The Same Score, A Different Childhood

On June 10, the National Center for Education Statistics released the long-term trend results for nine- and thirteen-year-olds in reading and mathematics.¹ The headlines tell two stories at once. The nine-year-olds are quietly gaining back what the pandemic took. Their reading score rose three points to 218 and their math score climbed four points from 234 to 238, both statistically significant moves.¹ The thirteen-year-olds did not move. Reading sat at 256, the same as in 2023. Math came in at 270, one point lower.²

The harder sentence in the report has nothing to do with this year. It is that reading achievement for today's thirteens has effectively flatlined at the level it held in 1971, the year the long-term trend assessment first ran.³ Math, after climbing for thirty years, has slipped back to roughly where it sat in the early 1990s.³ Two children, a half-century apart, lifting their pencils in two different rooms and circling roughly the same answers.

What the Instrument Was Built For

The long-term trend assessment was designed to do one thing well. It holds its items constant across decades, so that a score in 1971 and a score in 2025 can be read as comparable. That stability is what makes the headline possible. It is also what makes the headline misleading.

A score the same as fifty-four years ago says nothing about how the thirteen-year-old in 2025 actually got there. The fifth-grade year she sat through during the COVID spring was not the fifth-grade year her grandfather sat through. The book she did not pick up at home was not the book his older brother might have read aloud. Fourteen percent of her cohort now say they read for fun on a daily basis, the same share as in 2023 and roughly a third of what 1984's thirteens reported.⁴ The same number, in different rooms, made by different days.

The Press Conference Sentence

Matthew Soldner, acting director of the Institute of Education Sciences, said the most careful thing anyone has said about the report so far. The pandemic, he conceded, complicated the picture for the thirteens. "However, we can clearly see that this isn't just a pandemic story."²

That single concession is the place to dwell. The cohort now testing at the same level as their grandparents was halfway through second grade when the schools closed. They have lived the rest of their elementary years under teachers worried about screens, debates about phones, conversations about cheating, and most recently the arrival of a generation of chat tools that will write the paragraph for them in two seconds. None of those texture-of-the-year facts can be read off a score that holds its items constant.

What the Number Cannot Carry

This is the limit every annual instrument runs into. A score taken every two years can tell a country that something happened. It cannot tell the country what. The drop in reading for fun is the clearest secondary signal in the report.⁴ It suggests, at minimum, that a thirteen-year-old in 2025 lives with words differently than her grandparents did. But it is still a survey item, asked once a year, after the year is over. It is not the record of what she actually did on a Tuesday in October.

The thing schools do not yet have at scale is the Tuesday record. There is no widely deployed instrument that captures whether a student paused for sixty seconds before her next sentence, whether she rewrote her opening twice, whether she asked the chatbot and then asked it again differently, whether she opened the book and closed it after a paragraph. Without those, the score in May or June arrives as a verdict on a year nobody could quite see while it was happening.

The Year That Could Be Seen

The most useful response to the long-term trend results is not alarm at the flat line. It is to ask, quietly, what would have to be true for the next decade's line to bend. The answer, in the schools where teachers are most honest about it, is the same in every staff room. We need to see the work happen, not just the year close.

That is a harder ask than any reform agenda has yet absorbed. It is not a curriculum change or a new ban on devices. It is a change in what the school chooses to keep a record of. A draft instead of a final paragraph. A revision history instead of a sentence count. A pause logged instead of a percentile. A conversation with the model captured and read instead of erased after the page closes. A learner whose practice, not just her performance, is visible to the adult in the room.

None of that, by itself, will move the flat line. But the line will not bend if the only instrument honest enough to measure it is the one designed for stability across decades. The thirteen-year-old who scored a 256 last year is a real child who, at some point during a real year, did something or did not. The score, however careful, cannot say which.

If we measured a year by what the student actually did inside it, instead of what she could recall on a single morning in May, would the report card read the same?