Stats phd - realistic for me? If so, what schools?

[sisyphus1]

All, I am very interested in pursuing a phD in statistics and I had a few questions. To give a little background, I was an undergrad Math/Econ major at a decent school (top 10) and my GPA was > 3.9, but this was mostly because I took the easiest classes in order to graduate (i.e. no ‘honors’ level classes, just the bare minimums like real analysis, number theory etc.). I believe I am mathematically apt, but definitely no where near what my GPA would suggest. Upon graduating in 2009, I got a job in quantitative finance, but while I found the applied aspect of the job fascinating, I felt that I didn’t have the necessary mathematical tools. To enhance my quant skills, I decided to do a part-time Masters in statistics, and I will be graduating from the said program May of next year. While I’ve learned a lot from the program, I felt that there was not enough depth (e.g. we learned to use GLM, but didn’t show why it worked), and this led me to look into phD programs in Stats. Now, my questions (sorry for the lengthy background!):

• While I am fascinated by Statistics, I don’t see myself in academia – I see myself working jobs either in the public/private sector (not necessarily in finance though) that utilize high level statistics. Will this be a problem in my statement of purpose if I state that I do not want to be in academia? It seems like a lot of Math phD’s explicitly state that the goal of the program is to prepare their students for a career in research/academics.
  
• For many phDs (especially the sciences), undergrad research is paramount – how true is this for Stat phDs? I have 0 research experience (at least in an academic setting).
  
• Related to above, I feel like my recommendations (from my Master’s program), while good, will not be stellar – I did well in classes (my MS gpa is ~3.9) but didn’t really do anything beyond what was required (part of it was because I was/am working full time). My classes were also quite big and I feel that my recommendations will be impersonal. How important are recommendations? Will my work experience be given any (positive) consideration?
  
• Does the GRE matter for the verbal section? I didn’t have time to study for the GREs so my verbal score is on the low side (580), but my quant is 800.
  
• Which universities have a focus in applied statistics?
  
• Finally, given the above stats, do I have a realistic shot at the really top programs (Berkeley, Stanford, Harvard etc.). If not (which I fear will be the case), what is my realistic reach?
  

Thank you!

[sisyphus1]

Oh and my math/stat-related courses include:

number theory, real analysis, linear algebra, statiscal inference, intro-level stochastic processes, linear regression, time series, econometrics, some data-related classes (e.g. data mining) and some quant finance classes.

I also have experience in R/SAS/MATLAB (not sure if this would add anything to the application).

[OH YEAH]

If statistics is anything like theoretical computer science or mathematics, then no, you don't have a shot at the top programs. Did well in class recommendations mean nothing, and research experience is paramount. The fact that you have absolutely no research experience will really, really hurt you.

[sisyphus1]

If statistics is anything like theoretical computer science or mathematics, then no, you don't have a shot at the top programs. Did well in class recommendations mean nothing, and research experience is paramount. The fact that you have absolutely no research experience will really, really hurt you.

yeah that's what im wondering too (i.e. if statistics departments are like math departments)

[Poisson]

• While I am fascinated by Statistics, I don’t see myself in academia – I see myself working jobs either in the public/private sector (not necessarily in finance though) that utilize high level statistics. Will this be a problem in my statement of purpose if I state that I do not want to be in academia? It seems like a lot of Math phD’s explicitly state that the goal of the program is to prepare their students for a career in research/academics.
  
  

Most PhD programs will state that their goal is to train academics, but realistically there are not enough tenure-track positions for every graduate, and a field like statistics has a lot of private/public sector options. That said, I wouldn't explicitly state on your statement of purpose that you don't want to work in academia, as it will not help your case, and could possibly hurt your chances.

• For many phDs (especially the sciences), undergrad research is paramount – how true is this for Stat phDs? I have 0 research experience (at least in an academic setting).
  
  

I don't think that research experience is as important as the sciences. It is difficult to do statistical research as an undergrad, because it generally requires a fairly sophisticate knowledge of statistical software/theory/applications, which are skills that one does not typically acquire until after they have already started a graduate program.

• Related to above, I feel like my recommendations (from my Master’s program), while good, will not be stellar – I did well in classes (my MS gpa is ~3.9) but didn’t really do anything beyond what was required (part of it was because I was/am working full time). My classes were also quite big and I feel that my recommendations will be impersonal. How important are recommendations? Will my work experience be given any (positive) consideration?
  

Recommendations are fairly important. If you cannot find three recommenders from the Master's program, try to find some from undergrad if they will write enthusiastic letters. Work experience might be a slight positive, but not substantial.

[*]Does the GRE matter for the verbal section? I didn’t have time to study for the GREs so my verbal score is on the low side (580), but my quant is 800.

Quantitative is all that matters. I don't think a single person in my department has a verbal score above 600.

[*]Which universities have a focus in applied statistics?

From what I can tell, Stanford, UChicago, and Harvard are the most theoretical programs (possibly UNC as well). Most other programs are a mixed of theoretical and applied, but biostats programs are generally more applied than pure stats. Virginia Tech and NC State seemed to be the most applied programs.

