MS statistics Profile Evaluation + needed funding advice!

[dunleavy005]

Hi All,

I'm looking to apply to MS statistics programs in the fall, but am unsure of the difficulty of getting into good programs that may serve as PHD stepping stones as well as applied programs if I decide I want to work in industry after. I'm also interested in knowing about funding opportunities as the info on the various websites say its scarce, if at all. I've done my homework, but would like some opinions of people who have applied or know the process of applying for statistics grad programs. Below is my profile:

Type of Undergrad Institution/Major: UC Berkeley, B.S. Business Administration, B.A. Statistics

Undergrad GPA: 3.925 (overall)

GRE: 800Q, 510V, 5.0 AWA

Math Classes: Calc 1,2,3, Linear Algebra 1 & Differential Equations, Numerical Analysis (all A's or A+'s)

Stats Classes: Intro to Business Stats, Intermediate Probability Theory, Math Stats, Statistical/Machine Learning, Stochastic Processes (in Op.Research dept) (all A's or A+'s)

Other Classes:Microecon theory, macroecon theory, game theory, econometric theory (all A's except A- in game theory)

LOR: 2 year RA with a good behavioral economist (above-average LOR), my tutoring leader at the student center who has written many letters for math/stat PHD and masters students (great letter), math stats prof (ok LOR, didn't converse with him much, I was just top student in his class for probability and math stats), work boss (CS phd who is head of analytics at a real estate predictive/statistical analytics firm)

Programming Experience: R (great), MATLAB (great), STATA (ok), SAS (ok), Scheme (great), SQL (great)

Teaching Experience: 3 year Math/Stats Tutor and have been a TA for probability theory through the student center

-In the Fall, I am returning to school and will be taking a R computing course, Time series course, and a seminar to finish off my major while applying.

Any advice on good programs that might fund a student like myself would be greatly appreciated, as funding is a big deal for me and would not want to do the whole 100k loan stuff. Thanks for the help and advice everyone!

Edited June 30, 2011 by dunleavy005

[dunleavy005]

bump.... does anyone have any advice for me?? any help would be greatly appreciated!!

[coffeeintotheorems]

Berkeley--your UG institution--has one of the best Stats programs in the world. Have you tried asking any of the professors there how competitive your application is?

I'm not qualified to evaluate your chances, since I didn't get into a single one of the seven Stats programs I applied to. You seem to have a strong profile, however, so I would guess you could get in with funding at least at a second-tier school.

The main thing I see lacking in your academic background is a course in Real Analysis. I hear that's pretty important for a competitive application at the top schools.

[dunleavy005]

Thanks Coffee for replying! It is true I have not taken Real Analysis, but I did have to write proofs in my Machine Learning class so I was hoping that would suffice. Am I at a disadvantage because I havent taken math's like abstract algebra and real? Linear Algebra I'm pretty comfortable with (at least for the statistics applications so that isn't a big concern of mine)

[coffeeintotheorems]

I did have to write proofs in my Machine Learning class so I was hoping that would suffice.

Any mathematically rigorous course like that is going to look good. Getting to the heart of the matter, though, it is no replacement for Analysis. For example, in real analysis you make a deep, close study of integration and all kinds of convergence questions related to series and to sequences of integrable functions. These issues arise naturally in theoretical statistics (e.g. delta method and other asymptotics, high-level probability theory), and that is why Stats departments like to see it on your transcript.

If you're mostly interested in an MS and in applied careers then you certainly won't need to have made such a deep study. All I'm saying is that it's academically stronger for some important theory in prob/stats.

Linear algebra is also extremely important for Stats--perhaps even more so than real analysis--because it is useful in both the theory and practice of the field (e.g. matrix algebra and matrix representations of vectors of coefficients to be estimated, to name just one example.)

[sidneysamson]

Any mathematically rigorous course like that is going to look good. Getting to the heart of the matter, though, it is no replacement for Analysis. For example, in real analysis you make a deep, close study of integration and all kinds of convergence questions related to series and to sequences of integrable functions. These issues arise naturally in theoretical statistics (e.g. delta method and other asymptotics, high-level probability theory), and that is why Stats departments like to see it on your transcript.

If you're mostly interested in an MS and in applied careers then you certainly won't need to have made such a deep study. All I'm saying is that it's academically stronger for some important theory in prob/stats.

Linear algebra is also extremely important for Stats--perhaps even more so than real analysis--because it is useful in both the theory and practice of the field (e.g. matrix algebra and matrix representations of vectors of coefficients to be estimated, to name just one example.)

The following is all stuff my stat professors have told me, not my personal thoughts.

For Berkeley and Stanford's programs at least, they require some real analysis background. I know this because people at both schools were telling me. However I guess you could take a crash course the summer before entry and be okay. I should mention too that their programs are also very theoretical so they wouldn't be good fits for you if you're into applied stuff. For master's, it won't destroy your chances of getting in without that background but it would help tremendously if you did.

On the other hand, analysis is a must-have for good Ph.D. programs. It's mandatory and you'd be pressed to find a good program that doesn't force its graduate students to take some analysis or have some under their belts already.

And yeah second what coffee said...linear algebra is crucial.

Also some (at least) Canadian universities do provide funding for master's programs. However, getting into those master's programs is comparable in difficulty to getting into a ph.D. program in the US.

Now personal thought:

Don't bother with abstract algebra unless you find that stuff interesting. Otherwise, it will be a waste of your time.

Edited July 16, 2011 by xiaoxin

[dunleavy005]

Thanks for the replies guys! I'll probably just self-learn Real Analysis theory, not for applications but just so I would understand statistics at a deeper level. Outside of Berk/Stanford, which are considered theory-heavy schools, what are other good American schools that have an applied focus but not overly a "professional" program, as I do want to learn some theory, just not a completely focused theory school as those listed above (I'll probably apply to some Canadian schools, because it does seem that accepted students get funded more often than not).

Edited July 25, 2011 by dunleavy005