GRE and gender

[TheMercySeat]

I've been doing some leisurely reading on the gender gap.

Does this disturb anyone else?

http://psycnet.apa.org/psycinfo/2013-09150-001/

https://research.collegeboard.org/sites/default/files/publications/2012/7/researchreport-2001-6-differential-validity-prediction-college-admission-testing-review.pdf

[GeoDUDE!]

What's disturbing about that ?

 

(In other words, why don't you put a few line summary under each article and actually articulate a point other than a feeling if you would like to have a good discussion about something.)

 

You might think what you find disturbing is obvious, but then, that thought would be contradictory to your question "Does this disturb anyone else?"

 

 

Men tend to do better on standardized tests, have higher IQs ect. Its been that way; does it make men smarter? No. 

[TheMercySeat]

I didn't think to put a summary under the assumption that prospective graduate students would probably want to read primary source content and draw their own conclusions.

Anyway, since I think the point is missed... I would like to draw your attention to differential item functioning/differential validity and over/under prediction. These are psychometric flaws in the test with real life implications that systematically exclude qualified women from higher education.

Edited September 15, 2014 by TheMercySeat

[bsharpe269]

A quote from the abstract of the first link: "Transferred onto a 4-point grading scale, women earn college grades that are 0.24 points higher than those of men with the same admission test result."

 

If women and men are getting similar test results and women are getting higher grades then it seems to be like this is more related to choice of field than it is to an issue with the GRE. Men tend to study fields with reputations for lower GPAs (STEM) while women graviate towards education, humanities, social sciences which have a reputation for higher GPAs. It makes sense to me that men would have lower GPAs while getting similar test results based on this.

 

As a woman in STEM, I am not trying to be sexist with the above comment but this has been an observation of mine. I am also not suggesting that humanities and social sciences are any less important or worthwhile to study. I am only making a very generalized statement based on my personal observations.

[TheMercySeat]

Good point, bsharpe. I'd like to think that some researcher(s) somewhere controlled for major in some of the studies feeding into the meta analysis (second link), but I honestly haven't canvassed all of the articles.

Still... It doesn't quite resolve the gender gap on the SAT/GRE-- why is exists, how it came about, and why it is particularly bad in the US, relative to other industrialized nations. In fact, some literature suggests reverse-directionality in the US (females score lower in quant, which leads to them avoiding traditional 'hard science' majors; in fact, there's quite a bit out there that's specific to women in physics).

This topic has been a tangential interest of mine.

Edited September 15, 2014 by TheMercySeat

[TakeruK]

Here is an example case that does control for various fields of study. The Physics subject GRE is often a required exam for many astronomy grad programs. Here is a blog post from an astronomy professor at Caltech who collected data from their accepted applicants: http://mahalonottrash.blogspot.com/2014/06/the-gre-test-that-fails.html

 

The prof shows the distribution of the PGRE scores of accepted applicants, sorted by gender (the black and red histograms). It shows the median score for male students is 740 while for females, it's 660. If you are not familiar with subject tests, the scores range from 200 to 990. In addition, the blog post also writes about some other studies on the distribution of the General GRE scores based on race and gender; and also an old study (back to 1996) showing basically no correlation between GPA and Physics GRE score. 

 

The prof provides all this as evidence to back up his claim that the GRE is a better predictor of gender / race than it is of undergraduate academic performance or ability. 

 

I think there are at least two reasons to be disturbed:

 

1. The first might be that if schools use very strict GRE / PGRE cutoffs then some students from disadvantaged groups (in terms of GRE scoring) might miss out on opportunities that they are qualified for and would excel at. I feel that many schools in my field do not weigh the GRE scores very highly so this may not be as big of a problem, however it's still something to worry about.

 

2. The second is something the prof also brings up in his blog post: "stereotype threat". This is the concept that if a student knows they are part of a group that is expected to do poorly, then it can influence that student's performance negatively. There is more information in the linked blog post as well as other studies of stereotype threat. 

[TheMercySeat]

^ bingo.

See also: http://www.ets.org/s/gre/pdf/snapshot.pdf

For every nearly discipline, men outperform women on the GRE. Granted, these data are just descriptive, but still...

Why does America use metrics that disproportionately benefit males in college admissions?

Edited September 15, 2014 by TheMercySeat

[GeoDUDE!]

Men also outperform women on IQ tests. Its all standardized tests I'm fairly sure (perhaps you have an example of where women do better.) There are probably biological reasons, or whatever, i'm not sure.

[TheMercySeat]

Men also outperform women on IQ tests. Its all standardized tests I'm fairly sure (perhaps you have an example of where women do better.) There are probably biological reasons, or whatever, i'm not sure.

Men tend to cluster on both extremes-- just as there are more male geniuses, there are also more men with intellectual disabilities.

Also, women started outperforming males on Raven's Progressive Matrices (RPM) maybe about three years ago. It is generally agreed by psychometricians that RPM is a much more valid indicator of IQ than the SAT/GRE. First and foremost, RPM was actually designed to be an IQ test, whereas the SAT/GRE are aptitude tests that are largely contingent on rote memorization.

Edited September 15, 2014 by TheMercySeat