Statistics PhD Application Evaluation

[az25340]

Hi all,

I'm planning on submitting applications to statistics PhD programs this upcoming fall, and I was hoping to get some feedback on the strength of my application. My research interests are in statistical machine learning and latent variable models, though I could see myself branching out in the future.

Undergrad Institution: Top five liberal arts college
Majors: B.A. Statistics & Economics
GPA: 3.8/4.0

Type of Student: Domestic Male

GRE General Test:
Q: 167
V: 167
W: 5
GRE Subject Test in Mathematics: Not taken yet.

 

Programs Applying: Statistics

 

Research Experience:  

My primary research experience has been in my job as a bioinformatician. My main focus has been trying to develop subtypes for disease progression and predict disease progression more generally. I'll be an author on a paper that I'm hoping will be submitted for publication by the time applications roll around. I also did an honors statistics thesis in undergrad comparing the performance of Bayesian inference algorithms.
Awards/Honors/Recognitions: 

Graduated magna cum laude, award for top statistics thesis in undergrad.
Pertinent Activities or Jobs: 

18 months of work as bioinformatician at time of submitting applications.

Letters of Recommendation:

I have three professors from undergrad who I believe would write strong letters, and I would also expect to be able to get good letters from my boss and the director of the department at work.

Coding Skills: 

R, Python, Java

Relevant Classes, Grades:  

Undergraduate Classes:

 

Mathematics -  Linear Algebra (B+), Probability (A-), Combinatorics (A-), Real Analysis (A-) (I took AP Calculus and Multivariable Calculus in high school, so they won't show up on the college transcript)

Statistics - Intermediate Statistics (A), Theoretical Statistics (A), Data Science (A-), Advanced Data Analysis (A-)
Other Relevant Courses: Intro Computer Science I (A-), Intro Computer Science II (A-), Machine Learning (A), Advanced Econometrics (A-), Data Structures (A)

 

I'm still torn between going into industry or academia after grad school, though I've been starting to lean more towards academia.

 

Additional comments/concerns:

1. Looking at my application, I believe my biggest weakness will be my math background. The linear algebra class was in my first semester of college, but I still don't very many other courses for them to go off of, and I didn't ace any of them. Do you think this could torpedo my application? Additionally, given that I am out of college now and am not interested in getting a master's before a PhD, are there any steps I could be taking between now and the fall to strengthen this part of my application? Would a good score on the GRE Math Subject Test help? If so, what score should I be looking to get? I also think I could retake the General GRE to go for the perfect score on the math section -- I was stupid and didn't keep track of time when I took it the first time -- but I don't know if that would make a big difference.

 

2. Are the top-30 of the USNWR a reasonable range of programs for me to be applying to? I'm basing these schools primarily based on conversations with my professors in undergrad, who recommended CMU, Columbia, Cornell and some others as possibilities, however I don't know where in the reach-safety range these would fall. Realistically, I don't believe I'd be all that competitive for the Stanfords/Berkeleys, but I'm not opposed to submitting an application just to have a shot at it. I've thought I would probably be more competitive for the 10-20 range, with Duke, Wharton, Columbia, NC State, UNC, etc. being interesting to me. Do you think these programs are good potential fits? If not, where should I be looking instead? Ultimately, while a prestigious program is nice, I'm trying to prioritize finding a program that will be a good fit for my interests and abilities, and will give me opportunities after completion.

 

3. Maybe I don't know too well where to look, but how should I be evaluating which programs would be best for me beyond the rankings? I know there is a split between more theoretical and applied approaches, but I've been having a harder time telling which programs would fit into which groups, and so on.

 

 

Thanks for reading, and I appreciate any comments! 

[bayessays]

I think the schools you listed are reaches with your math background (CMU, Wharton, Columbia I would say are extremely unlikely and the others are slight to moderate reaches - I think NCSU/Cornell/UNC would be the top range you should be applying to as reaches). Have you thought about biostat programs? You could probably go to a top 10 biostat program.  Your math background would be more typical there and they'll value your bioinformatics background. I'd probably start looking at biostat programs ranked 3-15 and stat programs in the 25-60 range.

[Geococcyx]

My math background was pretty similar to yours; it was pretty light, and I'd taken linear algebra my first semester of undergrad as well.  Bayessays pretty much hit the nail on the head for me, since I got waitlisted and admitted at NC State, and was likely waitlisted and admitted at Duke.  You went to a better school, and did better in Real Analysis, so you might end up being marginally stronger.  Even so, biostat programs are definitely places to look -- just as an example, UNC biostat has Dr. Kosorok doing machine learning, and lots of statistical genetics folks that might mesh well with you. 

Beyond that, I'm sure folks can help you with specific fits as needed.  I recall Cornell being good at high-dimensional statistics, which meshes pretty well with your topics, so it makes sense that your professors and Bayessays recommended them.  Wharton takes very, very few domestic students each year (and very few students in general), so they're a tough for most people to get in, hence why applying there as a reach might not be as helpful as you'd like.  

As to your other questions, you don't need to retake the regular GRE.  Unless you really think you would knock the Math GRE out of the park, and it would be a really clear benefit to you, then I wouldn't bother either (provided you aren't applying to Stanford, of course, in which case you'd need it).  The theoretical/applied split is hard to tell, but you can pick up some ideas of it in past posts on these forums, or else just by asking.  UNC statistics seems to have a reputation as pretty theoretical, and being good at stochastic processes and whatnot.  NC State would tend to be more applied.  Places like Columbia and CMU probably fall somewhere in the middle, and Duke might fall a little more applied than those two.  Ultimately, it's a case-by-case basis, though, and just because you might go to UNC statistics doesn't mean you couldn't work with somebody on, say, Kalman filtering and data assimilation for numerical weather prediction.  

