers: teacher pay. Specifically it looks at how teacher compensation compares with compensation in nonteaching occupations. It also delves into an aspect that has received increasing attention recently, which is whether teachers work multiple jobs, and what share of teachers supplement their earnings by moonlighting. When we look at what quitting teachers were doing the year before they quit, we find that their pay and moonlighting activities are different than for teachers who remain in teaching. Teachers who end up quitting their jobs received, on average, lower salaries, they participated less in the kinds of paid extracurricular activities that complement their professional development (activities like coaching students or mentoring teachers), and they participated more in working options outside the school system than did teachers who stayed at their schools. In high-poverty schools, teachers face compounded challenges. Relative to teachers in low-poverty schools, teachers in high-poverty schools are paid less, receive a smaller amount from moonlighting, and the moonlighting that they do is less likely to involve paid extracurricular or additional activities for the school system that generate extra pay but also help them grow professionally as teachers.

We discuss the likely influence of the large and rapidly growing “teacher pay gap”—how much less teachers earn than their comparably educated peers in other professions—on the weakening attractiveness of public education as a profession, and on the rising rate of teacher attrition. These findings suggest that efforts to address teacher shortages must consider attending to the deterioration in pay and working conditions for teachers, especially in high-poverty schools.

Low wages and other compensation

Teaching has never been a particularly high-paying profession, but in recent decades, teachers have lost ground to their peers in other professions, even other public-sector professionals who are also paid less than private-sector equivalents. Historically, teaching and nursing were the two main professional occupations open to women. Low pay was thus in part a consequence of occupational segregation by gender. This occupational segregation continued even as women began to enter other fields in large numbers: Teaching remained a female-dominated profession and teachers’ pay fell further behind pay in other, more gender-balanced, professions.2 However, low pay can also be a cause of gender segregation, and this explanation for low pay begs the question of why teaching remained a “pink collar” job.

Public school teachers represent about 87 percent of all K–12 teachers in the United States (NCES 2018a). Because public school teachers are paid by governments, their pay is only indirectly affected by market forces. School districts that pay below-market wages may be understaffed and underperform, but they will continue to operate nonetheless. Of course, teacher pay is not set in a societal vacuum, and school districts do take into account teachers’ other public- and private-sector job opportunities.3 As a result, experienced teachers may not immediately leave in response to worsening pay and working conditions, even if they feel undervalued and unmotivated. Declining relative pay and prestige and other factors may thus not have an immediate effect on staffing. However, increased turnover and vacancies (and less visible declines in credentials held by the teaching workforce) will appear with a lag as some teachers become so dissatisfied that they leave, and as more young people are dissuaded from obtaining teaching degrees and entering the profession.4

Whatever the exact causes, the teacher wage penalty—how much less teachers earn in weekly wages than similar college graduates (i.e., after accounting for education, experience, and other factors known to affect earnings)—has grown over time, as shown in Figure A. The figure, reprinted from a major EPI study of this problem (Allegretto and Mishel 2019), shows that the average weekly pay penalty hit a maximum of 21.4 percent in 2018, a sharp increase from just a quarter-century ago (1996), when it was just 6.3 percent. Even when teacher benefits—including pensions, which have been used to compensate, in part, for educators’ relatively low salaries—are taken into account, the total compensation penalty was 13.1 percent in 2018, up from 9.8 percent in 2004 and from 2.9 in 1993.5 Note that the compensation penalty would be even larger than measured by Allegretto and Mishel if it weren’t for health and pension benefits becoming more costly to employers without becoming more generous in practice for teachers, due to the rising relative cost of health insurance.6

Figure A

The teacher pay penalty is even stiffer for teachers in high-poverty schools. A teacher is in a low-poverty school if less than 25 percent of the student body in his/her classroom is eligible for free or reduced-price lunch programs; a teacher is in a high-poverty school if 50 percent or more of the student body is his/her classroom is eligible for those programs. As shown in Figure B, the base salary in high-poverty schools is substantially lower than in low-poverty schools, with the gap of about $5,600 representing 9.5 percent of the 2015–2016 base salary of teachers in low-poverty schools. In other words, on average, teachers in high-poverty schools make about 10 percent less than teachers in low-poverty schools.

