Originally posted by Eladar
http://www.randalolson.com/2014/06/25/average-iq-of-students-by-college-major-and-gender-ratio/
I found this interesting. Notice that the IQ doesn't drop below 100, so we are talking above average intelligence here.
Perhaps it isn't societies' gender beliefs that are creating the gender difference in areas like Physics and Engineering. Based on this link, the gender differences may simply be the result of IQ.
The IQ scores were estimated from SAT scores. There was
no correlation between the tests on verbal reasoning and subject. However there was a correlation between quantitative reasoning and subject. Unsurprisingly, people with good quantitative skills tend to do subjects like maths, physics, and engineering. People with poorer quantitative skills tended to go for non-quantitative subjects. None of this should come as a surprise. However this effect is more pronounced with men than it is with women.
Now we get to the problematic part. They estimated the IQ from the SAT score. This is a dubious piece of methodology as it introduces a bias towards physicists (not that I'll complain too much about that). With verbal reasoning it is generally difficult to get either a very bad score or a very good one. In quantitative tests it is a lot easier to get full marks, and it's easier to score really badly. This means that students who are not interested in maths will tend to score relatively badly in the quantitative tests - and this is not necessarily related to IQ.
What you are not taking into account is that men with good quantitative skills tend to do a quantitative subject. The graphs do show that women, on the other hand, are more likely to do a qualitative subject,
even if they have good quantitative skills. The method for estimating IQ is biased towards the quantitative subjects so that a gender difference in IQ appears as an
artifact of the methodology.