(final note:  Bayessays has been around rather longer than I -- if they or another older poster argue with anything I say, I'd take their answer above mine)

[ENE1]

I wonder whether an excellent maths GRE score could make up for your light maths background. If so, there might be a case for taking the maths GRE.

[bayessays]

There are like 5 schools that really look at the subject GRE.  Based on OP's math background, they'd probably have to study many 100s of hours to get a score that would help them and even then, they're competing against a bunch of international students who get near perfect scores at the top schools.  Probably not worth it.

[az25340]

Thanks for the great feedback! It's much appreciated. I'll definitely need to look at biostat programs more closely. As I start to learn more about them, I do have the following questions:

• I was surprised to hear that biostat programs tend to have lighter math requirements than their stat counterparts. Do you know why this might be? Initially, it seems like there might be more people coming in with a greater science background rather than pure math, but I'm not sure if that would explain it all.
• Do you know how big the differences in curricula between the stat and biostat programs are? My basic understanding is that there aren't usually too many differences in the fundamental coursework covered by the programs, but that biostat may place a heavier emphasis on study design or other biology/health related topics.
• Does a statistics degree versus a biostat degree make a big difference in terms of placement outcomes after completing the degree? Do the range of options change depending on whether you go into academia or industry? As far as academia goes, I know that at my school, many of the stat professors had biostat degrees, but I'm not sure how common it is for biostat PhDs to teach in statistics departments elsewhere. I don't know as much about biostat PhDs in industry beyond pharmaceutical companies. I think I ultimately hadn't been looking too much at biostat programs due to worries about being pigeonholed after completing the program, so it would be great to hear a better perspective on what options are out there.

Thanks for your thoughts!

[bayessays]

The basic answer to all three of those questions is that, in general, biostatistics departments have less of the very theoretical end of the spectrum. Math requirements for admission are more lax because you likely won't have to take years of extremely theoretical probability classes.  You probably won't be publishing in Annals of Statistics. You might take a class in survival analysis instead. Not a huge impact on your placement - if you want to do theoretical work and get a job in a stats department, make sure you go to a biostat with professors that do that type of work. If you want to teach undergraduates, make sure to go to a place that will give you the opportunity to do so.

 

[Stat Assistant Professor]

I think you should have an above average shot at NCSU and Duke. After all, you did go to a top liberal arts college and did pretty well in the math and stats classes you did take. Your profile is by no means a "shoe-in" at these places, but I think that if you can get excellent letters of recommendation, I wouldn't discount your chances at either of these schools. Your profile reminds me: there is an outstanding statistics researcher James Johndrow (now an Assistant Professor at Penn Wharton) who got his PhD in Statistical Science from Duke even though his undergrad degree was in Chemistry (also obtained from a top liberal arts college). And I also know of a young woman whose Bachelor's is in Psychology/Pre-med from Columbia but she also got her PhD in Statistics from NCSU (now a postdoc at Columbia). I think that Duke and NCSU are more open to accepting applicants from "lighter" math backgrounds and different majors than other places -- I know of at least a couple students/alumni from NCSU and Duke Statistics who did not have extensive math backgrounds. However, they did come from strong pedigrees, so your chances may be inversely proportional to how prestigious your undergrad is (that is, you can possibly get away with a lighter math background if your undergrad institution is very prestigious. But if you went to an unknown school, then you need to have very strong performance in math classes to be competitive).

However, UPenn Wharton, CMU, and Columbia may be tough for you to break though without a lot more math, and I would recommend that you apply to a wide range of PhD programs, like the other posters have suggested. I also think you can definitely get into a top 10 Biostats PhD program, no question, if you think that this aligns better with your interests than traditional Statistics programs.

Edited January 21, 2020 by Stat PhD Now Postdoc

[rfan]

@Stat PhD Now Postdoc I can confirm that what you have said about Duke is at least somewhat true. Over the past 2 years, almost every admitted (matriculated, that is) student was a math major. However, it was frequently paired with another major (think economics, psychology), and the department seems to value that educational diversity. For what it is worth, the few students without as significant of a math background found the first year coursework exceedingly challenging. It has become the case that students with less mathematical maturity take some of the master's classes (for example, master's inference instead of PhD inference or delaying Measure Theory for a year), which delays the qualifying exams until the following year. I think there also is an associated challenge that, from my observations, professors tend to prefer the more mathematically inclined students, so those without such a background often have less advising options (although definitely still some!) available. That's just something to think about...

[az25340]

Hi all,

I figured I'd follow up with a question regarding the list of programs I plan to apply to. The list as it stands right now is as follows:

Biostat:  Harvard, Johns Hopkins, UNC Chapel Hill, Minnesota, Columbia, UPenn, Brown, Duke, Boston U, Pittsburgh, U of Rochester

Stat: NC State, Yale, Ohio State, Rice, WUSTL, UT Austin

 

Given my background (I also bumped the GRE score up to 168Q and 170V if that makes a big difference), does this list seem reasonable? I know it's a lot of programs but I'd ideally like to end up with a couple options and I'm concerned with the size and selectivity of some of the programs. Though I think I would prefer a smaller program, I would ultimately be happy going to any of the programs if accepted and want to be sure I don't get shut out. Would I help myself by focusing more on larger programs at public schools, or applying to programs further down the rankings?

Thanks for your feedback!

[kingsdead]

I think that's a good list (for biostats at least, I don't know as much about stats). I'm confident you'll get into many of the biostats schools. For reference I think your profile is better than mine was and I got into many of the schools on your biostats list. Under normal circumstances I would say your biostat list might be a bit bottom heavy, but given that funding seems a bit more uncertain this year it might be good to have that just in case.

[az25340]

@kingsdead Thanks for your response -- I appreciate it!