Figure B

As Figures A and B show, teachers in high-poverty schools face a double disadvantage in their pay: first, relative to comparable college-educated workers in nonteaching jobs, and second, relative to their peers in low-poverty schools.7 Put plainly, teacher salaries are just too low “to support a middle-class existence” (Podolsky et al. 2016 and 2019), a reality that is especially true in our hardest-to-staff schools.

More evidence that teacher pay is not only low but falling

The above discussion of the growing teacher pay penalty shows that teacher pay is relatively low and suggests that it is not improving in recent years. The evidence that teacher salaries have in fact deteriorated over the years is found in several sources. The survey data used in this report and in the rest of the reports in this series produce a short-term comparison of salaries in the current decade. Specifically, we find that the average base salary for teachers decreased slightly from about $55,400 in 2011–2012 to $55,000 in 2015–2016 (in 2015 constant dollars).8 Allegretto and Mishel (2019, Figure A) found that the average weekly wage of public school teachers was $1,216 in 1996 and $1,195 in 2018 (in 2018 constant dollars, using the Current Population Survey Outgoing Rotation Group data). In 2018, Time reported that “the country’s roughly 3.2 million full-time public-school teachers (kindergarten through high school) are experiencing some of the worst wage stagnation of any profession, earning less on average, in inflation-­adjusted dollars, than they did in 1990, according to Department of Education (DOE) data” (Reilly 2018).

Moonlighting is another indicator that teacher pay is low

Another indicator that teacher pay is low and even falling is that a growing number of teachers are taking second jobs, on top of their full-time job at school. While the surveys providing the data we use do not ask teachers why they take on this extra work, we can speculate, based on news reports, that some teachers are not merely looking to supplement their incomes or to engage in activities that help them build their careers but to make ends meet (Reilly 2018; Long 2019; Talley et al. 2018). According to CNN, teachers in Denver are living with roommates and pursuing side jobs such as serving as ride-sharing drivers to be able to afford basic living expenses (Zdanowicz 2019). Further evidence that some teachers need extra outside income to make ends meet comes from Boser and Straus (2014, 2), who described situations of teachers relying on food stamps and other public benefits: “Mid-career teachers who head families of four or more in multiple states … qualify for several benefit programs, including the Children’s Health Insurance Program and the School Breakfast and Lunch Program.”

As shown in Table 1, in the 2015–2016 school year a large share of teachers (59.0 percent) reported receiving additional compensation aside from their base salary (this is up slightly from the 2011–2012 school year, when it was 55.6 percent). A second job within the school system was the single largest source of moonlighting: 44.1 percent of teachers received additional compensation for activities such as coaching, student activity sponsorship, mentoring other teachers, or teaching evening classes. The share of teachers working jobs outside the school system was smaller but still substantial: almost one in five (18.2 percent) worked jobs outside the school system. Smaller shares earned income from other sources within the school system such as “state supplements” (8.5 percent),9 and from merit pay or other sources based on student performance (5.7 percent).

Table 1

Figure C shows the compensation moonlighters receive from moonlighting activities as a share of their combined base salary and moonlighting pay. For moonlighters, the total compensation from moonlighting activities was $4,100, which represents 7.0 percent of teachers’ base salary plus compensation from moonlighting during the 2015–2016 school year.

