Everyone who cares about racial justice should be focused on doing what’s best for students and their learning—not on school buildings or the employment impact of closing some of them.
Exactly one decade ago, a national coalition of activists released “Death by a Thousand Cuts,” a report making the moral case against school closures. The subtitle—“Racism, School Closures, and Public School Sabotage”—made clear that racial disparities would be a central talking point in political efforts to keep underenrolled buildings open.
The strategy apparently worked so well that the latest “action kit” for opposing efforts to close buildings in the face of declining enrollment and expiring pandemic-era federal support leans heavily on the same arguments. “Nationwide, Black and Latine (sic) students disproportionately experience school closures,” the new strategy document asserts on its first page.
It is true that the public schools shuttered in the first two decades of this century disproportionately enrolled students of color. But this does not imply, as closure opponents claim, that these buildings closed because of the racial composition of their student bodies. Nor does it show that the closures were a bad policy or that they made minority students worse off academically. Behind the numbers is a much more complex story about the inequities built into our public education system and the dysfunctional politics that creates them. Continuing to divert scarce resources to keep half-empty buildings open is likely to worsen many of these disparities.
Lying with statistics
In a May letter to the U.S. Department of Education, several civil rights groups boldly asserted that they had “compelling evidence that districts are disproportionately and intentionally opting to close schools serving majority Black students” (emphasis added). They went on to cite a new Harvard Education Review study by Stanford professor Francis Pearman (who discussed it recently on Fordham’s podcast). There’s a problem, however: That study does not show what the civil rights legal groups and Pearman have claimed.
Pearman and his co-author collected records on school closures from the National Center for Educational Statistics and then combined these data with a variety of other information about individual school buildings, including test scores in some grades. Looking at the simple bivariate relationship, they found that majority-Black schools were a shocking three-times more likely to have closed. But statistically adjusting for a small number of other observable school characteristics dramatically reduced the apparent disparities—eliminating nearly 90 percent of the initial racial gap! The remaining 10 percent, they concluded, must be evidence of intentional racial discrimination.
That is not the case. Rather, the residual disparities simply reflect the fact that the authors could include only a limited set of observable characteristics in their analysis and their statistical model was missing other key factors that drive closure decisions.[i]
Perhaps the most important variable is the district in which a school is located. Historically, closures were rare. In a typical school year, only about 3 percent of all districts shuttered even a single school. (Recent enrollment losses suggest that closures will become more common in coming years.) As it happens, districts that have at least one majority Black school were far more likely to also experience a school closure, creating an apparent correlation between the racial composition of students and the probability that their school closes.[ii]
This correlation tells us only that districts with large minority student populations were more likely to close schools, not that they intentionally targeted these closures in a discriminatory way among the buildings that they operate. Comparing the student bodies of schools that close to buildings in other districts—most of them located in other parts of the country—that don’t close any tells us nothing about which factors actually influence the key decisionmakers in the affected communities. It’s like trying to determine whether snow lifts are sited in a racist way by comparing the areas around publicly owned ski resorts to neighborhoods in parts of the country that don’t get any snow.
To establish that districts are intentionally closing schools serving majority Black students, as the civil rights groups have asserted, we would need first to limit the analysis to districts that actually close schools, and only in the years in which such closures occur. Then we need to compare the racial composition of schools that close to the student bodies of schools in the same district that remain open. I carry out exactly this analysis in my forthcoming book, which includes a chapter on the politics of school closure.
Overall, I find no evidence that districts are significantly more likely to close heavily minority schools.[iii] Instead, school enrollment size is by far the most important factor. Average test scores matter somewhat, as well, but only on the margin. Unfortunately for those of us who wish that districts used closures strategically to improve academic outcomes, student achievement growth—the best measure of actual school quality—does not seem to affect district consolidation decisions at all, at least not for elementary and middle schools, for which we have available data.
Root causes
The evidence makes clear that the real story is not about districts intentionally targeting heavily minority schools for closure. Rather, it’s about heavily minority districts facing the greatest closure pressures. What explains these district-level differences?
Opponents of closure blame fiscal challenges—arguing that these districts are “chronically underfunded”—and a history of racist policies. Once again, however, such critics miss the mark.
Start with finances. It is true that, historically, a system of funding schools based on local property taxes created disturbing inequalities. Fortunately, that is not how we fund education in most states today. State governments have stepped up in recent decades, and state dollars tend to be distributed in a progressive way to compensate for differences in local tax bases. As a result, Black and Hispanic students now attend school districts that spend more per student compared to districts attended by white students in the same state. Communities facing some of the most serious building under-utilization issues and the most vocal opposition to closure—including Boston, Chicago, and New York—are also among some the highest spending.
As other evidence of racist policy contributing to closures, opponents also weave together complex webs of hypotheses and conspiracy theories that attempt to link modern enrollment losses to gentrification pressures. To be sure, the widespread shortage of affordable housing is real—but its connection to school closures tenuous at best. (And well-meaning efforts to increase minority political representation in local government seem to have made the housing shortages worse.)
Closure opponents ignore what is perhaps the most important factor—a long history of dysfunctional governance in heavily minority districts that continues to deprive students of color of equal educational opportunities. This dysfunction has contributed both to these districts’ enrollment declines and to the inability of school leaders to respond appropriately. In other research, we have shown that few voters who elect school boards in such communities have kids of their own and that these voters look nothing like the students being educated in local schools. Such communities also tend to have powerful teachers unions and highly prescriptive union contracts, including provisions that result in the most heavily minority and disadvantaged schools being assigned the least experienced teachers. During the pandemic, these districts were among the last to resume in-person learning, delays that exacerbated already large racial achievement gaps and produced substantial enrollment losses that have directly led to the current conversations about rightsizing.
These are all examples of outrageous racial injustices. But school closures are mere symptoms of the underlying political and governance challenges. Preventing closures does nothing to address any of the root causes or to resolve the many other issues that deprive low-income students and students of color of the education they deserve.
Lessons from Chicago
No city has attracted more attention in the school closure debates than Chicago, which went through several waves of consolidation over the past two decades. It appears as a case study in the new strategy document under the section titled “They Defeated School Closures, You Can Too!” Chicago is indeed instructive—as a cautionary tale about how misleading and cynical rhetoric and claims of racism can be weaponized in school closure debates.
Today, the Chicago school system is again suffering from serious excess capacity, and advocates again blame underfunding. “I’m no politician, but I do know that schools underutilized are schools underfunded,” a Chicago parent is quoted as saying in a book about school closures in that city, a quote that is repeated in the new strategy document.
Thanks to new federal requirements, we now have access to school-level expenditure data that definitively debunk such underfunding claims. Chicago’s most underenrolled high schools spent several times more per student than the district average last year, with half a dozen spending more than $40,000 per student, on par with the tuition in some of the country’s most exclusive private schools. The district’s smallest high school, Douglass Academy High School, enrolled only thirty students last year and cost the district nearly $70,000 per student to operate. Just keeping this single building open required the district to spend over $1 million per year more than it would have cost to educate its students at a different school—money that was diverted from other programming serving students district wide.
As an example of how they say racism has contributed to school enrollment losses and, ultimately, closure, activists in Chicago also point to the demolition of the city’s notorious high-rise public housing in the early 2000s. This is a bizarre case of revisionist history.
Chicago’s public housing projects were themselves a product of racist policies, federal courts found. The buildings were so violent and living conditions so horrendous that they were the setting of a 1990s horror movie and best-selling book. Most importantly, research has shown that the demolition clearly benefited the former residents, almost all racial minorities. These residents moved to better neighborhoods and their children went on to graduate from high school at higher rates, were less likely to be arrested, and more likely to be employed in early adulthood. Hardly evidence of racism.
Another oft-cited claim is that previous school closures in Chicago negatively affected academic outcomes of the displaced students. Again, that is not what the evidence actually shows. It is true that test scores and attendance rates of the affected students declined. However, these declines occurred the year the closures were announced—before a single building had actually been closed! The timing of the declines suggests that adult political conflict and efforts to keep schools open were responsible, not the closures themselves—something anti-closure advocates today should keep in mind.
