> males preferentially looking at/interacting with mechanical things, females faces/dolls

Then why are most politicians male?

The problem with this line of reasoning is that it can be used to support almost any status quo because there's no obvious connecting line between childhood preference for "mechanical things" and STEM career preference later on. In particular, modern software development is highly collaborative and involves no "mechanical" parts at all. It's often closer to prose than mechanical engineering - focused on structure and organization rather than numbers and analysis. Not to mention the plethora of counter-examples in the form of "technical" degrees and careers where women are closer to 50%: http://www.onwie.ca/resources-tools/statistics/canadian-engi...

Because politicians require hyper-competitiveness to reach the top of the political hierarchy. Unfortunately most women would be considered far too agreeable, or conversely gender stereotypes portray females as being out-of-character (bossy) pursuing such cut-throat roles.

I suspect the amount of hatred against the German chancellor Merkel I can hear in casual conversations here in Italy (by males and females equally) is mostly to be attributed to her being perceived as bossy just because she's a woman. An equally bossy man would be normal

Yes.

Another example, I have read research that ADHD might not necessarily be less common in women, just under diagnosed. A little boy running around the classroom and talking to everyone is potentially referred to a doctor for ADHD. A little girl running around the room talking to everyone is considered "social."

Bingo.

IQ variance might explain part of the difference in nobel prize winners, but is minuscule at average IQ in tech (~90th percentile).

Yet Marie Curie was the first person to receive two prizes. In fact, she was the only person who received two science Nobel prizes for 61 years.

That really doesn't invalidate the statistical argument. She was one person.

A better counter-argument would have been to focus on how historically women have been actively and passively discouraged from entering the sciences at practically every level of education. This more clearly has a causal role with Nobel rates and likely played a large role in the disparity.

"...The mental test applied was the Armed Services Vocational Aptitude Battery (ASVAB),..."

Well, ok, then.

"Males have only a marginal advantage in mean levels of g (less than 7% of a standard deviation) from the ASVAB and AFQT, but substantially greater variance. Among the top 2% AFQT scores, there were almost twice as many males as females. These differences could provide a partial basis for sex differences in intellectual eminence."

Now if we only knew the AFQT scores for, say, computer programmers. Oh, wait, we kinda do (http://www.military.com/join-armed-forces/asvab/asvab-and-ai...).

Air Force-wise, Electronic signals intelligence requires a G72, Cyber Systems Operations and Cyber Security requires G64. The ASVAB score is a percentile,

"Thus, an AFQT score of 90 indicates that the examinee scored as well as or better than 90% of the nationally-representative sample of 18 to 23 year old youth. An AFQT score of 50 indicates that the examinee scored as well as or better than 50% of the nationally-representative sample." (http://official-asvab.com/understand_coun.htm)

So being smarter than 65-70% allows you into those jobs.

(I am, by the way, loving Fig. 1. If I'm reading this right, "Auto and shop information" is a better predictor of g than "General science" knowledge.

"Participants were administered the Armed Services Vocational Aptitude Battery (ASVAB) which has 10 subtests: science, arithmetic, word knowledge, para- graph comprehension, numerical operations, coding speed, auto and shop information, mathematics knowl- edge, mechanical comprehension, and electronics infor- mation."

Yeah, that's the ASVAB I remember.

"Four of the subtests comprise the Armed Forces Qualification Test (AFQT) which includes only the more general, less vocationally-specific tests: arithme- tic, word knowledge, paragraph comprehension, and mathematics knowledge."

Well, at least intelligence, as defined, is less vocationally specific.

Aha! Table 1!

Someone please correct my understanding here if I've missed anything.

The male mean AFQT89 score was 38.7, the female mean was 38.2. The standard deviations were 30.1 and 27.7, respectively. (Wait, what?) The male-female standard deviation ratio was 1.09.

So those are the kinds of numbers we're talking about.

Genuine question: is is possible some selection bias is at play if only people interested in that given career will undergo this testing?

There are all kinds of biases involved, but the ASVAB has been normalized against a (the?) general population of high school students.

I think that the ASVAB was designed as a job-choice tool is more of a problem than that it was designed as a military job-choice tool.

The SAT was shown to be a flawed test in that it is racially, gender, and income-level biased. I'd be interested to look into the Armed Services Vocational Aptitude Battery with a keen eye to bias. I have to say that anything designed by/for the armed services makes me suspicious of bias for obvious reasons. I will see if I can find any studies on this-- would be good to know.

Found something about gender bias on this test:

"However, many widely used standardized test scores - ranging from the Scholastic Aptitude Test (SAT) to the Armed Services Vocational Aptitude Battery (ASVAB) and beyond - also reflect substantial differences based on gender. Minority females suffer a double jeopardy as they often score lower than both white females and males of their own racial or ethnic group. Nonetheless, the law regarding both gender issues and combined gender and race issues is largely undeveloped. Only one lawsuit, Sharif v New York State Education Department, " has been brought to challenge any use of a standardized test on the grounds of gender bias. Legal scholars have only recently begun to devote attention to this issue and have focused on Sharif. 2 Virtually no attention has been focused on the particular issues raised in connection with minority girls and women."

This is from a paper published in 2013 in Berkeley's Journal of Gender, Law, and Science. http://scholarship.law.berkeley.edu/cgi/viewcontent.cgi?arti...

Jumping from a few studies about infants and monkeys to "...you can assume that there are at least some biological differences in behavior between the sexes - and thus differing workplace representation is not necessarily 100% derived from social pressures." is far, far from enough evidence and extremely hand-wavy to make a scientific argument here.

There isn't really a leap. All I'm saying is that we have evidence for some gender differences that are rooted in biology, and thus we cannot assume that other differences have no biological component. Which is what asserting that 100% of a gender gap is caused by social pressures does.

It's very convenient to hand-wave away the problem as "biological differences". Until we have some direct link between biology and preference for software development I think the wiser course of action is to try to get more women interested in the field. I mean why not? That's not to mention that open discrimination against women in numerous fields is a thing that happened within the lifetimes of people that are still alive today. It seems foolish to think that's all been fixed already.

> I mean why not?

Why not have a broad-based program to appeal to all demographics? Black Americans are also underrepresented. Rural children have less access to certain kinds of resources like magnet STEM schools.