Early assessment for a non-traditional applicant

[aluc]

Hi all, I am planning to apply to stats PhD programs for Fall 2021, but wanted to ask for an evaluation early to calibrate my expectations for where to apply, and to see if there is anything else I should absolutely do in the coming months while I still have time. I would like to get into a PhD program, but am willing to put the time into a masters if that's what I have to do based on my profile. I'm open to biostats departments but I enjoy math and theory, so I don't want to settle for a program that's too applied if I don't have to.

Type of Student: Domestic, White Male

Undergrad Institution: Small LAC (Known for lack of grade inflation and not playing the rankings game FWIW.)
Major: Linguistics
GPA: 3.3
Relevant Courses: Calc I (A-), Calc II/Intro to Analysis (B), Calc III (B-), Linear Algebra (B+), Intro Stats (A). All of the math classes were largely proof-based.

Further non-degree coursework at Directional State U (quarter system)
GPA: 3.96 UG, 3.66 Grad
Classes: Real Analysis I,II (A, A), Math Stats I,II (A, A), Applied Regression Analysis (A), Applied Diff Eq. (A), Group Theory (A), Numerical Optimization I,II (A, A), Computational Methods in Stats (Grad) (A), Bayesian Statistics (Grad) (B+)

Additional Info: I'm a fairly good coder. I did a decent amount of CS coursework with good grades and worked for several years in industry doing NLP/ML.

GRE General Test: Taken several years ago before the additional math coursework. Expect to retake with a higher quantitative soon.
Q: 163
V: 170

Programs Applying: Statistics

Research Experience: Working on an applied ecological stats paper with the professor I took Math Stats and Bayesian Stats with.

Letters of Recommendation:  One from undergrad thesis supervisor, one from professor I'm writing the paper with, choosing the third between the professor I took Numerical Optimization with (would probably be generic) or my work supervisor (has a PhD, but not in math or stats).

Coding Skills: R, Python, C/C++

Research interest: Bayesian stats

Main questions and concerns: I had weak undergrad grades, taking math outside of my major. I've developed more mathematical maturity since then and taken a lot more courses, though at a less than stellar institution. Will that upward trend be enough to make up for my undergrad performance? If I can do well, will taking the GRE subject test help convince admissions committees of my mathematical ability? Are there any other major holes I should try and plug by the end of the year? With the non-degree school being on a quarter system, I don't have a huge choice of classes I could take before applying in the fall. Would one more quarter of Real Analysis make much difference?

I'm having a very hard time getting a sense of where I stand. Would I have a shot at PhD programs at the lower end of the top 50 like Mizzou, UCI, UPitt, Ohio State, or Rutgers?

[Stat Assistant Professor]

What does your comment "not playing the rankings game" mean exactly? As it stands, you could probably get accepted into one of the schools that you listed, or at a similarly ranked school like UFlorida or UIowa. I wouldn't say it's a "sure thing," but if you can get good recommendation letters and emphasize the strong performance in upper division math classes, I could see you getting in somewhere around that tier of schools (and lower-ranked ones too).

It may be helpful to mention in your application somewhere -- or have one of your LOR writers mention -- the textbooks you used in your courses (e.g. if you used Casella & Berger for the Math Stat I&II classes, Rudin for Analysis, etc.). Otherwise the adcoms may not have a good sense of how rigorous the coursework was at "directional state university." Scoring well on the math subject GRE would definitely help your application too, but in my opinion, your effort would be better spent strengthening other parts of the application.

Edited March 27, 2019 by Stat PhD Now Postdoc

[aluc]

Thank you for the reply! The school has a contentious history with US News and supposedly has a deflated ranking as a result of not readily providing surveys, certain numbers, etc. I threw in the comment to justify not providing a more precise ranking, but it probably wasn't the most necessary detail.

It's hard for me to tell how well the rigor of the coursework compares to that at a ranked university. The sequence of topics seemed typical in both Math Stats (using Wackerly) and Analysis (using McDonald/Weiss), but there's a possibility the difficulty of assignments and exams didn't line up with that of a tougher university. How much of a "penalty" can I expect if the content is judged to be less rigorous? This is why I thought taking the subject GRE might be particularly helpful.

[Stat Assistant Professor]

49 minutes ago, aluc said:

Thank you for the reply! The school has a contentious history with US News and supposedly has a deflated ranking as a result of not readily providing surveys, certain numbers, etc. I threw in the comment to justify not providing a more precise ranking, but it probably wasn't the most necessary detail.

It's hard for me to tell how well the rigor of the coursework compares to that at a ranked university. The sequence of topics seemed typical in both Math Stats (using Wackerly) and Analysis (using McDonald/Weiss), but there's a possibility the difficulty of assignments and exams didn't line up with that of a tougher university. How much of a "penalty" can I expect if the content is judged to be less rigorous? This is why I thought taking the subject GRE might be particularly helpful.

Yes, scoring the math subject GRE would be helpful but only if the score is 70th percentile or higher or so (otherwise I wouldn't send it). I could see a mid-tier or lower ranked school possibly taking a chance on you, but I also think you would be well-served by getting a Masters degree at a flagship public university or a more reputable private university. That would certainly improve your chances of getting into a more reputable PhD program. It seems as though competition has gotten a lot more fierce since I first applied to grad school (just over 5 years ago).

If you could ace the Casella & Berger theoretical statistics sequence and one or two additional upper division math courses (e.g. some math departments teach first-year graduate/Masters students real analysis from the Rudin textbook) in a respectable Masters program, then that would certainly position you well to get into a Statistics PhD program and assuage possible concerns about your undergrad performance or the rigor of the school where you took your non-degree coursework. Also, you would be a lot better prepared for doctoral study too (in my personal experience in Statistics programs, there is less attrition among domestic PhD students who have Masters degrees than those who only have a BS/BA). The downside of this obviously would be cost and two additional years of study before you start your PhD. But if you *really* want to get a PhD in Stat, then I think you should also strongly consider Masters programs as a "gateway" to that eventual goal.

Edited March 28, 2019 by Stat PhD Now Postdoc

[bayessays]

I think you have a shot at those schools. FWIW, I understood from your description which LAC you went to and I'd personally be pretty impressed by your profile.

I wouldn't bother taking the math GRE, since schools that don't require ot aren't likely to care enough to make it worth the effort.  The upward trend will help. 

I think you have a similar profile to me in a lot of ways, and I think the Florida/Rutgers level would be reasonably attainable for you now - BUT you're going to see a lot of variance and should apply to a range of schools. If you could get into a funded MS program at a good school and kill it, you would be more competitive for top 20 PhDs. If I were you, I'd apply to like 10 schools in that 25-50 range. It only takes one.

A top 10 biostatistics program will probably have plenty of math for you, and I think you would be competitive outside top few.

 

Edited March 28, 2019 by bayessays

[Stat Assistant Professor]

Ah okay, yeah, I think I know which SLAC you are referring to as well. I checked and the average GPA at this school is under 3.2, so I definitely do not think the lower GPA will hurt your application -- as this particular institution is known for lower average grades and has a strong track record of sending its graduates to top grad schools.

OP: I do think you would be more competitive for slightly higher ranked PhD programs with a Masters (considering most of your upper division math classes were taken at a regional state school), but the schools you listed should be attainable for a PhD with your current profile.

Edited March 28, 2019 by Stat PhD Now Postdoc

[aluc]

This is incredibly helpful advice. This sort of feedback is hard to come by without many connections in the field, so thank you both. I'll definitely be adding funded masters programs to my list.

3 hours ago, bayessays said:

If you could get into a funded MS program at a good school and kill it, you would be more competitive for top 20 PhDs. If I were you, I'd apply to like 10 schools in that 25-50 range. It only takes one.

A top 10 biostatistics program will probably have plenty of math for you, and I think you would be competitive outside top few.

@bayessays, just to clarify: do you mean I should apply to these for their MS programs, or that I could reasonably apply for their PhD programs as things stand now?

[bayessays]

3 minutes ago, aluc said:

@bayessays, just to clarify: do you mean I should apply to these for their MS programs, or that I could reasonably apply for their PhD programs as things stand now?

I think it'd be reasonable and a good idea to apply to PhD programs now.  I'm not sure what your budget is, but I would apply to a LOT of programs. I probably wouldn't bother with the top 10 stat or top 3 biostat programs, but I don't think applications below that would be a waste.

