This week, the Center for Education Policy Research (CEPR) at Harvard University and the Educational Opportunity Project (EOP) at Stanford University released the Education Scorecard, which uses data from the National Assessment of Educational Progress and state assessments to compare reading and math scores at the district level in 40 states from 2015 to 2025. It’s a sophisticated methodology that provides an important, if gloomy, reflection of student achievement from before, during, and after the pandemic.
A primary goal of the report is to push Americans to think beyond pandemic learning loss. To put a finer point on it, the researchers renamed this year’s report: Previous versions were called the Education Recovery Scorecard, but the authors removed the word “Recovery” from the title to emphasize the “learning recession” that began well before the pandemic.
This learning slump is most glaring in reading comprehension. Between 2017 and 2019, reading scores declined at the same rate as they did during the pandemic years. That is worth restating: Reading declines in the two years before the pandemic were comparable to those during the pandemic itself. Worse, between 2022 and 2024, reading continued to decline at roughly the same rate. This year’s scorecard shows the beginnings of reading recovery between 2024 and 2025, but that’s where the good news ends: From 2015 to 2025, 83 percent of districts lost ground on reading, and one in three districts fell behind by a whole grade level. That is indeed the stuff of a learning recession.
The pattern in math was different. The pandemic was a much bigger driver of losses in math, and recovery in math was evident immediately after the pandemic. Even still, about 70 percent of districts scored lower in 2025 than in 2015.
The report is full of useful information and has been covered extensively in the media. The New York Times, for instance, published an extensive feature alongside an interactive tool where you can look up your own district’s achievement trajectory. I recommend reading the whole report, but I’ll briefly highlight three important observations the researchers make.
The first of these is the shape of the recovery since the pandemic. We are accustomed to thinking that low-income districts take the brunt of education downturns, but the scorecard finds a “U-shaped” recovery by income: The highest- and lowest-income districts improved more than middle-income districts, which saw the weakest average gains from 2022 to 2025. The authors credit high-poverty districts’ recovery advantage to the substantial (and disproportionate) federal pandemic relief funding they received, without which their recovery would have been much lower. The bad news is that federal pandemic relief funds dried up by 2025, leaving open the question of how high-poverty districts will fare moving forward. Regardless, concern about high-poverty districts should not distract from the especially slow rebound in middle-income districts.
A second key point is the beginnings of a reading resurgence seen in 2025. The authors tentatively attribute that turnaround to states’ “science of reading” reform packages. Not all states that implemented a science of reading reform before 2024 saw reading improvements, but those that did see improvements had all implemented these instructional overhauls. In contrast, none of the states that had not implemented science of reading reforms saw their reading scores rise between 2022 and 2025. This suggests that science of reading laws may be effective, maybe necessary, but they don’t appear to be a magic bullet for juicing reading achievement.
The report also highlights post-pandemic student absenteeism. The report uses my Return to Learn chronic absenteeism data and finds recovery would have been meaningfully larger across the board if chronic absenteeism rates had not skyrocketed during the pandemic. The good news here is that chronic absenteeism rates have been improving, if slowly. More concerning, the rate of improvement across the nation is steadily slowing. At the current pace of improvement, US schools are unlikely ever to return to pre-pandemic absenteeism rates without substantial and sustained intervention.
Like the scorecard authors, I have been working on these issues for some time, and I am alarmed at what the data are showing. As I told The New York Times, “I cannot be more emphatic: This is an enormous problem that’s not getting enough attention.”
Another year of data shows some signs for hope, but the Education Scorecard puts the focus squarely where it needs to be. There has been a dramatic decline in student learning since 2013. It started well before the pandemic, and it is both deep and widespread. This is the central problem in American education. Schools, families, and policymakers ought to pay it far more attention.
