mGRE is only recommended when applying for a Ph.D in statistics. But obviously sometimes it's better not to send a score than to report a bad score. Stanford, which is the only place that actually requires mGRE, post that their average is the 82% quantile. But this also includes graduate students in probability, with much higher level in mathematics.

How would you cut the threshold between send/not send/indifferent?

  • Does the website provide any other statistical data, like maximum, minimum, or standard deviation?
    – The Hiary
    Commented Nov 24, 2013 at 6:22
  • Nope, just that it's 82%. And stanford program is way above all the other statistics programs in the world. statistics.stanford.edu/academics/…
    – Stats Grad
    Commented Nov 24, 2013 at 6:55
  • If they require it, they do not leave you a choice though so you should send it. Commented Nov 24, 2013 at 15:06
  • 3
    If you do not have the required math background, you should not attempt the program that would be way over your head, and that you will drop out of having failed the qualifying exams one or two years later. *Way above * is an overstatement, although of course most meaningful rankings will put Stanford on the 1st place in the US. However, if you want to do social science statistics, you would be better off at UWashington or UCLA, and if you want to do something Bayesian, at Duke or Harvard. It may be difficult to get through in these programs, too, if you math is not so strong, either, though.
    – StasK
    Commented Nov 24, 2013 at 20:01

1 Answer 1


It really depends. Obviously if the score is required you need to send it. If you are applying to top programs, a 68% at anything is not going to help your case so I would not send it if it wasn't required. At lower ranked schools, a 68% might be okay, but again it is probably not going to impress anyone. I would probably draw the line at the 90th percentile for providing something that is not required.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .