There’s Something Wrong With the TIMSS Advanced Math Test
Mother Jones
<!DOCTYPE html PUBLIC “-//W3C//DTD HTML 4.0 Transitional//EN” “http://www.w3.org/TR/REC-html40/loose.dtd”>
Excellent news! The 2015 TIMSS test results are out. This is one of two international math tests for 4th and 8th graders (the other is PISA), and it provides us with yet another chance to bemoan the shoddy education of American students.
I’ll get to that later tonight. First, though, I want to point out an odd thing about the TIMSS test. This year, for only the second time, they decided to add a third “advanced” math test for high school seniors who were in advanced math courses. Eight countries participated, and the United States did pretty well. We lagged behind only Lebanon.
Lebanon? You bet: their average score was 532, a whopping 50 points ahead of the two second-place countries (Russia and the US). But then I noticed something: only 3.2 percent of Lebanese students were in advanced math courses compared to 34 percent of Slovenian students. It makes sense that if you compare the top 3.2 percent of one country to the top 34 percent of another, the former is going to do a lot better.
So are differences in these scores just due to differences in how selective different countries are in accepting students into advanced math courses? Here’s the scatterplot you’ve been waiting for:
Selectivity doesn’t account for everything, but it does have a significant impact. If you restrict your classes to only the very brightest students (like Lebanon, Russia, and the US), they’ll do well. If you open them up to more than a quarter of your students (like Italy, Portugal and Slovenia), the average kids will drag down the mean score. But which country is actually doing a better job of education? It’s hard to say.
Regardless, there’s always something to complain about. Here is Jeffrey Mervis in Science:
Students taking the most challenging math and science courses in their senior year were found to have performed progressively worse as they moved from elementary to middle to high school. The U.S. cohort, for example… deteriorated over time, from 29 and 9 points ahead of the midpoint in fourth and eighth grade, respectively, to 15 points below as seniors. Italy recorded the steepest drops, a startling 126 points below the midpoint in physics and 78 points in advanced math by the end of high school.
It’s not clear to me that the “midpoint” of the TIMSS test means anything at all. In the advanced math test, every single country except Lebanon scored below it. What kind of midpoint is that? A pretty arbitrary one, I’d guess.
See original article: