Of Boys and Men
On the recommendation of Tyler Cowen, I’ve been reading Richard Reeves’ book Of Boys and Men. The most interesting parts of the book are the evidence that Reeves gives for boys doing much worse in school and university than girls, and the solutions that he suggests for improving things. I’m generally impressed by the evidence, but not so impressed by the solutions.
Why should we be concerned about boys and men?
Reeves presents a few of the most compelling reasons to be concerned about boys’ educational outcomes: boys are increasingly doing much worse than girls are at school, and interventions to help boys do better at school seem to basically be failing.
The evidence that boys are doing significantly worse than girls when it comes to education is pretty striking. Reeves points out that forty years ago in the US, there was a thirteen percentage point gap in the percentage of undergraduate degrees going to men compared to women (in favour of men). Now, there is a fifteen point gap in favour of women. In the UK the figures (from 2017) are similar, with 27% of young men going to university, compared to 37% of young women.
Reeves points to data from Finland that shows that boys make up the majority of children with the lowest reading scores: 20% of boys get very bad reading comprehension grades compared to only 7% of girls. Could this be a product of boys just generally having more variability in their performance, being more likely to be among the worst performers but also more likely to be among the best? Probably not, given that girls also make up the majority of the highest performing pupils, with 20% of Finnish girls getting the top scores compared to only 9% of boys.
This doesn’t seem to be cherry-picked data at all, and lines up with most of what I know about educational outcomes - see below for a figure, also taken from the book, that shows that a higher proportion of girls go to university than boys across the OECD. The evidence comes thick and fast, and is convincing: boys are really doing pretty badly at school, getting into university at lower rates, and performance isn’t improving much.
And something that’s especially interesting that Reeves points to is the fact that interventions that appear to have miraculous results seem to basically only work for girls while being useless for improving outcomes for boys. The Kalamazoo promise is a useful example here: in 2005, anonymous donors decided to pay 100% of college tuition fees for pupils in a number of public schools in Kalamazoo, Michigan. What happened? The number of women getting a college degree increased (by between 45% and 49%), whereas the number of men getting college degrees didn’t change.
I checked the original paper looking at the Kalamazoo promise to see if Reeves’ interpretation was correct, because it seemed completely wild to me that the percentage of men getting a college degree wouldn’t increase at all if they were offered 100% of their tuition covered. I think Reeves’ interpretation is correct, but there should be a caveat here. The promise also failed to increase the number of white children going to college, so the effect here seems to be basically concentrated among nonwhite children (probably due to the baseline of white children going to college was already pretty high).
So, the takeaway should probably be that the promise led to a huge increase in the number of non-white girls going to college, but had little effect on non-white boys, resulting in the numbers looking very good for girls overall and not good for boys, but really the story here is a bit more complicated. (I think this sort of bolsters the point Reeves makes about how there are often interactions between male outcomes and the outcomes of other groups that do worse in school).
Other studies Reeves points to have similar findings: the ‘Stay the Course’ (STC) mentoring and support program at Tarrant County College attempted to increase college completion rates. It worked, sort of. The rate of completion for female students increased a lot. For men, the numbers didn’t change.
Again, I looked at the original paper. This one is great because it’s actually an RCT, so we can take it pretty seriously. And yes, it does look like it had a significant effect for women and not for men. The authors of the study offer a few possible reasons for this. It could be that the study is underpowered, and would have found an effect on men if the sample size had been larger (there were half the number of men in the study as women). But it isn’t the case that the effect is positive but just doesn’t reach the threshold for significance, there is actually not much to suggest that there is an effect for men at all.
One other reason the authors offer up is that all of the mentors offered to pupils as part of the STC scheme were female, so if it’s the case that men are particularly responsive to male mentors but not to female mentors, we would expect to see this sort of result.
I’m kind of convinced by the evidence that educational interventions often work for girls/women but don’t work for boys/men. A few other examples are given: preschool programs, a summer reading intervention, college scholarship schemes, etc. They all seem to show similar results, with girls seeing a large improvement with the policy intervention, but boys seeing no result (or even negative results). All of the interventions seem to be in the United States, and I wonder if the research looks the same in other countries. But putting that aside, the evidence is generally compelling.