Former Chicago superintendent Arne Duncan, who later served as President Obama’s secretary of education, has a very different memory of what happened in his hometown. In his book, Duncan recounts attending meetings of community members opposed to the closures he oversaw. At one such meeting, Duncan was confronted by an angry mother upset that her child’s school was on the chopping block.
After she accused Duncan of racism, the superintendent walked her through years of achievement data from the school—showing her how it had continued to fail generation after generation of students in her neighborhood. “If I were a racist, then I would leave this school exactly as it is,” the superintendent told her. “That’s not what I want. They’re children; I believe in them as much as you do. They can’t wait for things to improve any longer.”
This is an important point, one that often gets lost amid the disingenuous claims of racism that characterize many of the closure debates. Everyone who cares about racial justice should be focused on doing what’s best for students and their learning—not on school buildings or the employment impacts of closing them. Ensuring that districts prioritize academic considerations in their closure decisions—especially achievement growth over time—and that affected students receive spots at better schools should be our top priority, not keeping as many under-utilized buildings open as possible.
“Sometimes,” Duncan wrote in his book, “closing a school is the best thing that can happen to a group of students.” The best available research supports him.
[i] Although not discussed in the article, the authors’ full set of results actually show that school size was by the far the strongest predictor of school closure. Their estimates imply that increasing student enrollment by a standard deviation decreases the odds of closure by at least seven times more than the predicted effect of being a majority Black campus.
[ii] The authors do try to control for district demographics in a limited way, but this does not fully address the problem.
[iii] Like Pearman’s study, I used the official closure indicators in federal data. Tulane University economist Doug Harris recently released an alternative dataset manually identifying several thousand additional schools that may have been closed but not flagged as such in official statistics. Repeating the analysis using Harris’s data does show that majority Black elementary schools are significantly more likely to close—but the gap is orders of magnitude smaller than Pearman and his co-author found. There is also no difference among middle schools, and Hispanic student enrollment doesn’t predict closure for either type of school.
By now, we’re well familiar with critiques of standardized testing opponents: tests rob schools of critical instructional time, encourage teaching to the test, place undue pressure on students and educators to perform, are educationally irrelevant, only provide a snapshot of student achievement at a specific moment in time, and are largely driven by family income levels, parents’ education, and other non-school-related factors.
For decades, we’ve used standardized tests to measure student achievement and progress in U.S. schools. That’s because they offer a host of important benefits to students, parents, educators, and policymakers.
First, tests provide an essential source of information for students and parents about student learning, alongside grades and teacher feedback.
In reality, most parents and guardians aren’t all that well-attuned to how their child is doing academically. While the vast majority believe that their children are on track academically, that’s not actually true. The National Assessment of Educational Progress (NAEP), a.k.a. “The Nation’s Report Card,” reports that only about a third of all students are proficient in both reading and math. Test data help bridge this disconnect by providing parents and caregivers with important, objective information about their students’ academic achievement and progress (in addition to report card grades and teacher feedback). As the saying goes, “we can’t fix what we don’t see.” Just as I wouldn’t skip my child’s annual physical at the doctor’s office, I wouldn’t opt out of testing that provides important data about how my child is doing and progressing academically.
Test data also supply useful information for vetting school options, via state report cards or helpful sites such as GreatSchools (a website providing families with updated, transparent data on student test scores and academic progress at specific schools, among other information).
Second, test scores help counteract grade inflation in schools.
One reason that parents believe their children are on track in school is because their child’s report card says so. In contrast to objective test data, student grades can be more subjective and less related to content and grade level mastery, and grading is often uneven within and across schools. Grade inflation has long been a problem in schools, as evidenced by rising GPAs even as ACT and SAT scores, and other measures of academic performance, such as NAEP, have dropped or held steady. Grade inflation may look like it’s helping students by making them look better, but that’s an illusion: Students learn more from teachers with more rigorous grading standards. As my colleagues Meredith Coffey and Adam Tyner recently underscored, grading reforms “that water down expectations ultimately harm the students they are meant to help… Such policies tend to reduce expectations and accountability for students, hamstring teachers’ ability to manage their classrooms and motivate students, and confuse parents and other stakeholders who do not understand what grades have come to signify.”
Third, tests shed light on learning successes and gaps, and help teachers address students’ unique needs.
U.S. Secretary of Education Miguel Cardona once said that tests can be a “flashlight” on what works in education, and Virginia Secretary of Education Aimee Guidera has similarly noted that “actionable data is the foundation for improving performance and holding our schools accountable.” Alongside other indicators of student performance, tests provide teachers with actionable data that can help inform their instruction, such as whether students might need more help or acceleration in a certain subject or part of a subject. They indicate how well students are meeting grade-level standards and making progress over time, and shed light on which types of instruction or programs are effective in driving student learning. They can also provide a helpful data point for administrators about which teachers and schools are excelling at or struggling with helping students learn.
Fourth, state tests provide policymakers with consistent, comparable data about student learning statewide.
Standardized tests are the most reliable measures we have for gauging performance at the school level, shedding light on systemic inequities, and holding schools accountable for their academic performance. Correctly reported and analyzed, they show performance broken down by demographic subgroups (including race, English-learner status, and more), and can help direct support and resources to teachers, schools, and districts in need. For example, beginning in 2002, the No Child Left Behind (NCLB) Act shed a spotlight on academic progress, and particularly on outcomes for certain groups of students, including those from low-income families, English learners, students in special education, and students of color. As a result, student performance rose, particularly among younger children and traditionally disadvantaged populations.
Fifth, they’re an important indicator of college readiness.
While controversial of late, college entrance tests such as the SAT and ACT provide important information about schools and students that is distinct from the other elements of college applications. Research shows that the best predictions of a student’s success in college stem from a variety of measures, including admissions exams, grade point averages, personal essays, letters of recommendation, and more. Admissions tests are thoroughly vetted to ensure the exams do not discriminate against students’ race, ethnicity, class, or gender, and help higher education institutions identify students with high academic potential, especially those from underrepresented groups.
Sixth and finally, tests are also pretty good predicters of later life success.
There is considerable evidence that test scores are good predicters of later life outcomes, such as educational attainment, labor market outcomes, and earnings. Testing done right also incentivizes students to put in their best effort, demonstrate the knowledge they’ve gained, and work toward achieving educational goals. Testing need not foster fear and stress in children or adults, either. During testing season last spring, the principal at our son’s public charter school—a school that emphasizes academic achievement—stressed to teachers, parents, and kids that testing was simply a chance for students to demonstrate their knowledge and skills and show off their hard work. Teachers and principals are the primary communicators about why testing happens and how data are used, and this kind of positive messaging can go a long way in reducing stress and anxiety levels.
—
Tests are useful tools in the education toolkit, providing students, parents, educators, and policymakers with a full picture of student learning. Test data help us understand how individual students are doing; determine where gaps exist in terms of income, race, and geography; and identify effective educational models that help drive student learning. They are not some sort of time-consuming interruption of teaching and learning: One study found the average amount of time spent on mandated tests adds up to just over 2 percent of total school time. Yes, we can continue to work to keep testing time down, ensure state tests are high-quality and aligned to state standards, and work to ensure that any “test prep” aims at helping students master important content. But we shouldn’t throw the baby out with the bathwater.
Thankfully, the anti-testing movement has receded somewhat since its heyday almost a decade ago, and many top colleges have recently reinstated admissions tests, such as the SAT and ACT. No, standardized tests are not perfect, but they’ve improved greatly over time, and they’re certainly worth preserving, particularly as we’re still working to track and reverse the Covid-19 pandemic’s lingering disruptions on student learning. As my colleague Chester Finn wrote recently, “we have solid evidence over thirty years in America and beyond that students learn more when they—and their schools—are held to account for what and how well they’re learning. And while testing isn’t the only possible way to gauge this, nor is it a perfect way, it’s efficient, reasonably accurate, more objective than most of the alternatives, and relatively easy to explain and understand.”
So keep calm and test on!
According to a Goldman Sachs analysis of federal data, the college graduating class of 2024 is having a tough time finding attractive jobs. That seems counterintuitive, given that the national labor market is extremely tight, with unemployment around 4 percent. But that economic indicator masks a lot of variation, and the post-Covid-19 job market has been much stronger for lower-skill, lower-wage workers than for new college grads.
