Stat Assistant Professor

What are the reputations of your undergrad and your Masters institutions (i.e. range of rankings in USNWR)? It would make a huge difference in your chances whether you attended Duke or Emory for your BS/MS vs. Charleston Southern University. That said, your research experience is good (one publication is great) and it seems like your academic performance is solid, regardless of the prestige of your undergrad/Masters school(s). I would say Pitt and UC Denver are very safe choices and you could aim higher (e.g. schools like UCLA and UMinnesota Biostat do not seem totally out of the question).

How did you perform in your undergrad? Assuming you did well enough, I would say you have a very good shot at a top 20 Statistics (not Biostat) PhD program. The top 10 Statistics programs (i.e. University of Michigan and above) might be a reach because your math background may be a bit lacking compared to other outstanding applicants, but you could apply to a few of them and see what happens (you could get lucky). Just don't make your list of schools *too* "top-heavy." Now, if you were to apply to Biostat PhD programs (and not just pure Statistics), you would probably be able to get into a top 5 Biostat program and definitely any program below the top 5. I know someone who got into Harvard Biostatistics with a non-traditional background similar to yours (they majored in Biology as an undergrad, but then took the prerequisite math courses and completed a Masters degree in Statistics -- and this was all *while* working full-time at a software engineering job, and their MS school wasn't as prestigious as yours either).

I think your GRE score is perfectly fine, actually. A 170 on the verbal is also impressive, as most applicants to Stat PhD programs do not score nearly this well. The Q score of 164 is also fine -- I wouldn't bother to retake it.

While Statmaniac has made some valid points, methinks that they have extrapolated a bit too much based on their personal research. For example, they dismiss detailed study of GLM's, but argue that information theory and functional analysis are things that are "more important and fundamental to learn." Many stat students can get by and publish in top journals/conferences without having taken an entire course on information theory or functional analysis -- they can pick up on the things from these areas that they need for their research *if* they ever need it (like the various entropies and divergences, for example). And students who are doing more applied statistics have little use for those subjects. Anyway, it is a matter of opinion what is "most important and fundamental."

To OP: I wouldn't worry too much about the coursework. Once you pick a PhD advisor, they will advise you on additional courses to take for your research (if any). For example, the department I got my PhD from was well-known for its work on theory for MCMC. If you wanted to work with one of the MCMC professors on this, they would ask you to take a functional analysis class in the math department. So you can always take courses that are immediately relevant to your research, especially if your PhD advisor encourages you to do so. I think Statmaniac makes a few good points, but I disagree with some of what they are saying and think some of the things are a stretch (for example, their comment, "But why does it need to waste students' time and effort if there is something more important and fundamental to learn?" seems to be a little bit too subjective). Also, most Stat/Biostat departments are incorporating more "modern" topics into their curriculum.

Regarding the utility of measure theory, etc.: the relevance of it to a PhD student's research depends on what the research is. If the Applied Math or IE PhD student is doing their dissertation on mathematical/computational biology and neuroscience or on queuing theory/mixed-integer programming or something like that, then a lot of measure theory and functional analysis is not going to be directly useful to their research. Yet most programs in Applied Math require two semesters of graduate-level analysis. The ones that do not are the elite programs where the students entering have *already* taken graduate-level courses like measure theory, functional analysis, etc. Most PhD students (at least domestic ones) have not taken these classes as undergrads, hence why they are being taught it in their PhD program. In general, I support PhD students needing to take two semesters of measure-theoretic probability, although a deep knowledge of it is certainly not needed for all statistics research (especially not if the research is more on the applied side of stats). I myself have published in top stat journals like the ones you mentioned, and I don't use that much measure theory (and I work on statistical methodology *and* theory). Other researchers might need to use more of it -- it really depends on the topic. In spite of this, I have no complaints about the coursework I had to take as a PhD student (which included two semesters of measure-theoretic probability), as I mostly taught myself the stuff I needed to know for my dissertation research. For my postdoc and now being faculty, I also have to constantly teach myself new things.

