Published on

It Was Never COVID's Fault β€” U.S. Students Have Been in a 'Learning Recession' Since 2013

On May 13, 2026, researchers from Stanford, Harvard, and Dartmouth released a study that reframes a decade of debate about American education. Analyzing test score data from approximately 70 million students across more than 5,000 school districts in 38 states, they found that the decline in U.S. student achievement did not begin with the COVID-19 pandemic. It began in 2013 β€” seven years earlier.

Table of Contents

  1. Seven Years of Erosion Before the Pandemic
  2. The Reading Recession β€” Scores Down in 83% of Districts
  3. Two Culprits: Social Media and the End of Accountability
  4. States That Found a Way Back
  5. What This Means Beyond America

1. Seven Years of Erosion Before the Pandemic

The most quoted line from this study belongs to Harvard professor Tom Kane: "The pandemic was the mudslide that followed seven years of erosion in student achievement."

The research team built on the Stanford Education Data Archive (SEDA), a massive database linking test results for roughly 70 million students in grades 3 through 8, covering 2009 to 2025. By mapping results onto a common national scale, they could compare progress β€” or its absence β€” across different states, districts, and years.

The picture that emerged was unsettling. Math and reading scores began declining steadily around 2013. By the time the pandemic hit in 2020, scores had already fallen significantly. The pandemic years added another shock, but the foundation had been crumbling long before.

2. The Reading Recession β€” Scores Down in 83% of Districts

The numbers are stark. Reading scores declined in 83% of school districts where data was available. Math scores fell in 70% of districts compared to the prior decade.

Researchers labeled this a "reading recession" β€” a term that captures something more than a dip in scores. It signals a structural shift: reading is becoming less central to children's everyday lives. The drop in reading scores in the years just before the pandemic was as steep as the drop during it.

By 2025, only five states plus Washington D.C. showed meaningful growth in reading scores from 2022 onward. In math, nearly every state showed some improvement β€” but reading remained stubbornly resistant to recovery.

3. Two Culprits: Social Media and the End of Accountability

The research team identified two major shifts that coincided with the 2013 turning point.

The first was the dismantling of test-based accountability. The No Child Left Behind Act of 2001, which tied school funding and consequences to standardized test performance, was replaced in 2015 by the Every Student Succeeds Act β€” a law that gave states far more flexibility and significantly softened accountability mechanisms. Researchers point to this transition as a likely contributor to the academic slowdown.

The second was the rise of social media. Around 2013, smartphones became ubiquitous among teenagers, and image-driven platforms like Instagram and Snapchat exploded in popularity. Social psychologists β€” most notably Jonathan Haidt β€” have long argued that heavy social media use correlates with reduced reading time and shorter attention spans. This study's timeline aligns with that claim.

The relative weight of each factor remains an open question. But the timing of both shifts overlapping so precisely with the start of the decline is difficult to dismiss.

4. States That Found a Way Back

Among the bleak data, there are genuine bright spots. States that saw meaningful improvements in reading β€” Louisiana, Maryland, Tennessee, Kentucky, and Indiana β€” share one thing in common: they mandated the adoption of phonics-based literacy instruction, known as the "science of reading."

This approach, which emphasizes phonemic awareness and letter-sound connections, had fallen out of fashion in many districts. Its comeback is part of a broader evidence-based education movement β€” the idea that teaching methods should be tested and validated before being scaled up. The results suggest the approach works.

In math, the picture is more broadly encouraging. Almost every state in the analysis showed gains from 2022 to 2025, suggesting that post-pandemic math recovery is more tractable than reading recovery.

5. What This Means Beyond America

This study is not just an American story. OECD PISA data shows similar declining trends across multiple countries over the same period. South Korea, for example, has experienced falling reading scores alongside rising smartphone use β€” a pattern that rhymes with the U.S. findings.

The deeper challenge this research poses is a framing problem. Policymakers around the world are currently focused on "pandemic learning loss recovery." But if the decline started before the pandemic and for reasons unrelated to it, then recovery-oriented policies may be addressing the symptom rather than the disease.

The pandemic was the headline. The real story started a decade earlier.

Sources

Discuss on Twitter β€’ View on GitHub
It Was Never COVID's Fault β€” U.S. Students Have Been in a 'Learning Recession' Since 2013 | MINSSAM.COM