You can tell from your profile that you're intelligent, have a strong math background, lots of coding experience, research in statistics - I think you could do really well but you'll have to apply widely to get over the fact that some people are going to be scared by some of your undergrad grades.  My undergrad GPA was a little higher than yours, but my math grades were worse (than your B's at the LAC), and I got into schools in your range, but I also had to deal with rejections from programs that I thought were safeties. 

Look for some MS programs as backups/springboards if you aren't happy with the results.

[penguinbombs]

This might be somewhat "sample size of 1," but I was similarly unconventional applicant this year (liberal arts major with even less direct relevance to statistics), and I did most of my non-degree math work at a regional state school as well. I think my application was kind of a wildcard, but I had surprisingly strong admissions results this year.

My undergrad GPA appeared stronger in an absolute sense, but it was only a hair above the school's average GPA, just like yours. Since I was not focusing on Math/Stats at the time, I don't think those grades played much of a factor in my results. Your strong grades since then and your research experience will be much much more important for your application. 

I agree that you should apply to a wide variety of schools, but if you find a really strong research match (anything related to environmental statistics, or maybe there is some statistics related to voice recognition technology, where your Linguistics background could be considered a plus?), it wouldn't be a total waste of money to try some applications to top 20 schools as well. Some larger programs like NC State might be more open to giving a chance to a less conventional applicant, and UCLA for example highly values people with strong training in adjacent disciplines. Also, given your academic trajectory and what you've done with stats so far, you would probably be a great candidate for a masters at top statistics programs (UChicago, Duke, ect). If you check the admissions statistics at Duke for example, you'll see that the admission rate for domestic applicants is pretty generous.

My recommendation to you would be to nail the Statement of Purpose, clearly tracing your intellectual development and your unique motivation for entering statistics. I gather that for most applicants, the Statement of Purpose is not hugely consequential, but it is much more important for applicants from unconventional backgrounds.

Overall, some admissions readers may be somewhat agnostic about your undergraduate years, but some might see them as a major plus. It will depend a lot on how you present yourself as an applicant. I feel very strongly about opening statistics to people from other disciplines, so I would be happy to talk to you more over DMs if you like!

[aluc]

Hi again,

With deadlines getting closer I was hoping to get some advice on a more detailed list of schools. Two important updates that affect my application: I don't think the paper I am working on will be submitted by the earliest deadlines, and I scored a 164Q/167V on the GRE. Could anyone shed light on whether it is worth taking again? It seems like the general quantitative cutoff mentioned on this forum is 165. It'd be a pain to do it for the few point increase, but if the current score precludes me from even being considered by most schools on my list, then obviously there is no choice.

I find it tough to label departments as reaches or safeties with any certainty, so I'll just list them roughly in the order of their rankings. I took @bayessays advice and am trying to cast a wide net in the 25-50 range, with a few lower ranked. There are a few places I'm not considering for geographic reasons, but just in case are there any obvious Bayesian departments I am missing, or any reaches I should outright remove to save myself the application fees?

• NCSU
• Iowa State
• UC Davis
• UCLA
• Ohio State
• Rutgers
• UC Irvine
• UT Austin
• UPitt
• Mizzou
• UCSB
• Oregon State
• UCSC
• Columbia Biostat
• Penn Biostat
• Minnesota Biostat

Thanks again for any advice.

[bayessays]

You did a really thorough job of finding the Bayesian places. I'm assuming you left out Duke as too much of a reach, not sure if you want to apply there just to have no regrets. Michigan biostat has a decent number of Bayesians and I think you might have a chance getting in there. 

[Joe VL]

Why Fall 2021 man?

[aluc]

4 hours ago, Joe VL said:

Why Fall 2021 man?

Oh wow, my mistake. This whole time I've meant 2020.

 

On 10/24/2019 at 11:31 PM, bayessays said:

You did a really thorough job of finding the Bayesian places. I'm assuming you left out Duke as too much of a reach, not sure if you want to apply there just to have no regrets. Michigan biostat has a decent number of Bayesians and I think you might have a chance getting in there. 

You read my mind. I'll add Duke back on. And I don't know how I missed Michigan. I'll definitely apply there.