Figure C

Data from both Table 1 and Figure C provide details on how moonlighting for teachers in high-poverty schools compares with moonlighting for teachers in low-poverty schools. As shown in Figure C, Teachers in high-poverty schools report lower-than-average base pay ($53,100) than their peers in low-poverty schools ($58,800) and the amount they receive from moonlighting is $300 lower than the pay generated from moonlighting by teachers in low-poverty schools ($4,000 vs. $4,300). As shown in Table 1, a higher share of teachers in high-poverty schools receives income as state supplements (9.3 percent, versus 7.0 percent, a 2.3 percentage-point gap). However, the share of teachers who work a second job within the school system is 3.9 percentage points lower in high-poverty school than in low-poverty schools. And the share of teachers who work a second job outside the school system is 2.4 percentage points lower in high-poverty schools than in low-poverty schools. Despite the fact that moonlighting brings in more extra pay for teachers in low-poverty schools than in high-poverty schools, the lower base salaries of teachers in high-poverty schools overall (as shown in Figure B), and of moonlighting teachers in these high-poverty schools (as shown in Figure C), means that moonlighting income is nevertheless very important for teachers in high-poverty schools. As seen in Figure C, moonlighting compensation represents roughly 7 percent of total base salary plus earnings from moonlighting for teachers in high-poverty schools.

While financial hardship is likely leading to much of the teacher moonlighting we are seeing, there are other reasons why teachers would undertake extra jobs voluntarily. Some of these within-the-school-system second jobs allow teachers to engage more deeply with their schools, enjoy enhanced collegiality with other peers, or further their professional development. Examples of such paid profession-building activities include mentoring other teachers, coaching students, sponsoring student activities, and engaging in other similar activities. But moonlighting can increase stress and drive disengagement, as teachers are forced to juggle multiple schedules and have their family and leisure time reduced. And if moonlighting occurs outside the school system, the challenges of juggling the extra work are likely greater. For these reasons, the causes and conditions under which this moonlighting occurs determine whether it makes teaching more or less attractive (and thus whether it helps or hurts recruitment and retention).

In addition, when teachers in high-poverty schools supplement their base teacher pay with outside sources of income, lower shares are compensated based on what we call “profession-building” extra work, such as coaching student activities or mentoring other teachers. Teachers in high-poverty schools either don’t have as much access to these profession-building activities (or aren’t paid for them). These disparities in opportunities and compensation for second jobs between teachers in high- and low-poverty schools make teaching less attractive for teachers in high-poverty schools and make high-poverty schools less appealing work destinations for potential new teachers, which in turn contribute to shrinking the pool of applicants for high-poverty schools and to expanding the population of teachers leaving high-poverty schools. In this way, moonlighting can exacerbate the inequitable distribution of highly credentialed teachers across schools.

Teacher pay and the struggle to attract and retain teachers

Low salaries and the perceived need to supplement income by moonlighting—at least among some teachers—make teaching particularly unattractive for the strongest teachers. This deterrent to teaching among the most-credentialed teachers explains in part why not all teachers currently teaching are fully credentialed, and also drives some of the gap between qualifications of teachers in low- and high-poverty schools (for more on the shares of teachers who are not fully credentialed and on the gaps in credentials see García and Weiss 2019a).

We next show that there are gaps between the pay and access to profession-building activities of teachers staying in the profession and the pay and access to profession-building activities of teachers leaving the profession. These gaps point to a relationship that suggests that low pay and a lack of access to paid profession-building activities make it more difficult for schools to attract and retain highly credentialed teachers.

Figure D compares salaries and engagement in moonlighting activities of teachers who stay with teachers who quit. Teachers who quit have lower salaries and engage less in paid profession-building activities than teachers who stay. Annual salaries of those who quit were $2,500 lower than salaries of those who stayed ($50,800 vs. $53,300; these salaries come from the 2011–2012 Schools and Staffing Survey data). Larger shares of teachers who quit had reported in the year before they quit that they were supplementing their base pay with work outside the school system (18.4 percent versus 16.3 percent of teachers who stayed). A smaller share of the quitting teachers were supplementing their base teacher pay by doing paid profession-building activities within the school system the year before they quit (33.3 percent vs. 42.7 percent, a gap of close to 10 percentage points).

