Should I Retake the GRE for Cornell or MIT Admissions?

[Ben Niehoff]

Tom earlier mentioned that ETS engineers tests to produce bell curves. This is mostly true, but not completely true. The GRE produces bell curves except on the Quantitative section, when compared among students going into math-intensive fields (physics, computer science, mathematics). Within these subjects, the curve instead has an exponential shape, with a delta function right on the highest score. That is, among quantitative science-geared students, the vast majority of them get perfect scores on the Quantitative section. If you earn less than a perfect score, you are below average among this particular subset.

Graphs are available within the PowerPrep software, broken down by major, which you can look at for proof of this. The worst misbalance occurs in computer science, with some 1100 data points on the 100% mark, and quite few below.

Oddly enough, the Verbal section does exhibit nice bell curves for every subset, including English majors, Journalism majors, Linguistics, etc.

I'm not sure whether this is the result of a flaw in the test, or the fact that math always has clear-cut, logical answers (whereas Verbal questions can appear to be debatable if you think about them wrongly).

Of course, the result is that in quantitative science departments, a good GRE Quant score doesn't really have a lot of weight (but a bad GRE score can have a negative weight).

 

[G01]

What I have heard in my research of the GRE and what it is used for in regards to physics grad school is the following:

The quantitative GRE is absolutely meaningless for picking physics grad students except to pick out applicants who really cannot do basic math. I have been told to get >700 on quantitative and that should be fine. How far above 700 your score is really doesn't matter, since this score is used mostly as a threshold to pick out candidates who may have basics math problems.

The verbal GRE score is even less meaningful. Get somewhere on the positive side of the bell curve and you should be fine.

More of you studying time should be spent studying for the GRE Physics test. The reason ETS even offers this test is because Physics grad programs know how useless the general test is in picking applicants. They might as well be rolling dice to pick students. The GRE Physics is scored out of 990 and I have been told that the average for American applicants at top 10 programs is somewhere in the vicinity of 850.

If any of the information I have provided is incorrect, someone please correct me.

Anyway, My scores are:

Q- 730 (I'm a slow worker and had to guess on the last few.)

V- 560

From what I was told (what I said at the beginning of these posts), these scores won't prevent me from getting into top 10 programs, but they also won't seal the deal. I think that most of my time is better spent working on upping my physics GRE score, rather than retaking this to get a 800Q. Is my analysis correct?

 

[leumas614]

I don't know if those scores are enough for a top ten program in engineering because there is no subject test. I think you're right for physics though. From what I've heard grad schools place much more emphasis on the subject test than the general test.

I think Ben is incorrect in saying that the vast majority of science-geared students get perfect scores on the Quantitative. ETS averages scores according to majors and the average score for physical science is a 690 and for engineering a 720. This would be impossible if most people got 800's.

 

[proton]

What I have heard in my research of the GRE and what it is used for in regards to physics grad school is
The verbal GRE score is even less meaningful. Get somewhere on the positive side of the bell curve and you should be fine.

where are the sources of your 'research'?

also, what's considered the 'positive' side of the bell curve for the verbal GRE?