There are *many* faculty in Statistics/Biostatistics departments conducting research and publishing in the top journals and the top ML conferences in the areas you mentioned (high-dimensional statistics, causal inference, etc.). Most students are capable of self-teaching themselves these topics (or taking electives to gain some exposure to them) after they begin their research. I think it is somewhat unreasonable to expect programs to teach students all they need to know for their research through courses (when a PhD is largely about teaching yourself and contributing new research that isn't already covered in classes) or to tailor coursework around what's "trendy" at the moment. For one, not all students are interested in the same things. Classes on optimization likely have little relevance to students who are interested in applied probability/stochastic processes, for example. Nevertheless, as I also mentioned above, most Stat and Biostat programs are taking it upon themselves to 'update' the curriculum to also include the more current topics. Secondly, the other fields you mentioned might have a lot of coursework too that isn't directly applicable to students' research. For example, in an Applied Math PhD, students might need to take two semesters of graduate-level Analysis with measure theory, Hilbert spaces, functional analysis, etc., as well as a lot of classes like numerical analysis, partial differential equations, etc. These students typically also need to pass several written qualifying exams. An EECS student might need to take classes on computer architecture, theoretical analysis of algorithms, etc. But if these Applied Math or EECS students then go on to conduct research in machine learning or global optimization, then it's not like all of their classes are immediately relevant to their research. Now, some of the top programs in these other fields (like Stanford CS, Princeton Applied Math, Berkeley EECS) likely do keep the coursework requirements to a minimum (so most students are largely done with classes by the end of their first year, and students also have greater flexibility in what classes to choose -- so they probably do only take a few classes that are immediately relevant to their research). But that's mainly because the types of students that are admitted to these kinds of programs have already completed extensive graduate-level coursework as an undergrad and have already done research as an undergrad that got published in major journals or conferences. But these are exceptions rather than a general rule. Most PhD programs in Applied Math, EECS, and IE have at least two years of coursework, and certainly, not all of it is relevant to every student's research.

I didn't know that undergrad probability and statistics were hard requirements for MS programs in Statistics. Since you have a Bachelor's in a STEM field, I would think you would have a decent shot at most Statistics MS programs provided you have a decent GPA and general GRE Q score, and grades of B- or higher in Calc I-III and Linear Algebra. Nevertheless, if an MS program has those classes as prerequisites for admissions, then if you want to attend that program, you'll need to take those classes for a letter grade from a college/university. Given the current pandemic, it would be fully understandable for you to take these classes remotely, but they should be offered by a university for a letter grade and for which you can obtain an official transcript (not through an MOOC). If the programs that interest you don't explicitly list probability and statistics as prerequisites, then it might be worthwhile to take the MOOC to prepare for these classes when you take them at the Masters level.

I assume the requirement is waived for Stat PhD programs. I think only Stanford had it as a requirement before, and the Subject test was used as a screening tool (so anyone who scored under [x] percentile was not seriously considered for their program). In my opinion, it should not be required to begin with for Stat programs, since it tests things like abstract algebra, number theory, topology, etc. that have little relevance to statistics. But clearly, some elite PhD programs feel differently. I assume that Stanford and other schools that "strongly encouraged" the subject test will just weigh the other factors more heavily -- GPA, grades in upper division math classes, class rank, letters of recommendation, reputation of undergrad institution, research experience, general GRE, etc.

In addition to the above: I will also add that a lot of Statistics and Biostatistics programs do recognize the need to "update" the graduate curriculum to include more "modern" topics. Most statistics/biostatistis departments are aware of this and have either already done so or are in the process of doing so. So while you might still encounter a lot of 'classical' topics such as UMVUE, UMP test, James-Stein estimator, etc., the coursework often *does* give a splattering of more recent topics too, like high-dimensional regression, multiple testing with FDR control rather than FWER, etc. But also, there is only so much that you can cover in classes. The subject matter of each class (e.g. probability theory, linear models, etc.) has enough material that you could easily spend a whole year or two covering subtopics in depth. You have to pick and choose what to emphasize and trust that once you give some basic foundation, the students will be able to learn other things on their own and pick up what they need for their own research.

Good point as well. There are a lot of PhDs in various disciplines who end up going into high school teaching, especially at private schools where the money is comparable to (or even more than) the salary of a math prof at a directional state school. Also, the payscale is higher for teacher who have higher levels of education.