Figure D

Conclusion

The teacher shortage is real, large, and grew rapidly in a very short number of years. We see no signs that it will go away. Unfortunately, this overall teacher shortage is not the only issue of concern: In our two previous reports, we documented that the problem is more acute when teacher qualifications (credentials, relevant training, experience, etc.) are taken into account; and that high-poverty schools suffer the most from the shortage of highly credentialed teachers. This third report in the series helps shed light on the role that pay and other compensation play in the teacher shortage.

Based on the evidence provided in this report, low relative teacher pay and school poverty are clearly implicated as factors behind the teacher shortage. This relationship between teacher pay and the teacher shortage suggests that policy interventions and institutional decisions to invest more in teacher pay and to widen teachers’ access to paid profession-building activities such as coaching and mentoring could help address the teacher shortage. In addition, given the inequities reported in the first two reports of the series, and the new pay and moonlighting inequities between high- and low-poverty schools described here, we must provide extra supports and funding to high-poverty schools and their teachers, not only to support students directly, but also to reduce teacher shortages.

In our forthcoming reports in the series, we will continue to discuss the factors that make teaching an unattractive profession for current and potential new teachers, going beyond pay to discuss working conditions and the deteriorated prestige of teaching. We argue that policymakers need to think holistically about how to address these factors, and to act fast in ways that support teachers economically, at their workplaces, and enhance the prestige and the professionalism of teaching.

About the authors

Emma García is an education economist at the Economic Policy Institute, where she specializes in the economics of education and education policy. Her research focuses on the production of education (cognitive and noncognitive skills), evaluation of educational interventions (early childhood, K–12, and higher education), equity, returns to education, teacher labor markets, and cost analysis in education. She has held research positions at the Center for Benefit-Cost Studies of Education, the Campaign for Educational Equity, the National Center for the Study of Privatization in Education, and the Community College Research Center; consulted for MDRC, the World Bank, the Inter-American Development Bank, and the National Institute for Early Education Research; and served as an adjunct faculty member at the McCourt School of Public Policy, Georgetown University. She received her Ph.D. in economics and education from Columbia University’s Teachers College.

Elaine Weiss is the lead policy analyst for income security at the National Academy of Social Insurance, where she spearheads projects on Social Security, unemployment insurance, and workers’ compensation. Prior to her work at the academy, Weiss was the national coordinator for the Broader Bolder Approach (BBA) to Education, a campaign launched by the Economic Policy Institute, from 2011 to 2017. BBA promoted a comprehensive, evidence-based set of policies to allow all children to thrive in school and life. Weiss has co-authored and authored EPI and BBA reports on early achievement gaps and the flaws in market-oriented education reforms. She is co-authoring Broader, Bolder, Better, a book with former Massachusetts Secretary of Education Paul Reville that will be published by Harvard Education Press in June 2019. Weiss came to BBA from the Pew Charitable Trusts, where she served as project manager for Pew’s Partnership for America’s Economic Success campaign. She has a Ph.D. in public policy from the George Washington University Trachtenberg School and a J.D. from Harvard Law School.

Acknowledgments

The authors are grateful to Lora Engdahl for her edits to this piece and for her extraordinary contributions to structuring the contents of this series of papers. We are especially thankful for Monique Morrissey for her contributions to framing the discussion on teacher pay and pay more broadly, and for her notes on teacher health and pension benefits. We also appreciate John Schmitt’s supervision of this project and Lawrence Mishel’s guidance in earlier stages of this research and his later comments on this piece. We acknowledge Julia Wolfe for her assistance with the tables and figures in this report, Kayla Blado for her work disseminating the report and her assistance with the media, John Carlo Mandapat for the infographic that accompanies this report, and the rest of the communications staff at EPI for their contributions to the different components of this report and the teacher shortage series. We appreciate EPI Communications Director Pedro da Costa’s coordination of all the steps required for the publication of this report and of the series.

