Boys Can Read

Whitmer is right about the gender gap in reading in Michigan. But it’s a national story, as Claire Cain Miller showed recently in the New York Times, drawing on NAEP numbers crunched by Sean Reardon over at Stanford. (If you’ve not been following Miller’s series on boys and men, check it out. Superb stuff.)

In a striking chart, she showed that by the end of high school, boys are on average roughly a year behind girls a gap that has remained stubborn over time:

One question that often gets asked at this point is: But what about math the other way? It’s true that on the NAEP math scores, boys are ahead of girls. But a couple of points of worth noting here.

First, the math gap favoring boys is much smaller that the one favoring girls in reading about half as wide, in fact, so closer to about half a year of learning:

Second, and perhaps counter-intuitively, math scores are much less of predictive of the next stage of the education journey, as Brookings work by Sarah Reber and Ember Smith in 2023 shows. They examine which, if any, measures of academic preparation in high school explain the gender gap in college enrollment.

After showing the large gender gap in enrollment at both two-year and four-year colleges, they control for various measures to estimate their contribution to the gap:

Strikingly, and in contrast to their results for class and race, the math test score results don’t move the needle at all. It is other outcomes that matter, especially grades. As they write:

GPA explains essentially all, and math test score explains none, of the gaps.

Subsequent work by Reber, this time with Simran Kalkat, and Gabriela Goodman, shows similar results. In fact their measure of academic preparation was able, statistically speaking, to explain pretty much all of the increase in the college enrollment gap:

The general thrust of the findings in this area of scholarship is that GPA, course-taking and literacy matter for subsequent academic progress: math does not.

One question I have about the Reber et al findings is how far the causal relationship between high school academic engagement and college enrollment might go the other way. In other words, one reason why boys might not take harder classes or worry too much about their GPA is that they are already disinclined to go to college. It might be less than boys don’t do to college because they don’t do well in high school, than that they don’t do well in high school because they don’t want to go to college.

(I’m having my new best friend Claude have a look at that: let’s see what he’s come up with my the time I’ve finished this post….)

As Miller’s oped acknowledges, the reading gap is large all the way through the K-12 years. Another Brookings article, just published, suggests that the gender gap in reading is “fully baked” even before school. I worry this framing could lead to some fatalism. “Fully baked” sounds dangerously close to gender determinism.

And in fact there are some encouraging signs that some programs could help to close the gap. A couple of the initiatives mentioned by Miller at the end of her Times piece offer potential models here, specifically the NC Ed Corps tutoring program and Response to Intervention (RTI), an instructionally-aligned group tutoring scheme in a large SE school district.

The scholars evaluating Ed Corps write:

Exploratory analyses reveal that instructional alignment appears to have a greater impact on boys.

And the team examining RTI write:

While we find positive (though not significant) impacts for a wide variety of student and school subgroups, there is some evidence that boys may benefit more than girls.

(A nice additional finding from the RTI study is that the outcomes were worse when the tutor had Master’s degree..)

We should be careful here. The gender different in outcomes are not themselves statistically significant, and these studies don’t have much power simply because of sample sizes. An AIBM assessment of tutoring programs overall did not find much evidence of a gender difference; thought frustratingly, only a minority of studies disaggregated their results by gender.

Quick rant at education scholars: COULD YOU PLEASE FOR THE LOVE OF GOD DISAGGREGATE YOUR DATA BY GENDER

Rant over; thank you.

With Governors like Whitmer and others taking the gender gap seriously, and renewed attention on reading and literacy in part because of the “Mississippi miracle” fever, there’s an opportunity here. If we truly want to boost reading scores, we better start with the boys.

P.S. OK, update from Claude: couldn’t do this analysis because we need to physically go get the data. (Is anybody else noticing how quickly things have gone from “the AI can do the grunt work for me” to, “I have to do the grunt work for the AI”?)

But even the descriptive results on college-going expectations in 9th grade are telling:

I’m willing to bet a fair amount that even controlling for starting academic position, the kids who have lower college-going expectations don’t do as well in high school. Somebody out there want to do this research?

Richard Reeves: Writer, wonk, Dad. Founding President, American Institute for Boys and Men via Brookings. Tweeting from @richardvreeves. Books include Of Boys and Men (2022), Dream Hoarders (2017), and John Stuart Mill (2007). www.richardvreeves.com.