States are the “laboratory” of democracy, opined Supreme Court Justice Louis Brandeis. That laboratory swung into action when states introduced substantially different policies as Covid-19 swept across the country in early 2020. In some places, schools closed their buildings and switched to remote learning for a year or more, while in others the school door reopened by the start of the fall. Surprisingly, school closures and other social distancing policies often had more to do with the political coloration of a state or district than with the risk that children and teachers would suffer severely from the virus.

The politicization of school policy proved unfortunate for children. Yet the fact that pandemic mitigation measures were heavily shaped by politics allows one to obtain a rough estimate of their effects on student learning simply by comparing trends in states with varying hues—red, blue, or purple.

Here we use data from the National Assessment of Educational Progress, or NAEP, to compare the pandemic-era learning loss across states grouped by the share of the two-party vote that went for Kamala Harris in the 2024 presidential election. We divide states into three near-equal categories (see table) and give them the conventional colors for Republican, Democratic, and swing states. We consider trends in student performance from spring 2019 to 2024 on each of the four tests that NAEP regularly administers: 4th- and 8th-grade reading and math. We take into account variation in student demographics across states and over time by relying on adjusted NAEP scores published by Matthew Chingos and Kristen Blagg at the Urban Institute.

Findings

Student reading and math performance nationwide in 2024 remains substantially below pre-pandemic levels in both 4th and 8th grade. The same is true in all three groups of states. But the downward slope is considerably steeper for students residing in blue states than for those in red states, with those in purple states falling somewhere in between (see figure). Each of the reported differences between red and blue states in the slope of the trend from 2019 to 2024 is statistically significant.

In 2019, the average reading scores of 4th-grade students after adjusting for demographics are highest in the deep blue states, lowest in the deep red states, with purple states hovering in the middle. By 2024, the reverse is true. Red states rank highest, blue states lowest.

In 8th-grade reading, students in blue states in 2019 have a narrow edge over those in purple states and a sizeable one over red-state students. By 2024, the three sets of states are essentially tied.

In 4th-grade math, students in purple states score higher in 2019 than those in red and blue states, which are essentially tied. All three sets of states take a terrible tumble in 2022, then recover somewhat in 2024, at which point red states catch up with (and barely surpass) purple states. Blue states trail well behind the other two groups in the most recent data.

In 8th-grade math, purple states begin in 2019 with the highest average score, holding a narrow advantage over blue states and a larger one over red states. All three sets of states suffer a severe drop between 2019 and 2022. By 2024, the blue and purple states continue to slide downward, while red states recover somewhat. By this point red and purple states are essentially tied, well ahead of blue states.

Possible explanations

The differential extent of school closures during the pandemic is the most likely explanation for the differential achievement trends across states. Numerous studies show greater learning loss in places that closed schools for longer periods of time. However, we cannot rule out the possibility that the differences across states also reflect other pandemic-era regulations both in schools and in society more generally. Social and emotional distress registered during the Covid years could have been driven in part by general social isolation and have had its own adverse impact on learning.

Collective bargaining arrangements may help to explain both the extent of school closures and their consequences. In some blue states, schools could not open without the approval of teachers unions, and in some left-leaning areas, unions threatened strikes unless time-consuming school improvements—air filters, larger spaces to facilitate distancing, and so forth—were implemented. Also, teacher absenteeism may have been facilitated by generous sickness and personal leave policies negotiated by unions. Right-to-work laws in many deep red states may limit the ability of unionized employees to restrict school opening and related measures.

School choice ranks among the possible, but less likely, explanations for red-blue state differences. Public schools have suffered sizeable enrollment declines since the pandemic, homeschooling is on the rise, and several red states have recently passed laws that help finance access to the private sector. Public schools were more likely open when they faced greater competition from Catholic schools, and the wave of school choice laws that spread across red states after 2018 may have enhanced public schools’ responsiveness to demand for open schools. On the other hand, the choice sector remains a small part of the education landscape even in red states, and state-level NAEP data exclude students attending private schools. More fine-grained analyses may discern whether school choice affected rates of learning loss.

Greater federal funding for southern schools is another possible, if less likely, explanation. Red states receive more federal education dollars than other states, on average, because they have more low-income students eligible for the compensatory education program. The $185 billion increase in federal education funding during the pandemic was administered according to the same formula, likely resulting in a disproportionate allocation to red states. However, recent research shows that the effects of these additional funds on post-pandemic student performance are too small to account for the substantial differences between the rates of decline in red and blue states.

The “science of reading” has exploded in popularity since the pandemic. The “Mississippi miracle” and recent gains in Louisiana may both be attributable in some degree to a return to an old-fashioned approach to instruction that may be more popular in Trump states than Harris states. On the other hand, science-of-reading laws have not evoked partisan debate and have become popular across the country, even in New York City. And, in our data, it is in math rather than reading that 4th-grade students in red states make the largest relative gains.

It is tempting to attribute academic decline to students spending ever more time on smartphones, tablets, and other tech gadgets. Perhaps it is the students in the wealthier blue states that are especially likely to buy and use these devices. We doubt it. The cost of access to screens of all kinds has dwindled, and students everywhere are enticed to use them to connect to peers. Still, we cannot rule out the possibility that tech and social media are also part of the story.

Conclusion

The high-level results we have shared are only a starting point for understanding the effects of state policy choices on student learning. We focus on math and reading achievement in just two grade levels, as these are the only domains for which comparable data are available across states. We do not break students into subgroups, because the Urban Institute only reports demographically adjusted trends for all students. When scholars can draw upon data from districts, schools, and individual students, they will be able to test with greater precision the extent to which differential state policies contributed to the steep decline in student math and reading performance documented by the NAEP.

Some might wish to celebrate the decline in long-standing geographical disparities in achievement between red and blue states. But most will be alarmed that so many U.S. students suffered such a severe decline from the pandemic, producing losses that remain sizeable four years after the Covid outbreak in math—and in reading now show signs of further slippage.

If our results are validated by further research, the policy implications are immense. The risk of pandemics remains great. Should such a dreadful event occur again in the future, governments will, we hope, realize they need to balance short-term public health consequences against long-term educational ones.

Michael Hartney is an associate professor of political science at Boston College, the Bruni Family Fellow at Stanford University’s Hoover Institution, and a Senior Fellow at the Manhattan Institute. Paul E. Peterson is the Director of the Program on Education Policy and Governance and the Henry Lee Shattuck Professor of Government at Harvard University, and a Senior Fellow at the Hoover Institution.