Data sources used in this report

The analyses presented in this report mainly rely on the Schools and Staffing Survey (SASS) 2011–2012, the Teacher Follow-up Survey (TFS) 2012–2013, and the National Teacher and Principal Survey (NTPS) 2015–2016. The surveys collect data on and from teachers, principals, and schools in the 50 states and the District of Columbia.10 All three surveys were conducted by the U.S. Census Bureau for the U.S. Department of Education. The survey results are housed in the National Center for Education Statistics (NCES), which is part of the Department of Education’s Institute of Education Sciences (IES).

The NTPS is the redesigned SASS, with a focus on “flexibility, timeliness, and integration with other Department of Education data” (NCES 2019). Both the NTPS and SASS include very detailed questionnaires at the teacher level, school level, and principal level, and the SASS also includes very detailed questionnaires at the school district level (NCES 2017). The TFS survey, which is the source of data on teachers who stay or quit, was conducted a year after the 2011–2012 SASS survey to collect information on the employment and teaching status, plans, and opinions of teachers in the SASS. Following the first administration of the NTPS, no follow-up study was done, preventing us from conducting an updated analysis of teachers by teaching status the year after the NTPS. NCES plans to conduct a TFS again in the 2020–2021 school year, following the 2019–2020 NTPS.

The 2015–2016 NTPS includes public and charter schools only, while the SASS and TFS include all schools (public, private, and charter schools).11 We restrict our analyses to public schools and teachers in public noncharter schools.

Endnotes

1. For a more detailed review of the media coverage of the shortage, see García and Weiss 2019a. Research on costs come from Carver-Thomas and Darling-Hammond 2017 and the Learning Policy Institute 2017, reports which estimated that filling a vacancy costs $21,000 on average; and Carroll 2007, which estimated that the total annual cost of turnover was $7.3 billion per year. According to Strauss 2017, that estimated annual cost of turnover would exceed $8 billion at present.

2. See Figure 10 in NCES 1993 for a historical overview of the share of female and male teachers in primary and secondary schools. The most recent estimate from NCES is that 76.6 percent of teachers are female (NCES 2018b).

3. However, teachers’ investments in formal and on-the-job training, as well as pension benefits designed to reward tenure, makes turnover costly not only for school districts but for teachers themselves. When a workforce is less responsive to changes in wages than that workforce would be in a competitive market it means the workforce is operating in a “monopsonistic” labor market. See the discussion of monopsony as it applies to teachers in CEA 2016. Researchers distinguish between a monopsony model as technically defined (one with only one or, stretching the definition, relatively few employers) and a market that does not meet the assumptions of a “perfectly competitive” labor market. For a recent discussion of monopsony and its effects on wages see Bivens, Mishel, and Schmitt 2018; for a discussion of monopsony power and the effects on wages, see Bivens and Shierholz 2018.

4. Though public-sector labor markets behave differently from textbook competitive labor markets, and these differences may explain some part of how teacher shortages and declining standards can persist without automatically driving up pay and improving working conditions, an adapted monopsony model does not by itself explain why teachers’ relative pay and working conditions declined in the first place. Instead, the teacher pay gap appears to have both policy and political roots. Anxiety over global competition and widening pay gaps between more- and less-educated workers in the United States has increased pressure on public schools to raise the bar academically. These pressures and lack of reforms or solutions have increased teachers’ stress and reduced their autonomy without any associated increases in teachers’ pay, bargaining power, or prestige.

5. Their report provides comprehensive information on the gaps by gender and by state, showing significant differences on both dimensions. A recent report by the National Education Association also shows the salaries for public school teachers and instructional staff by state in the two most recent school years, the change between these years, and the total growth from 2009–2010 to 2018–2019 (NEA 2019).

