Grad programs in Statistics (not pure math)--min/good score?

[Ergodic]

I'm thinking about applying for graduate programs in Statistics, rather than traditional (pure) mathematics. I've looked at admissions standards/requirements at top programs, but they don't give any info about typical GRE subject scores of incoming students. I was wondering if any of you were or are in the same boat and could evaluate at a glance my chances for admission into either a prestigious Statistics program (e.g. Stanford, Berkeley, Michigan) or moderately strong program (e.g. Purdue, Iowa.)

Here are my "stats," so to speak:

1) I expect to score at least in the 80% percentile on the GRE subject test (though that could change come next month)

2) My general GRE scores are Q800 and V740 (99th percentile on the Verbal.)

3) I already have an MA in Mathematics, albeit from a low-tier state university and with only a 3.0 GPA (see #5)

4) I have one research paper in elementary number theory (not yet sent for publication, though)

5) I worked for a year as an Actuary (two exams passed), and I'm hoping this plus recommendations will demonstrate enough maturity to make up for my low GPA.


Any thoughts or comments would be much appreciated

[Nameless]

Hello,
You should get at least 90% for GRE math if you are heading to UCB, Stanford, ... For those schools, GPA is not a problem. You still get admitted if you have VERY STRONG letters of recommendation from well-known mathematicians. You should be prepared for GRE subject . I know some friends whose GPA is just 3.0 or little bit higher but still got admitted to UCB since they got VERY Strong Letters. His profs wrote like that : he was truly impressived by that student, the students seem did not care much about grade, only did what he wanted to do...even though his GPA is low but his knowledge is very much higher some PhD students he was advising......he not only understand mathematics deeply but also can work well in physics ....blah,blah.

I believe that guy got admitted because of such letters .
From my experience, only ask for LORs from profs. you know they will absolutely write for you good letters. Otherwise, it will kill you

Good luck

[Ergodic]

Thanks, Nameless!

Are you sure you're referring to graduate programs in Statistics, rather than pure Mathematics? (They are different.) For example, the Berkeley Statistics program (ranked #2 by U.S. News report) doesn't even require the GRE subject mathematics test, nor does Harvard (#4.)

Nevertheless, it's good to know that GRE and recommendations can make up for low GPA.

[diogenes]

Just to hop on the recommendation band-wagon, I think good LORs will go a long way to bolstering your application.

[ana3a]

hi ergo,

i found this at duke:
http://www.gradschool.duke.edu/about_us ... itstat.htm
(unfortunately nothing on SUBJECT test scores)
and from upenn:
http://www.math.upenn.edu/grad/graddata.html
(unfortunately that is a math program, not stats)

[diogenes]

i wish i could get a clear sense of the expectations at the different schools. i'd save time and money applying only to realistic places and know which schools to count on as safeties.

Depending on the alignment of the stars, relative temperature, and other unknown unknowns, you might be able to get a general competitive/middle/weak response to a hypothetical candidate that is shocking close in profile to you.

[ana3a]

ok thanks

[Ergodic]

ana:

I will reply in more detail later. (In fact, I'm thinking about making a single post that will compile as many links to GRE stats at various top schools.) Tonight, however, I plan to drown myself in beer, karaoke, and women.

For now I shall leave you with the following link: the GRE subject scores of the incoming class of 2005 for University of Minnesota (a top-20, or maybe top-10, school in Applied Mathematics, and also a top-20 in Pure):

http://www.math.umn.edu/grad/faq/index.shtml (Scroll down to Question 7.)