There are plenty of pure math PhD's who began their careers as high school teachers. Shouldn't be a problem. The main issue is if you will be able to score decently enough on the math subject GRE, which takes a decent amount of preparation (especially if you have been out of school for awhile) -- you really need to master subjects like abstract algebra, real analysis, linear algebra, etc. to do well on this test. You also need strong letters of recommendation, in addition to good grades in upper division math courses. I doubt the M.Ed. would be a "black mark" on your transcript -- most likely, it would be viewed as a slight positive, or at worst, viewed neutrally. The main things you need to convey in your application are mathematical maturity and research potential. That said, the academic job market for math PhDs has been incredibly tight for many years and is likely to be more so now because of covid. Even before covid, doing 4+ years of postdocs was the norm. There are math PhDs from Harvard, Stanford, Princeton, MIT, etc. who have had to do multiple postdocs before taking a job at Montana State or Kansas State. I'm not saying that these are bad schools, just that it underlies that if you want to stay in academic mathematics, you *really* need to be in it for the long-haul as well as geographically flexible. You can't be like, "Oh, I would never take a job in [state/geographical region]." The situation is a bit better in statistics/biostatistics, but pure and applied math are really competitive.

I think admissions is very competitive to get into Cornell and Yale too. Of the 3 Ivies mentioned, I think Columbia and Yale are reaches, and Cornell is a slight reach. I think Yale especially is a bit underrated relative to its most recent ranking in USNWR. This department has some heavyweights like Harrison Zhou, Andrew Barron, and John Lafferty, and it has placed some of its former PhD students in TT faculty jobs at UPenn Wharton, UC Berkeley, University of Chicago, as well as other good schools like UIUC, etc.

Thanks for the clarification. I think you should aim for schools around the general tier of Ohio State, UIUC, University of Florida, Rutgers. I think you would have a decent shot at those schools. You can apply to a few that are ranked above that (as you correctly conjectured in your original post, your chances would be better at large state school programs) and a few ranked below 40 like UConn, UIowa, etc. and maybe a few even below that just for good measure.

It is really hard to say. I will say that for most of the schools in the top 40 or so (i.e. Rutgers, UIUC, UFlorida), *most* of the international students are from top universities in their respective countries like ISI, Peking, Fudan, Tsinghua, SNU, etc. You will probably be at a disadvantage compared to these students, which I why I think top 40 might be a tad of a reach (unless you get a Masters first, that is). I'm not sure how UC Irvine compares to the other aforementioned schools, but if it is similar to them in terms of student makeup, then UCI would also be a reach for you. I think you might have a shot at many schools around the rank of Florida State and below, and your chances should be fairly okay for schools around the rank of Virginia Tech.

I don't think you need to do a Masters in math or stats. However, I would recommend taking a semester of real analysis, since many schools do strongly recommend this. Ideally, you would be able to take real analysis this fall (or is it already past the add/drop period at your school?), and you can point out in your SoP that you are currently enrolled in Real Analysis when you apply. If not, maybe it will be just sufficient to state in your SOP that you plan to take Real Analysis in the spring, *but* you have taken other proof-heavy courses like abstract algebra and stochastic processes, so you have received training in doing advanced math and writing mathematical proofs. Also ask a LOR writer to point out that you have taken these advanced proof-heavy classes. Of the schools on your list, I think Penn State, Purdue, and UCLA are "safely" attainable. Wisconsin also seems like a relatively safe bet. I could see you getting into Duke, Michigan, or Washington, but the admissions process might be a bit noisier at schools of that tier, so it's good that you're applying to several of these. Columbia is a reach but it doesn't hurt to try that one.

In that case, your list looks reasonable. However, I think Michigan State is more of a "target," while UC Davis is a "reach." UCD is actually a very good, competitive program -- one of their PhD graduates had 3 papers in AoS/Biometrika/JRSS-B and during the 2019-2020 academic job cycle, they got a ton of campus interviews for Assistant Prof positions at all sorts of top-tier departments like University of Washington, UNC-Chapel Hill, UPenn Wharton, Cornell, etc. I also think you should add more programs in the range of 60-80 of the USNWR rankings to be safe -- schools like University of Missouri and University of South Carolina, for example. I would apply to some Masters programs too, if I were you. If you were to get a Masters degree from a respectable R1 school like Rutgers or GWU, you might be able to aim a bit higher. The downside of the Masters is having to pay tuition for two years of a MS.