Finally, given the above stats, do I have a realistic shot at the really top programs (Berkeley, Stanford, Harvard etc.). If not (which I fear will be the case), what is my realistic reach?

Probably not at the very top programs, but it couldn't hurt to apply if you think it would be a good fit and have the funds. Since you have good grades from a top university with a solid math background, I think you would have a good chance at the programs in the next tier like CMU, Duke, Washington, Cornell, UNC, and NC-State.

[cyberwulf]

Strongly concur with most of what Poisson said.

If you're considering biostats, I suspect you would be competitive for all the top programs (Harvard, Hopkins, Washington, Michigan, Minnesota, UNC, etc.) given your strong quantitative background. If you're interested in the biological applications of statistics, you might be better off at a top-tier biostats department than in a second-tier stats program. Also, a biostats degree can lead to a plethora of non-academic research positions.

When I was applying to graduate school, I applied to both stat and biostat departments. I ended up in biostat, and have never once regretted the decision. sisyphus, feel free to PM me if you have more questions.

Edited November 21, 2011 by cyberwulf

[wine in coffee cups]

Wait, so there is consensus that someone with:

• a 3.9+ GPA
  
• from a top 10 school
  
• with a math major
  
• and a good number of relevant classes
  
• and solid GRE scores
  
• and an MS in statistics(!)
  

doesn't have a realistic shot at top 10 statistics programs? For real? Maybe this is just hitting a little close to home and I'm in denial, but isn't that just a tad pessimistic?

[cyberwulf]

Wait, so there is consensus that someone with:

• a 3.9+ GPA
  
• from a top 10 school
  
• with a math major
  
• and a good number of relevant classes
  
• and solid GRE scores
  
• and an MS in statistics(!)
  

doesn't have a realistic shot at top 10 statistics programs? For real? Maybe this is just hitting a little close to home and I'm in denial, but isn't that just a tad pessimistic?

This certainly wasn't my indented message. However, the OP did acknowledge a few potential weaknesses in her/his application, and an adcom at a top-5 program might prefer a student with a 3.7 with glowing letters and a track record of success in challenging (i.e. "honors") courses.

But I have noticed that opinions here do tend towards the very pessimistic when people inquire about their chances for admission to stat/biostat programs. The reality is that, at least in comparison to other fields like math/physics/CS, the student talent pool in statistics and biostatistics remains quite shallow, and you don't have to be a superstar to be admitted to even the best programs.

[dendrogirl]

I think you could angle your work experience to make up for your lack of undergrad research. You're reasoning for going to graduate school seems really sound (you had a job that you found interesting, but your lack of math knowledge became a barrier). The fact that you went through a master's program while working and you still want to learn more is great. It doesn't matter that you "didn’t really do anything beyond what was required" because you were working! In your personal statement, try to connect what you learned in your MS to your job and describe why you feel you need the PhD. Before you ask for letters of recommendation, sit down with your professors and talk about your motivations for getting a PhD. Ask their advise. Go over your application highlights with them. Then, when they sit down to write your letter, they'll have a fresh idea of your motivations and you'll hopefully get better letters.

I wouldn't state that you don't want to go into academia because its just not necessary. Just describe your true motivations. If you go onto get a quant job, you'll be doing "research" (i.e. coming up with new techniques), and you need a PhD to excel as a quant. All programs (top to bottom) want students who will complete the program. Your work experience and motivations are strong indicators that you'll be able to finish.

That being said, I would apply to a range of schools. The top ones and some middle ones. More important than the rank of the program is the availability of faculty doing research in econometrics (or whatever you want to study). These days, even middle programs are getting 500+ applicants for ~10 positions.... Yikes!

[dendrogirl]

. The reality is that, at least in comparison to other fields like math/physics/CS, the student talent pool in statistics and biostatistics remains quite shallow, and you don't have to be a superstar to be admitted to even the best programs.

I'm not sure about that... But I am offended...

Edited December 3, 2011 by dendrogirl

[cyberwulf]

I'm not sure about that... But I am offended...

I suppose it depends on your definition of "superstar". But I guess what I mean is that if you:

1) Went to an undergrad school with a decent reputation (say top 100 in US)

2) Graduated in the top 20% of your major (not necessarily math)

3) Have the math prerequisites

4) Got a GRE quant score > 85th percentile

5) Have positive letters that say something interesting about you

Then you will at least be in the discussion for admission at the top biostat programs, and many good stat ones (the top stat ones might have a slightly higher bar for math background/performance).

In other words, you don't need to be:

1) The top math student in your cohort (that girl/guy's probably going to a Math PhD program)... nor even a math/stat major

2) A graduate of an Ivy League or elite liberal arts college

3) Published

4) The best student any of your professors has ever seen/worked with (well, unless you're coming from a school nobody's ever heard of)

So, I stand by my "don't have to be a superstar" assertion. Of course there are a lot of good students in the field, but they are not (warning: math/semantics joke) uniformly exceptional.

Edited December 3, 2011 by cyberwulf

[wine in coffee cups]

Very helpful followup comments - thanks, cyberwulf and dendrogirl!