This is partly because some major employers—especially in the tech sector, which staffed up big time during the pandemic—probably overdid it, and now have to tighten their belts and shrink their payrolls. Meanwhile, the anticipation of an AI-driven revolution in staff productivity has some other employers thinking that maybe they can do without so many junior associates. Then there’s the age-old concern that recent grads aren’t worth hiring, now exacerbated by pandemic-induced skill gaps.
If history is a guide, despite whatever kinks we see in the hiring system right now, newly minted college graduates will turn out to do just fine, given that their skills continue to bring a significant wage premium. But in the short term, at least, the fact that tens of thousands of talented young people are struggling to find lucrative first jobs is great news for employers that need talent, have flexible job requirements, and pay decent wages for new graduates.
Guess who fits that description? America’s public schools.
This is a once-in-a-generation opportunity for our education system to hire people who otherwise might not give teaching a serious look. The upsides are significant. Research by Harvard’s Martin West found that teachers hired during recessions, when schools have greater access to talented people, tend to outperform their peers in terms of boosting student learning. But schools can only seize the day if they are willing to do things differently.
Teaching careers are more lucrative than you might think
It might surprise some readers that our public schools could be an attractive option for underemployed or unemployed college graduates. Isn’t it a lot of work for very little pay? But teacher salaries in most states can be higher than many of us think. In 2021, the journal Education Next asked a representative sample of 1,410 Americans to guess how much their local districts paid teachers, and they lowballed the real number by 50 percent, or more than $20,000. As of the 2022–23 school year, the average teacher salary had reached almost $70,000. In California, New York, and Massachusetts, it exceeded $90,000.
No doubt, teaching for a lifetime doesn’t offer as generous of a salary as many other professions do, given that top salaries rarely reach six figures. But someone who wants to teach for a few years in their twenties can earn a comfortable living, including summers off and great health benefits.
The problem, alas, is that many public-school systems aren’t hiring. Widespread teacher shortages have quickly turned into widespread reductions in force. That’s because, like those tech companies, schools were able to staff up big time during the pandemic, when they were flush with federal pandemic relief aid. That money is now mostly gone. So they need to shrink their headcounts, and most are doing so either through attrition or, if necessary, layoffs via last in, first out (LIFO) policies—rather than letting go of their least effective employees.
Schools will say that’s impossible thanks to tenure protections that have been won by all-powerful teachers unions. But teachers brought on during the hiring surge of the past few years don’t yet enjoy those protections, and those who have been unsuccessful in the classroom could be let go, making room for new talent, including recent college graduates. It just takes courageous local leadership to do it.
Unfortunately, given their unwillingness to rock the boat, many systems will likely take the path of least resistance instead. That leaves charter schools, which are well positioned to recruit from this larger talent pool, given that most are not unionized and don’t have to worry about tenure or LIFO. To be sure, those schools are dealing with the end of federal relief aid, too. But one reason charters tend to be so effective is that they enjoy greater flexibility over their budgets and staff. In this case, they can much more easily replace low-performing teachers with great new instructors.
Most charters can also hire noncertified or alternatively certified teachers, meaning they can bring in people who majored in an academic subject like math or English, give them some short-term training to get them up to speed on teaching methods, support them once they are in the classroom, and help them work toward getting their teaching license over the next couple of years.
Everyone would benefit from aggressive efforts to get talented college grads into America’s classrooms—our schools, the graduates, and especially the students. Now it’s up to our educational leaders to make it happen.
Editor’s note: This was first published by Forbes.
School closures and remote learning led to widespread relaxation of student accountability at the onset of the coronavirus pandemic. Lax requirements to turn in work, fewer graded assignments, and—most perniciously—policies mandating “no zeros” or “no failing grades” were adopted (or accelerated) to lighten the load of young people whose worlds had been turned upside down. Now that school schedules have returned to normal, a descriptive study from CALDER examines whether grading policies have begun to do the same, too.