So, what do we do?
The main solution to help with these educational problems offered by Reeves is redshirting. ‘Redshirting’, if you’re unfamiliar with the term, is the practice of letting certain children begin school later than their peers do, and in this instance it refers to letting boys start school later than girls do. Reeves refers to Malcolm Gladwell’s book Outliers, which apparently shows a load of evidence that being older than the rest of your cohort benefits a child for the rest of their life. So, the solution is simple: redshirt all boys by default, and introduce a one year chronological age gap between boys and girls in school to erase the developmental age gap.
Reeves points to some evidence about how this is likely to work out: reduced hyperactivity, higher levels of life satisfaction in later life, higher test scores, and so on. He then lays out the objections to redshirting: delaying school could put pressure on parents to provide childcare for longer, boys who start schools later may be more likely to drop out of high school, boys will lose a year in the labour market, and phasing in the reform may be difficult.
And I think he deals with these objections fairly convincingly, possibly with the exception of the labour market objection, where his response is that this applies to all policies that increase the amount of education that children get. While that’s true, it doesn’t really seem like a slam dunk - many just take the view that this is also a problem with other policy suggestions.
But the main objection I have to the redshirting idea, which isn’t really addressed, is that it’s unfair on girls! If girls are competing for university places against boys a year older than them, they’re going to lose out through no fault of their own. And Reeves’ view on this seems to be based on the fact that there’s some evidence that redshirting doesn’t have an impact on younger children in the same class as older children who have been redshirted.
But this… doesn’t deal with the objection at all. It seems like at the moment there is some small proportion of children being allowed to start a year later, and the evidence suggests that this doesn’t harm the grades of the younger children.
But if the grades of half of children are improving where the grades of the other half are staying the same, the half seeing no improvement will be less likely to get into college (assuming the number of college places remains fairly stable). Even if their grades aren’t getting worse in terms of the actual letter grade (they may not literally go from getting As to getting Bs), they’re still going to be worse relative to the grades that boys are getting.
So, to me this seems like it’s basically Affirmative Action, putting the thumb on the scale to make sure boys are more likely to get into university. And fair enough, perhaps Reeves wants to basically do Affirmative Action. But in fact, Reeves is extremely opposed to the idea of Affirmative Action in its usual form, he just supports a policy that gets the same result (while taking the scenic route). Reeves writes:
Even though I am deeply worried about the way boys and men are falling behind in education, affirmative action cannot be the solution. (Or perhaps I should say, not yet.) To a large extent, the gaps at the college level reflect the ones in high school. Differences in early attainment at college can be explained by differences in high school GPA, for example. Reading and verbal skills strongly predict college-going rates, and these are areas where boys lag furthest behind girls.
Unless I’m misunderstanding Reeves (and I may be, given how he doesn’t really go into this objection at all, making me think that I might be missing something), he would be strongly against, for example, a policy that would artificially boost the grades of boys to reflect what they would have achieved had they begun school a year later. But the policy he has in mind does the same thing, just by actually giving boys a costly extra year at home, ensuring they’ll do better by being older than their female classmates. I’m a bit confused as to why he supports one but not the other.
Another solution that Reeves proposes is to increase the number of male teachers in schools, and cites this study that was published earlier this year, looking at the impact of a primary school quota in Finland that required that 40% of teachers were men.
The finding is: “A 1 standard deviation increase in the share of male (quota) teachers makes pupils .09 standard deviations more likely to shift towards a higher skilled degree”, meaning that the policy resulted in a small increase in the number of children getting university degrees as opposed to just completing high school. The policy didn’t seem to impact boys more than girls, so I don’t really get how this would translate to redressing the imbalance between the number of men and women completing higher education.
This just doesn’t seem that compelling to me - one decent study on the impact of quotas in Finland doesn’t sway me much, and nor does much of the other evidence presented that this would be such a good thing - there isn’t really a serious attempt to think about whether the trade-offs here are actually worth it.
I enjoyed the book overall, I think it’s worth reading, and I think it highlights some problems that (as Reeves is keen to point out) some people would be anxious about discussing given general views about male privilege. But the solutions, as far as I can tell, just don’t seem that convincing. Let me know what you think in the comments.