6. Employer contributions toward employee pension benefits are supposed to cover the normal cost of the plan—the estimated cost of benefits accrued by workers in a given year—plus or minus an amount needed to pay down any unfunded liability or surplus over a given number of years. The normal cost of pension benefits has increased in recent years due to more conservative actuarial assumptions (lower assumed investment returns and rising life expectancy). In addition, the cost of paying down unfunded pension liabilities increased significantly after 2008 due to the temporary effects of the stock market downturn and a change in accounting standards requiring states to treat future retiree health benefits as liabilities in financial statements. Unlike accrued pension benefits, retiree health benefits are generally not legally guaranteed. However, the Government Accounting Standards Board decided that they should be treated as liabilities because states may have a “constructive obligation” to provide these benefits (GASB 2007). The switch from pay-as-you-go to advance-funded retiree health benefits would have temporarily increased outlays for these benefits under any circumstances, but the increase was accelerated by an assumption of rapid health cost inflation over the projected lifetimes of participants, even though rapid cost inflation invariably results in cost shifting from employers to workers and retirees. Many states also shortened amortization periods for paying down unfunded pension liabilities, which reduces long-term interest costs but increases short-term outlays. Whatever one’s opinion of these changes in assumptions and methods, they had the effect of increasing current outlays for retiree health and pension benefits even as the value of these benefits to workers and retirees actually declined due to benefit cuts in the form of lower cost-of-living allowances and other reduced benefits.

7. This disadvantage might more accurately be described as a tripled or quadrupled compounded disadvantage when one considers the challenging workplace conditions and poor supports teachers in high-poverty schools are more likely to face than teachers in low-poverty schools, as shown in the next two reports in this series. In other words, not only are teachers in high-poverty schools not receiving higher base salaries to compensate for all the challenges in these schools, but also teachers in these schools are paid less.

8. These statistics are raw descriptive statistics, and are not adjusted for any other factor or circumstance.

9. The survey asked teachers, “During the current school year, have you earned income from any other sources from this school system, such as a state supplement, etc.?” While it is not entirely clear what these supplements are, and we were not able to find a definition, it seems that many local school districts attempt to boost insufficient teacher pay through locally raised funds that they distribute to teachers, and that not all districts have the resources to achieve this compensation.

10.  The 2015–2016 NTPS does not produce state-representative estimates. The forthcoming 2017–2018 NTPS will support state-level estimates.

11. The forthcoming 2017–2018 NTPS additionally includes the private sector.

References

Allegretto, Sylvia, and Lawrence Mishel. 2019. The Teacher Weekly Wage Penalty Hit 21.4 Percent in 2018, a Record High: Trends in the Teacher Wage and Compensation Penalties Through 2018. Economic Policy Institute and the Center on Wage & Employment Dynamics at the University of California, Berkeley, April 2019.

Bivens, Josh, Lawrence Mishel, and John Schmitt. 2018. It’s Not Just Monopoly and Monopsony. How Market Power Has Affected American Wages. Economic Policy Institute, December 2018.

Bivens, Josh, and Heidi Shierholz. 2018. What Labor Market Changes Have Generated Inequality and Wage Suppression? Employer Power Is Significant but Largely Constant, Whereas Workers’ Power Has Been Eroded by Policy Actions. Economic Policy Institute, December 2018.

Boser, Ulrich, and Chelsea Straus. 2014. Mid- and Late-Career Teachers Struggle with Paltry Incomes. Center for American Progress, July 2014.

Carroll, Thomas G. 2007. Policy Brief: The High Cost of Teacher Turnover. National Commission on Teaching and America’s Future.

Carver-Thomas, Desiree, and Linda Darling-Hammond. 2017. Teacher Turnover: Why It Matters and What We Can Do About It. Learning Policy Institute, August 2017.

Council of Economic Advisers (CEA). 2016. Labor Market Monopsony: Trends, Consequences, and Policy Responses. Issue Brief, October 2016.

García, Emma, and Elaine Weiss. 2019a. The Teacher Shortage Is Real, Large and Growing, and Worse Than We Thought: The First Report in the “Perfect Storm in the Teacher Labor Market” Series. Economic Policy Institute, March 2019.

García, Emma, and Elaine Weiss. 2019b. U.S. Schools Struggle to Hire and Retain Teachers: The Second Report in “The Perfect Storm in the Teacher Labor Market” Series. Economic Policy Institute, April 2019.