Authors Dan Goldhaber and Maia Goodman Young examine evidence of grade inflation before and after Covid-19 in Washington State. In March 2020, Washington’s Office of Superintendent of Public Instruction (OSPI) advised schools to consider the adoption of “pass/no credit” grading policies. A month later, as it became clear that closures would be a long-term proposition, OSPI revised its guidance to specify that high school students (and middle schoolers taking high-school-credit-bearing classes) could not be given pass/no credit marks, but could be awarded grades of “incomplete,” with the hope that they would complete their coursework in the future. The bottom line, however, was that no student could receive a lower final grade than whatever they had already earned when schools shut down in March.
Come fall of the 2020–21 school year, when some Evergreen State districts switched to in-person or hybrid learning modes, teachers were allowed to assign failing grades again, but the flexibility of the directive led districts to implement a range of policies. Leaders in Seattle, for instance, instructed teachers to keep grades within an A-to-C-minus range (or incomplete).
The study examines the relationship between test scores and grades and how grades and average GPAs have shifted over time. The analysis uses student-level data—specifically, test scores and grades earned in high-school-credit-bearing courses in English, math, and science for students from the 2010–11 through 2021–22 school years, including middle schoolers who took courses for high school credit—primarily Algebra 1 and geometry. While the covered timeframe is just barely “post”-pandemic, it will be important later to compare these early trends to future ones.
Prior to the pandemic, the results show a slight uptick in math and English grades. But consistent with state guidance, almost no students received an F grade in the spring of 2020. Failing grades dropped from 9.5 percent to 1.3 percent in English courses, 9 percent to 1 percent in math courses, and 8.6 percent to 1.1 percent in science between the fall and spring semesters of 2020. The distribution of grades higher than F mostly increased in that time, with the share of A’s specifically jumping from 33 percent to 56 percent in math; from 36 percent to 60 percent in English; and from 33 percent to 59 percent in science. From fall to spring 2020, the average GPA in math jumped from 2.6 to 3.2, in English from 2.7 to 3.3, and in science from 2.6 to 3.3.
Fast forward a single year, however, and signs of stability emerged. By 2021–22, English and science grades had largely returned to pre-pandemic levels, though math grades remained high. (The average math GPA in 2021–22 was 2.7, 0.4 points higher than it was in 2018–19.)
As for shifts in the relationship between grades and test scores, Goldhaber and Young predict students’ placement in the test distribution based on letter grades received in respective years. They find that their position drops for all math subjects between the first year of observed test scores (2015–16) and the first year of testing after the pandemic (2021–22). In other words, a student who got an A in Algebra 1 was predicted to be in the 73rd percentile of the test distribution in 2015–16, the 68th percentile in 2018–19, and the 58th percentile in 2021–22. In Algebra 2, a student receiving an A was predicted to be in the 64th percentile in 2015–16, the 58th percentile in 2018–19, and the 54th percentile in 2021–22. But these predictions do not hold true in all subjects. In geometry, for instance, the shifts in grading standards mostly stabilized, and in English they were not as big as in math.
The report’s conclusion draws attention to the misalignment between grades and achievement in Algebra 1, given that course is a gateway class to higher math courses in high school. If students aren’t properly identified for help at this juncture, they may not succeed in those more challenging classes.
It’s a huge problem when students who earn top grades fail to demonstrate mastery of key skills and knowledge on a state test. But the need to right-size grades is about more than wrong signals. There’s growing alarm that an extended hiatus from grade-based accountability is a surefire way to “lock in” pandemic learning loss for the long term. That’s a sobering thought, given more-recent evidence on learning loss.
As educators and policy leaders finally take seriously the harm that cellphones do to learning, it’s fair to ask what it will take for them to recognize the same about grade inflation?
SOURCE: Dan Goldhaber and Maia Goodman Young, “Course Grades as a Signal of Student Achievement: Evidence on Grade Inflation Before and After COVID-19,” Center for Analysis of Longitudinal Data in Education Research—CALDER (November 2023).
Cheers
Jeers
On this week’s Education Gadfly Show podcast, Vlad Kogan, a professor at Ohio State University, joins Mike and David to discuss what role race, achievement, and enrollment play in a district’s decision to close a school. Then, on the Research Minute, Amber examines a new report that investigates the staffing difficulties and potential academic effects of class size reduction policies in New York City.
Recommended content:
Feedback Welcome: Have ideas for improving our podcast? Send them to Stephanie Distler at [email protected].
