Results from an international math test that many of us often forget exists came out last week. TIMSS is an international math and science assessment given every four years to 4th and 8th grade students in more than 50 countries. Disheartening, troubling, disappointing—take your pick of dismal adjectives, and you’re getting the vibes in the education world from this new data. US scores declined for the third time in a row (this has been well-covered, for example here and here, along with every other news outlet last week).
I dug into the data for myself to see if there’s more to learn about how we might start solving our national math problem. The good news: there’s some reasons to feel optimistic. The bad news: without the sense of urgency we all had in 2021 that student learning was in crisis (and $120 billion), I’m worried we won’t get far.
Our primary concern should be the students at the bottom of the achievement distribution. In 2011, just 4% of 4th-grade students were below even the lowest proficiency bar on the TIMSS test. Now it’s 13%. For 8th grade, the jump is from 8% in 2011 to 18% now. More than double the number of American students are so behind in math they can’t even touch the lowest proficiency level. (Read Kevin Mahnken at the 74 and Nat Malkus at AEI on this too.)
The gap is also clear between schools with more affluent students and those with more disadvantaged students. In fact, elementary schools reporting more affluence actually improved.1 In 2019, 4th graders in affluent schools scored 574; this time, it was 576. Good job affluent schools!
But in 2019, 4th graders in schools with mostly disadvantaged kids scored 516. In 2023, the score dropped to 484. Students in low-income communities are doing much, much worse. The gap between affluent and low-income increased by almost 60 percent. Let’s look at that gap again:
2019: Affluent 4th grade schools, 574. Low-income 4th grade schools, 516.
Difference: 58.
2023: Affluent 4th grade schools, 576. Low-income 4th grade schools, 484.
Difference: 92.
Fourth grade students in the spring of 2023, when TIMSS was given, were in first grade when COVID started. These students, who fared so poorly on this assessment, are now in sixth-grade. I wrote recently about a middle school here in DC where the principal said that the entering sixth-grade students came in with the lowest academic levels she’s ever seen. Well, here is that trend playing out nationally.
The test also asks students to self-report their school attendance patterns. Twenty percent of 8th grade students said they missed at least one day every two weeks. Unsurprisingly, students who missed the most school had the lowest scores. Students who said they miss one day every week (7%) had an average score of 414, almost two years worth of learning below the US 8th grade average.2 Students who say they almost never miss school scored ten points above the national average. Parents: send your kids to school every day.
If there’s one place to look for hope, it’s in the content area scores. The test has questions in three categories of math content—numbers, measurement and geometry, and data. America’s highest-scoring area was numbers, where our 4th-grade students fared better than peers in Denmark, Finland, and Australia. Our problem was measurement, where the average US score was almost twenty points below our overall average. That means that 4th grade students are missing close to a year’s worth of knowledge in this content area. But it could also mean that it’s possible to identify clear knowledge gaps, and work to fill those gaps. We have to believe it’s not too late to teach today’s sixth-grade students foundational math skills.
Similarly in 8th grade, it’s clear that geometry is a significant challenge. Meanwhile, American 8th grade students actually do better on algebra questions than students in Norway and Finland. (It’s worth noting that the 8th-grade students who took this test are now in 10th grade, and we are running out of time to teach them more before they leave K-12 education.)
Now, I don’t have access to content area scores by affluence, or with trend data over time, so my conclusions are certainly more suggestive than prescriptive. We do have a math problem—a very, very big one. But we can see from the data that there are some obvious places to start. Let’s revisit curriculum and instruction on measurement and geometry. Let’s keep funding and energy going for high-impact tutoring to support students in low-income schools who are mostly likely to be behind. Let’s look for examples and best practices from these 600 districts where 8th-graders outperform expectations. Let’s get kids in school every day. I don’t want to do the math on what it means for America’s society and economy if we do nothing.
TIMSS defines a “more affluent” school as one in which at least 25% of students are from affluent households, and less than 25% are from low-income households. Then there are schools that are “neither affluent nor disadvantaged”, and then “more disadvantaged” schools, with the inverse proportions of affluence, or at least 25% from low-income households and less than 25% of students from affluent households. The numbers are reported by principals.
On this test, 30 points is considered roughly one year’s worth of learning. But let’s emphasize the roughly.