The Economics behind the Math Gender Gap

Juan Sebastián Muñoz (the economist, not the golfer) has a new paper on this topic, focusing on Colombia:

The literature that has previously shown that boys outperform girls in math tests has failed to explain the underlying causes of the phenomenon. This math gender gap has been documented to vary across countries, and shown to grow as students advance through school. In this paper I suggest that these patterns may be explained by sample selection caused by gender differences in schooling’s opportunity costs, which lead lower-achieving males to drop out. I present and test the implications of a labor supply model that examines the opportunity cost of school attendance and, thereby, the observed math gender gap. Using an exogenous policy change, the launch of a conditional cash transfers program in Colombia,
I estimate that sample selection explains between 50 percent and 60 percent of the gap. Estimates of non-parametric bounds show that selection in the lower quantiles of the male distribution explains a significant portion of the gap.

Comments

Is this consistent with males having a slight higher stdev than females?

That was my thought, too--outliers seems little more likely to be male, the further out the bigger the difference. So plausibly more boys than girls just flat aren't cut out for school
and they drop out.

This was similar to my thinking as well:
"The literature that has previously shown that boys outperform girls in math tests has failed to explain the underlying causes of the phenomenon." I was under the assumption the literature suggested that boys both outperform and underperform girls in math tests. Where's Larry Summers?

Any parent with both boys and girls can tell you what the problem is. With few exceptions their interests take different paths and girls aren't interested in math and mechanical things. Girls make better nurses and handle administrative tasks better than boys. Boys are interested in mechanical things and how things work and like to take risks. They are genetically different and think different and want different things. This stays true no matter how much the adults force them to adopt the other genders natural instincts.

If you want more girls in math then push them into it in the lower grades. Give them mentors and additional math classes. It will work to some extent but mostly when they are 18 they will still want to get married and have kids a lot more than they will want to work 80 hour weeks in engineering and other math centric jobs.

>They are genetically different and think different and want different things.

Sure, this is 100% obvious to anyone who has been paying attention the last 15 centuries or so.... but the topic under discussion is who is going to get The Prize for being able to point to a piece of paper that says boys and girls are equally good at math, and any variance between the two is due to the fact than men are evil.

get that paper on the desk of Justice Roberts or Attorney General Sessions ASAP.

I've never met a nurses job or an administration job that doesn't depend on maths.

Women typically prefer maths about people - statistics, epidemiology - whereas typically boys prefer the maths of things - calculus, physics. It's just the latter has had higher prestige (and been around longer) but times are changing.

There is math and then there is "math". I can assure you that the math that nurses and administration jobs use is not the same math scientists use. The fact that you don't know this kinda points out the problem.

The maths that scientist use varies considerably across the different fields. I would certainly put nurses on a par with biologists and chemists even if they don't get anywhere near engineers or pure mathematicians.

A nursing friend of mine is doing a (nursing) PhD which is 85% geometry, trig, calc and programming.

But the article was about differences in maths performance at the high school level. Changing the context of the issue by redefining what maths means is simply a way to exclude people.

I would put nurses on a pedestal but the math that they take at any level in their education does not even come close to approaching the difficulty of the math a scientist or math major takes.

The article is about the difference between males and females as it relates to math (period). That it used the high school level as a talking point is irrelevant

I agree that the maths a nurse takes and the maths a math major takes is different but not compared to an archaeologist or botanist. Nurses do epidemiology (in my country) - the latter scientists can probably get away with doing no college level maths at all.

The article is about the gap in maths performance between males and females as high school students. We can't say much about it out of high school because there are no population-wide tests of performance once people leave school. And the tests that are in high school are simplistic measures of ability, on narrow fields in maths.

I'm a nurse. We take basic algebra and stats. We use these in chem and biology classes, and a little bit in calculating IV infusion rates and drug doses. Not much else unless you go on to do research. For most of us math is not much used after graduation.

This doesn't seem to do a good job at explaining the variability and/or the data suggesting boys BOTH over- AND under-perform. Also, my recollection is the data held at early ages for IQ testing rendering your assertion "If you want more girls in math then push them into it in the lower grades. Give them mentors and additional math classes." invalid.

I feel compelled to point out that this is just a troll post

That is why you will never be president of Harvard.

So, any pondering necessary on how this too could be an example of statistical discrimination?

Here are the differences which haven't changed much over the years:

SAT math scores; overall a 35 point gap. at the higher end:

700-800: boys 62%, girls 38%

600-700: boys 55%, girls 45%

oh...

750 - 800: boys 66%, girls 34%

Does SAT-maths tests maths ability?

It's not a perfect signal at the individual level, but it is highly correlated. For example, I would doubt that you could find a single tenured professor of mathematics at a top 100 US university who did not score above, say, 780 on their Math SATs. Heck, I got an 800 on PSAT, SAT, and GRE Math and I never for a minute deluded myself into thinking I could make a decent living as a Math Ph.D. in academia.

I don't think that's true; speaking from the experience of being a math graduate student at a top 20 university in the US. Obviously people don't share their high school examination scores (I do not even remember mine, although I remember doing not particularly well on the math, and very well on the reading).

The math SAT boils down to a relatively small list of rules to memorize and practice and bears no resemblance of what math actually is. It is also taken by a 17 year old. Indeed, even at the graduate level, I think programs care more how you do on the verbal portion of the graduate records examination (similar to the SAT). I have spoken to a physics professor who believes this correlates better with success.

The point is why has there been this consistent gap between boys and girls?

There is also a gap at the high end of IQ:

IQ over 130 1:2 girls to boys
IQ over 140 1:2.5
IQ over 150 1:4
IQ over 160 1:6

This is basic information that is easy to look up.

Three is a gap at the high end among people who have been tested. Is everyone tested in the USA?

That is result of thousands of IQ tests over decades taken around the world.

"Three is a gap at the high end among people who have been tested. Is everyone tested in the USA?"

At lest now I know I can discard any post by mpledger, s/he's not really seriously discussing the issue.

I come from a different country so I don't know the context. In my country the only use for IQ tests is for joining Mensa. The point is that if IQ tests are mandatory or self-selecting then it makes a difference on what conclusions can be drawn from the stats.

The meta-observation is that things we think we know might be based on less clean data than we thought.

What would you have to do to really track potential? Choose X thousand births and then methodically track down everyone later in life, no matter their academic path? Even then, you only discover their path, and not their potential.

I think that well over 80% of our data-based knowledge of human behavior is based on much less clean data than we thought. Tyler and other economists and non-economists with a high propensity to rely on data-based knowledge should verify the data, but they will not do it, either because it's quite expensive, or because they like the conclusion.

Give it some time, and we'll be collecting clean data like that on ourselves and giving it to people like google and facebook.

I can easily imagine a near future where your every meal, your sleep, how you spend your time, and measures of skill/aptitude are all automagically logged.

Why wouldn't the phenomenon of low-achieving males dropping out create a gender gap in favor of males in every subject?

One thesis: males underperform females in other subjects, so the low end dropouts merely reduce the measured differences.

Yeah, this seems like an obvious objection--to the extent the least able boys drop out more readily than the least able girls, that should give us verbal and math gaps in favor of boys.

I hate to be the person to point this out, but even larger than the claimed "success in math" gender gap (in which the mix between the 49-51 composition of the genders in our society, when counting newborn babies, winds up at approximately a 20 to 1, or 1 to 20 ratio, when counting who is creative at the top level in math) is the gap that becomes apparent when one considers the differential between the mathematical creativity, at top levels, of the people who grow up to be more attractive and have better hygiene than average (by definition, the "more" half in a 50-50 ) and the people who grow up to be less attractive and have worse hygiene than average (the "less" half in a 50-50). A lifetime of observation has led me to believe that the claim that "19 out of 20 people who are significantly creative mathematicians are from the approximately half of humanity who are male" is dwarfed by my belief that "39 out of 40 people who are significantly creative mathematicians" are from the approximately 50 percent of of humanity who are , in their particular nation, people who are less attractive and have worse hygiene than average.

I am not trying to be negative or unkind, but two things are true: (a) women can get away with being less attractive than average and still have a good chance at having a rewarding pair bonding experience than a similarly situated man can, and (but? - English is hard - I know what I am trying to say, and I leave it to you to figure out if I mean "and" or "but") (b) there is a significant possibility that women are less willing to sacrifice hygiene and attractiveness, as Newton and thousands of others did, for the single-minded pursuit of a little bit more mathematical fun and renown than they otherwise would have had. God bless them.

that comment purports to explain one component of the "drop out" incentive.

"still have as good a chance as a similarly situated man can, and (but? ...)

The logical value of 'but' is 'and'. You're doing fine.

It's not really spoken of much, but in addition to the different standard deviations, 14 for women, 16 for men, women have a slightly lower mean IQ than men, about one- to two-fifths of a standard deviation (three to six points), so something like 98 versus 102 for men. This small difference has no effect on anything, except at the extremes, where there are very few super high performing women mathematicians.

This area is quite taboo, but just in the past few years there has been increasing research. Traditionally, IQ tests were normed to give equal results just because it was assumed by test makers that IQ would not differ by sex.

Interestingly, women do perform higher than men on some cognitive subtests, but on most men perform higher.

