Also, from the original article:
That seems imminently reasonable, and fiddling with test length seems like a reasonable starting point to try to determine if test length has an effect on gender parity.
In well-designed tests, yes. But some tests really use time pressure to increase distribution. I’ve had a lot of kids–often very smart kids–increase scores a ton when they quit trying to show off by being the first to finish.
The ACT is the worst for this. The science section starts on page 40 of this PDF. Look at it: it’s clearly as much a test of reading speed/cognitive speed as it is of skill in science analysis. Ten kids with exactly the same science analysis skills could get radically different scores based on processing speed and test-taking strategy.
That’s a lot less likely outside the US, although sadly becoming fashionable.
What I’d like to know is from what elbow did the dons get the idea that women are slower at those particular subjects. This woman only ran out of time for an exam once in my whole life (third trimester, Electronics, ChemE). There was a female student in my K-12 class who always waited to hand things in until the teacher was hovering over her, but she’s famous for her insecurity, used to get 100% and managed to get a PhD in Pharma despite never being sure of anything.
Nava, aka “Ms. Lightning”
If you’re test results give you a bell-curve you are doing something seriously wrong. Bell-curves are for norn-referenced test. Criterion-referenced test should be skewed if the teacher is doing their job.
The theory is, and apparently this is measurable, is that certain steps during the development of male brains cut down the number of connection paths. This is thought to speed them up but throw away functionality. So you end up with on IQ tests, females score higher on average, with a smaller variance, while males have a lower average but a wider variance.
If you plot the two curves on top of each other, you do see more males on both the right and left sides of the curve. More retards and psycopaths and more geniuses, essentially. (from my perspective this seems like a rather questionable optimization)
So you would expect that at elite school like Oxford, there would be a large number of male students from that right side. They would have inherent advantages on certain tests over the typical female student at Oxford, for the same amount of study time.
These are population distributions and do not preclude the existence of “Ms. Lightning”, just that it means she probably has less peers if the theory is true.
My school was as elite as they get (considered one of the hardest majors in Spain, after Mines Engineering in Gijón and Telecos Engineering in La Salle) and the sex ratio has been 50:50 since the early 80s (the school was opened to female students in 1975). Telecos in La Salle also has a higher amount of female students than other engineering schools (I am not familiar with Mines). There were some things in which we could see differences between male and female students: they were more likely to have problems telling colors apart, we were more sensitive to electrical currents and to heat. None of those has an influence on our grades in Algebra, Multivariate Calculus, Electronics…
Begin by proving that there is a difference, before trying to correct it.
You do realize that your anecdotal experience is absolutely worthless, right? It tells nothing about population distribution.
From the article in the OP -
Which makes sense - a change that affects both sexes equally isn’t going to address factors that do not affect both sexes equally.
Regards,
Shodan
I’m mostly talking about standardized tests like the SAT that want, if not a bell curve, at least a lot of discrimination of scores.