The math gap

I’ve written a lot about the gender gaps in education, and it was the main theme of my TED talk. For a long time, my basic view was that girls were ahead of boys in reading and literacy, but boys were ahead of girls in math, and that these two sort of balance each other out. But that’s not true. In a previous post I showed the OECD data from PISA showing that there’s basically no gap in math at 15, and a big gap favoring girls in reading.

In the U.S., I’ve been struck by the fact that especially in poorer areas, girls have pulled ahead of boys in math in the earlier grades. Here’s a chart showing the change in the gender gap in 8th grade math by state, from 2009 to 2018, where blue indicates that boys are ahead and pink indicates that girls are ahead (with apologies for the lazy color-stereotyping):

Sean Reardon and his colleagues at the Stanford Center for Education Policy Analysis show that at the school district level (they study 10,000), there’s a big gender gap favoring girls in English Language Arts (ELA) in elementary and middle school and essentially no gender gap in math. As they write:

By eighth grade, in the average district male and female students score equally well on math tests, but females are a nearly a grade level ahead of their male classmates in ELA.

If you’d assumed, like me, that the English gap one way, and the math gap the other way, sort of balanced each other out, these findings are counter-intuitive and striking. As the CEPA team write:

Although the general absence of large, male-favoring math gaps would seem at odds with public perceptions of math gender achievement gaps, these patterns generally align with those estimated at the state or national level in other studies.

Class and race count

Reardon and team also examine the impact of race and socio economic indicators on the gender gap in school districts. There is no variation between rich and poor districts in the gender gap in ELA. But there is in math:

Reardon, S.F., Fahle, E.M., Kalogrides, D., Podolsky, A., & Zárate, R.C. (2018). Gender Achievement Gaps in U.S. School Districts (CEPA Working Paper No.18-13).

Boys are way behind girls in ELA everywhere. But boys in affluent areas are slightly ahead of girls in math, while boys in poorer areas are slightly behind girls in math. So while there’s a little of that “balancing out” in richer places, the opposite is true in poorer ones.

What about race? Districts with more Black students have larger girl-favoring gaps, and this association remains even after controlling for socio economic variables:

This is a key point in the discussions of gender gaps in education, especially in grade school. It is lower-income boys and Black boys who are the greatest disadvantage.

New Chicago study

A study just published by John Q. Easton and Briana Diaz for the University of Chicago Consortium on School Research, “Lasting Differences: Math Grades and Gender”, drills down into the gender gap in math. They study the gender gap in 9th grade math GPA in Chicago’s public schools. This narrow focus combined with a rich data set, courtesy of Chicago Public Schools (CPS), allows them to drill down into some possible causal factors.

There’s a good reason to focus on 9th grade. There’s a lot of research showing that 9th grade GPA is highly predictive of later outcomes. (For a good summary, see this episode of the “Make Me Care About” podcast, with guest Krystal Payne, Co-Executive Director of the UChicago Network for College Success).

Chicago has racially diverse public schools, with 47% Hispanic and 36% Black enrolment. Easton and Diaz find a similar gender gap in 9th-grade math across all racial categories:

Final grade points (ranging from 0-100 which maps to a final letter grade of A-F) for all CPS ninth-grade algebra and geometry students in school years 2016–17 and 2017–18 (29,229 students total). Source.

Although their main measure of interest is GPA, their dataset also includes test score results (specifically, 8th-grade MAP), as well as other behavioral indicators such as attendance and suspension. They show that by far the biggest gender gap is in GPA:

Why? Things get a bit technical here, but I’ll try to summarize as best I can. One reason for the gap is that boys are more likely than girls to be in a less advanced math class (algebra rather than geometry), and in those classes a higher weight is place on assignments (where boys are further behind girls) as opposed to assessments, where the gap is narrower:

Chicago girls are also more likely to be in an Honors levels class in both Algebra and Geometry:

Controlling for the level of class students are in, as well as behavior variables, reduces the gender gap in math, but does not eliminate it. As the researchers write:

Even comparing very similar students, young women’s math grades were still higher by a small but statistically significant amount.

Grade weighting varies a lot

The Chicago study confirms the basic Stanford picture: in lower-income school districts and/or those with more Black students, there is a gender gap in math favoring girls at least up to 9th grade. (It’s important to note that boys seem to catch up and even overtake in math in high school, at least in terms of test scores, even as girls remain ahead in most other subjects).

Tucked away in the Appendix of the paper, was one of those charts that gets permanently etched onto your mind’s eye, and which provides a salutary warning to treat all studies using GPA (including mine) with a degree of caution. Thanks to the richness of their data, Easton and Diaz are able to look at how much weight different teachers put on different things when grading their students. And boy do they vary:

For some teachers, assessments account for 100% of the grade, for others they account for 0%. Ditto for assignments.

There’s no reason to think this would impact the analysis of the gender gap, but it might impact other things, such as race. What if, on average, teachers in a majority-Black schools are putting a different weight on different contributors to the grade than teachers in a majority-white school? The researchers write that “students may benefit from better understanding how their grades are affected by their teachers’ grading categories and weights”. That’s true, but researchers would likely benefit too. I’m still thinking through what this variation in weights might mean for a whole bunch of things: certainly it’s a striking finding.

Easton and Diaz conclude their paper with the following suggestion:

School communities districtwide may benefit from rich discussions about how to improve young men’s experiences and outcomes in school. . . A system-wide examination about school-by-school variability may be a helpful start, followed by considering what changes schools and educators can make. At the individual school level, instructional leadership teams could evaluate available data. . .to understand the experiences of young men in greater detail, bring in teachers and students to discuss, and consider potential changes.

As you might expect, my response is: Amen. Let’s see if Chicago’s schools respond positively to this research.

Richard Reeves

President of the American Institute for Boys and Men