What about math tests?

IQ tests cognitive ability, which is kind of circularly defined as the things that, when you test them, correlate statistically with the other things. These include language skills like reading and vocabulary, math, various kinds of puzzles, spatial perception, pattern recognition, physical reaction to visual stimuli, short term memory tests and manipulation of items in short term memory. Something not cognitive is musical rhythm.

So any test of any of this stuff is in effect a math test. The further from what you want to specifically test it is, the more noise in the results. But as the number of samples approaches infinity, the distortion by noise approaches zero. This is why they can use something as noisy as highest education achieved as a proxy for IQ in large scale MRI or genetic studies.

The one thing that advocates of the theory that innate sex-correlated differences explain the math gap in countries like the US have to explain is why that gap looks so different in different countries.

I'm a mathematician, and judging by who I see at research conferences, Italian and Portuguese mathematicians are pretty evenly split between men and women. Other countries like Germany, the US, and France have very different splits. The other mathematicians I've mentioned this to generally agree that it seems to be the case. Since Europeans from different countries are basically genetically identical to each other, you'll have to bring in at least one social ingredient to explain the difference in math gaps (perhaps in addition to genetic ingredients, which are no doubt also important). Any explanation that doesn't do that can't possibly explain the math gaps in the countries where it exists.

The author of the paper acknowledges this upfront in the second sentence of the blurb quoted above, but pretty much every other comment here has ignored it.

"The one thing that advocates of the theory that innate sex-correlated differences explain the math gap in countries like the US have to explain is why that gap looks so different in different countries. "

The differences in this math gap in different countries are evidence that "innate sex-correlated differences" are not the whole explanation.

But there is plenty of evidence that there are important "innate sex-correlated differences", and that such "innate sex-correlated differences" affect the choices that people make.

'A gender equality paradox': Countries with more gender equality have fewer female STEM grads

QUOTE:
Dubbed the “gender equality paradox”, the research found that countries such as Albania and Algeria have a greater percentage of women amongst their STEM graduates than countries lauded for their high levels of gender equality, such as Finland, Norway and Sweden.

The researchers, from Leeds Beckett University and the University of Missouri, believe this might be because countries with less gender equality often have little welfare support, making the choice of a relatively high-paid STEM career more attractive.
END QUOTE

So should countries with a big gender gap in math strive to be more like Albania and Algeria?

Oops, forgot the link:
http://www.thejournal.ie/gender-equality-countries-stem-girls-3848156-Feb2018/

"I'm a mathematician, and judging by who I see at research conferences, Italian and Portuguese mathematicians are pretty evenly split between men and women."

I looked up the faculty listed in the mathematics department at the University of Lisbon and the male to female ratio ratio is 2:1 and all are Portugal. At Princeton, which attracts the best mathematicians from around the world, the male to female ratio is 9:1.

Looking at the faculty of Portugal and Italy's top 5 mathematics departments would likely show around a 60/40 to 65/35 split.

(all are _from_ Portugal)

The smart money says that if the analysis hadn't confirmed his thesis, the moderators would have ignored the paper.

Maybe there's not much difference on average in math ability between males and females. However, it does seem that females are scarce at the very top.

2018 International Mathematical Olympiad Individual results:

http://imo-official.org/year_individual_r.aspx?year=2018

About 600 participants, of whom about 300 received medals. 17 females received medals. About 50 gold medals, 2 gold medals for females.

2018 Raytheon MATHCOUNTS National Competition

https://www.mathcounts.org/sites/default/files/u1706/2018%20National%20Final%20Standings.pdf

Of those 56 students listed, only one has an obviously female name. There are probably a few more females whose names are not obviously female.

Those are both opt-in events. There can be reasons other than maths ability that make people chose to compete e.g. a parent may not let her 14-17 year old daughter travel overseas to compete but let her same age son.

"There can be reasons other than maths ability that make people chose to compete e.g. a parent may not let her 14-17 year old daughter travel overseas to compete but let her same age son."

And maybe most of the girls get struck by lightning on their way to the math contests. That's about as much of a factor as what you bring up.

Students don't participate in the IMO or National MathCounts because they wake up one morning and decide they feel like going. There is fierce competition to get onto the teams (national for IMO, state for MathCounts). Only a few girls have ever made the US IMO team. The National MathCounts competition is always in the US.

The 2104 IMO was in Cape Town. The 2017 IMO was in Rio de Janiero. Safety issues do matter.

A (now) collegue and her (then) two teenage friends were kidnapped on the way to the airport by their taxi driver when they went to South America as part of a school cultural party. They managed to get themselves out of that by being clever and aggressive.

bill maher vaporiza su vagina

Comments for this post are closed