I disagree that this test could ever possibly serve as a genius-detector... you could be the greatest genius ever, and know how to derive every formula that you've ever seen, but you'd never have time to do that on this evil test. This is a test to find out who has been drilled on how to answer multiple-choice from the day of their birth. If your university doesn't emphasize this kind of thing, if you're more of a meticulous person than a very fast worker, or if you're prone to freaking out on tests, it doesn't matter how smart you are!
I spoke with a lot of graduate students at my university after my experience with the test... They all agreed that in the end it didn't matter as much as they thought when they were in our position. They've all ended up with positions they like, and are doing work they mostly enjoy.
I also spoke with my quantum professor, who was on the writing committee for a couple of years not long ago. Apparently ETS gets about 5 physicists together (at their EXTREMELY LUXURIOUS FACILITIES
somewhere back east
... I'm telling you, those ETS people are just scamming us out of our money...). Each professor writes 20 questions. Then they peer-review each other's questions... and they all get them wrong because of various ways you can interpret these questions. So they sit down and debate each question until they have questions they agree upon. My professor is now of the opinion that GRE scores say absolutely nothing about the promise of a young student.
There is quite a bit of research in standardized test statistics published online at the ETS website. Almost nothing on subject tests, but lots on the general test. One paper I found particularly interesting was the one about the differences between male and female scores on the general math test: apparently, they can mess with the content of the exams so that females score significantly lower than males, as opposed to both groups getting comparable scores (this has to do with changing the number of spatial-thinking problems and the number of problems with more than one path to the solution). Now, I would say that for ANY test that can be skewed so easily (not just in the male/female sense, but in general), there is not a lot of meaning in whatever scores we all get--because we could simply go back and skew the test the other way and all the statistics on who gets what score will change.
Bottom line--ETS sucks, and the GRE is basically useless crap that they use to force us to part with our money.
opticschmoptics wrote:erc. You're totally right...
and maybe to serve as a crude genius-detector on the other. Of course, the test is crackable if you have enough money (for coaches) and/or time (off from your real life/job/worthwhile research), so even this virtual meritocracy is easily corrupted.
To reiterate: 1) evil, 2) required to get in. Think of it as hazing that hasn't been banned yet.