I'm close to a number of folks going through the 2012 application season, and we've been waxing philosophical about the use of GRE scores. There are plenty of posts on this (and other) forums about the GRE, but there is a dearth of information out there about how each of the scores is used.
Can we have a more in-depth conversation about:
1. The distinction and relationship between each of the three scores?
2. How are each of the scores weighed in different disciplines and departments?
Anyone who has served on an admissions committee, or been otherwise privy to the evaluation of candidates, is more than welcome to chime in here.
Having been paying attention over a few application seasons, I will summarize the conventional wisdom:
(note that most of these observations are with an eye to how things are done in programs with a lower admissions rate)
Quantitative Score-
There is a widely-held generalized belief that, for all programs outside the humanities, the Quantitative score is all that matters. Many programs in engineering, accounting, and other "applied" technical fields have been known to state as much publicly. Beyond basic command of English, Q score seems to be "the" yardstick above any other. Interestingly, with the rise of quantitatively-oriented inquiry in the natural and social sciences, hearsay has also expanded to encompass sociology, anthropology, political science, geography, & public policy programs in this generalization. Tellingly, the NRC rankings use only the Q score in their reporting of average GRE scores for all programs outside of the humanities (whereas only V scores are reported for humanities programs).
Yet, by ETS's own admission, the concepts tested in the Quantitative section do not test abilities beyond a middle-school or early high-school mathematics course. Many students therefore rightfully complain that its focus is irrelevant, either because their discipline does not require mastery of the Pythagorean Theorem, or - just as significantly - because they have moved on to far more abstract concepts in mathematics. Which broad disciplinary group scores the highest on the Quantitative section of the GRE? Is it life scientists or natural scientists? Nope, its the Engineers. In fact, engineers as a group have a higher mean score on the Quantitative section than any other group has on any section. That they dominate this portion of the GRE suggests some unevenness in the quantitative concepts being tested.
The recent overhaul of the GRE format has supposedly placed more emphasis on "analytical reasoning" in all three areas of the test. Perhaps ETS is trying to reinvigorate the validity of the rest of the exam. Nonetheless, I have heard applicants - from political science to demography to ecology - being told that the Quantitative portion of the exam is by far weighted the heaviest.
Verbal Score-
Traditionally, score ranges have skewed much lower on this portion of the exam. Part of the recent ETS overhaul of the test was meant to narrow the distribution at the top of the score scale. Still, a score in the 96th percentile is now "only" a 165/170 - whereas the same percentile in Quant yields a 168. Given it's greater apparent difficulty, it would appear much easier to distinguish among applicants using the Verbal score. How are verbal scores interpreted? MENSA has asserted that Verbal GRE scores have a stronger correlation with IQ than any other section. Yet, again, outside of the humanities, is anyone paying attention? For those of you who have served on admissions committees, what weight have you seen Verbal scores hold?
Analytical Writing-
Many years ago, this section apparently resembled the "logic problems" portion of the Law School Admissions Exam (LSAT). Now it exists as 2 essays, written in 30 minutes each and subjectively graded by human beings on a scale of 1-6 in half point increments. According to ETS the Analytical Writing section score is apparently a better predictor of research ability than the personal statement. Yet, the perception among students is that this section is the least important of the 3. Is this the case? Does an admissions committee see any appreciable difference between an AW score of 5.0 (87th%) and a 6.0 (99%)? I've heard admissions professionals state openly that they "don't know what to make of the AW score" so, barring it being terribly low, they end up just ignoring it. Is this the case?
Combined Scores -
Often GRE scores are used by a university as a metric to levelize across disciplines - in the awarding of University Fellowships for example. The revised GRE does a better job at equalizing numerical scores in Q & V with their percentile counterparts. Previously, a student in the top 6% of test takers on the Verbal section would score a 660, while a score in the top 6% on Quant would yield an 800, now the numerical scores are roughly equivalent by percentile. Yet, with all of this emphasis on Quant, how are students compared at the university-wide level? And does the Analytical Writing score even play into the ranking?
I was a DGS in a small econ program. We mostly ignored the AW section. For the quantitative we wanted to see a very high score as a lot of our applicants had 800. If you weren't in the top percentiles we would eliminate you unless there was something else really good on your application. We used the verbal score in the opposite direction. A low score in a native speaker would knock you out of contention. For non-native speakers we would excuse lower scores but our university had a minimum limit that we needed a waiver for. Some non-native speakers did have high verbal scores (lots of training for the test plus language ability). If you had high scores in both you could compete in the cross-campus competition for fellowships.