Governmental Accounting Standards Board (GASB). 2007. Other Postemployment Benefits: A Plain-Language Summary of GASB Statements No. 43 and No. 45.

Learning Policy Institute. 2017. What’s the Cost of Teacher Turnover? (calculator). September 2017.

Long, Cindy. 2019. “Money Still Matters: A Tale of Two Teachers,” National Education Association, accessed April 30, 2019.

National Center for Education Statistics (NCES). 1993. “Figure 10. Percentage of Elementary and Secondary School Teachers, by Sex: 1869–70 to Fall 1990.” In 120 Years of American Education: A Statistical Portrait. U.S. Department of Education, Office of Educational Research and Improvement, January 1993.

National Center for Education Statistics (NCES) (U.S. Department of Education). 2011–2012. Licensed microdata from the 2011–2012 Schools and Staffing Survey (SASS).

National Center for Education Statistics (NCES) (U.S. Department of Education). 2012–2013. Licensed microdata from the 2012–2013 Teacher Follow-up Survey (TFS).

National Center for Education Statistics (NCES) (U.S. Department of Education). 2015–2016. Licensed microdata from the 2015-2016 National Teacher and Principal Survey (NTPS).

National Center for Education Statistics (NCES) (U.S. Department of Education). 2017. Documentation for the 2011–12 Schools and Staffing Survey. March 2017.

National Center for Education Statistics (NCES). 2018a. “Table 208.20. Public and Private Elementary and Secondary Teachers, Enrollment, Pupil/Teacher Ratios, and New Teacher Hires: Selected Years, Fall 1955 Through 2027.” Digest of Education Statistics: 2017, Institute of Education Sciences, U.S. Department of Education, NCES no. 2018-070, January 2018.

National Center for Education Statistics (NCES). 2018b. “Table 209.10. Number and Percentage Distribution of Teachers in Public and Private Elementary and Secondary Schools, by Selected Teacher Characteristics: Selected Years, 1987–88 through 2015–16.” Digest of Education Statistics: 2017, Institute of Education Sciences, U.S. Department of Education, NCES no. 2018-070, January 2018.

National Center for Education Statistics (NCES) (U.S. Department of Education). 2019. “NTPS Overview” (web page), accessed March 2019.

National Education Association (NEA). 2019. Rankings of the States 2018 and Estimates of School Statistics 2019, NEA Research, April 2019.

Podolsky, Anne, Tara Kini, Joseph Bishop, and Linda Darling-Hammond. 2016. Solving the Teacher Shortage: How to Attract and Retain Excellent Educators. Learning Policy Institute, September 2016.

Podolsky, Anne, Tara Kini, Linda Darling-Hammond, and Joseph Bishop. 2019. “Strategies for Attracting and Retaining Educators: What Does the Evidence Say?” Education Policy Analysis Archives 27(38). http://dx.doi.org/10.14507/epaa.27.3722

Reilly, Katie. 2018. “‘I Work 3 Jobs and Donate Blood Plasma to Pay the Bills.’ This Is What It’s Like to Be a Teacher in America,” Time, September 13, 2018.

Strauss, Valerie. 2017. “Why It’s a Big Problem That So Many Teachers Quit—and What to Do About It,” Washington Post, November 27, 2017.

Sutcher, Leib, Linda Darling-Hammond, and Desiree Carver-Thomas. 2016. A Coming Crisis in Teaching? Teacher Supply, Demand, and Shortages in the U.S. Learning Policy Institute, September 2016.

Talley, Tim, Melissa Daniels, Michael Melia, and John Raby. 2018. “‘I Just Have to Do It.’ Teachers Struggle with Second Jobs,” AP News, April 15, 2018.

Zdanowicz, Christina. 2019. “Denver Is So Expensive That Teachers Have to Get Creative to Make Ends Meet.” CNN, February 11, 2019.