Skip to content

Raw data: Black learning loss on the 2023 TIMSS mathematics test

Here's an odd thing. While I was poking around in the TIMSS mathematics results, I took a look at the size of the decline in test scores during the pandemic:

Scores among Black students barely dropped at all. This conflicts with every other analysis of test scores I'm familiar with, which unanimously find that Black students (and low-income students in general) had the biggest learning declines during the pandemic.

What's even weirder is that results for 4th graders were exactly the opposite:

This is sort of inexplicable. How could Black 4th graders show big learning losses while 8th graders showed none? It's damn peculiar.

10 thoughts on “Raw data: Black learning loss on the 2023 TIMSS mathematics test

  1. Solar

    Since this is measuring decline from a baseline, the question is what were those baselines?

    If 8th graders were already doing very poorly to begin with, their decline would be minimal since there is not much room for further decline. Also, if I'm not mistaken the gaps normally seen between the various racial groups (which really are a proxy for income) tend to be smaller at younger years and then widen as they move through higher school years, which would also explain the big drop in the 4th year group.

  2. somebody123

    Because it’s all just noise. Racial categories are arbitrary and this isn’t controlled for income or parents’ educational level, which are hugely important to this kind of analysis. It’s basically a random number generator.

    1. memyselfandi

      Not if you are sampling a large enough and truly representative sample. You don't want to correct for income, because that would simply hide the fact that we are doing a lousy job of educating our minorities.

  3. Scott_F

    We have to make certain that the identification of students by race has not changed over the comparison period. Black vs Two Races may be fluid and shifting in one direction at a larger scale. These shifts, might also be cultural specific to groups - i.e. a wealthier cohort might shift in one direction and a poorer group show an opposite trend

  4. JohnH

    "This conflicts with every other analysis of test scores I'm familiar with, which unanimously find ..." I'd stop reading there and not ask us to puzzle over it. It doesn't sound "peculiar"so much as in dire need of credibility, especially when their own data cited here in the same post could almost contradict it (well, not logically, but suspiciously).

  5. cephalopod

    If it is real, I'm guessing it's partly differences in test refusal by parents and the behavioral issues that seem to plague the youngest kids post-pandemic. Parents who opt out of standardized tests are not evenly distributed. And if you talk to anyone in early childhood ed, they seem to see much less school readiness than before.

    Those 8th graders missed having a normal 4th/5th grade. That isn't a particularly intense period for new math concepts.

  6. jdubs

    Test prep is the least valuable part of the US education system and analysis of standardized test results by themselves have never been useful in drawing meaningful conclusions.

    Garbage in, garbage out.

Comments are closed.