> list, and the following accumulation of objective data regarding
> students going on to graduate school, based on 4 major tests.
> 1. The GRE: Aggregate GRE Test scores fell slightly from 1965 through
I don't know how much of what I say below applies to GRE data collected in the US, but on an international level the GRE tests, especially the Subject tests, were pretty easy to twist.
At least in India, where a lot of early learning consisted (in the late 70s) of rote memorization, it was easy for some of us engineering students with notoriously bad verbal skills to memorize long vocabulary lists for the GRE Verbal test, without having a clue on correct usage of the words. This didn't help us on the essay questions, of course. Naturally, students who got their schooling in English had a substantial head start.
Secondly, some of our campuses (especially the elite ones that sent most graduates to the US) had developed huge compilations of earlier GRE tests, collected by former students, that we practised on and contributed to. This was especially useful in the Subject tests (engineering/computer science/physics etc.), since ETS was notorious for re-using questions from earlier tests. I don't know whether Kaplan/Princeton and suchlike provide the similar services here. Our campus compendia were communally owned, hence free!
Thirdly, in the mad desperation to get to the US, cheating was quite common, especially by the students from the non-elite campuses, who did not have the right label though they might have had the brains. After fax machines became more common and accessible in the early 90s, question papers smuggled out after the test in Japan were faxed to Indian cities hours before the test there. Indeed, some were even retransmitted to the US. This was found out pretty quickly by the testing folks, and the GRE tests were cancelled for a full year or more in India, while ETS got its act together.
--prashanth