Could you give us an idea of the tier/ranking of your large public university? This matters a lot. If your public institution is a school like UNC-Chapel Hill or Georgia Tech, then I think you may be able to aim a little bit higher for your targets and your "reaches" may be attainable (despite being an international student).

3.85 GPA from an Ivy and a bunch of math classes including functional analysis, measure-theoretic probability, and stochastic processes, plus solid research exprerience. This is a very strong profile. If I were you, I would apply to mainly top 15 USNWR Statistics/Biostat PhD programs. With strong letters of recommendation, I think you will be able to get into virtually all Biostat programs including Harvard and JHU. For statistics, you have a very good shot at UC Berkeley, UChicago, Carnegie Mellon, UPenn Wharton, Duke, etc. and I wouldn't be surprised if you also got admitted to Stanford too (I heard Stanford Statistics Dept is waiving the math subject GRE requirement this year for its PhD program?).

Recommendation letters aren't as important for Masters programs as PhD programs. With a 3.96 GPA and having met the math requirements for Statistics MS programs, you should be able to get into most Statistics Masters programs with your current profile. If you are contemplating a PhD eventually, then you probably do need to take some advanced proof-based math classes, including real analysis. If you could fit some in to your senior year schedule and/or take two such classes (one of which has to be real analysis) in your Masters program, then you will have a better chance at Stat PhD programs.

What does your percentage at ISI roughly translate to on a 0.0-4.0 GPA scale? From what I gather, it isn't like 90-100 is an A, 80-90 is a B, etc. like in the U.S., because I saw that somebody who had 88 percent from ISI graduated "First Class with distinction," and they were admitted to UPenn Wharton. So is anything over 80 considered an A at ISI? I wouldn't be surprised if an 80 is considered very good at ISI given how rigorous this institution is. With a bit more context about your academic performance (especially in comparison to your peers at ISI), we can give more tailored advice. Thanks!

Good to know. The academic job market will be incredibly tight for the next few years because of covid, so I am wondering if it will get to a point that VAP's are basically required to get a TT job at teaching institutions. OP: if you think you might be interested in biostatistics, then you should apply to biostat PhD programs as well. You will likely be able to get into some good biostat PhD programs. Although you expressed interested in teaching, you may find that you love research too and may want to focus on jobs at research universities later on. And if you still have a strong interest in teaching but the program doesn't provide teaching opportunities, you can try to get some experience mentoring and/or collaborating with undergrads (I think that also counts as a positive for job applications to teaching schools) or you can consider doing a VAP (which are usually a combination of teaching and research) after your PhD rather than a purely research-focused postdoc.

That is good to know. Did these Biostat PhDs go directly from PhD to the teaching oriented institution, or did they do a VAP first? Because when I was on the market, I looked at a few job postings for PUI's/non-doctoral granting universities (didn't end up applying to any), and they asked for teaching evaluations to be submitted with the application. Maybe some PUI's like the idea of hiring Biostat PhDs because the interdisciplinary research they're involved with can also involve undergraduate students...

I was referring to the USNWR top 30 rankings inclusive of the biostat programs, but my comments were mainly concerning the statistics PhD programs. If you are interested in biostatistics, then you could apply to those too. Actually, if you applied to Biostat PhD programs (and not just Stat), you would probably be able to get into those ranked in the top 10 -- so inclusive of biostat, you would be able to get into more of the top 15 programs. The reason is because unlike Statistics PhD programs, admissions for Biostat is clearly skewed in favor of domestic applicants (the top Biostat programs only admit a few international students every cycle). Your math background would also be especially appealing to a Biostatistics PhD program. Now, if your ultimate goal is to teach at the college level, then I think you are more likely to get the requisite teaching experience as instructor of record in a Statistics program than a biostat one. PhD students who are looking for jobs at teaching institutions (e.g. PUIs) must have taught at least one course as an instructor of record to be seriously considered for tenure-track jobs at these types of schools (many will ask you to submit a teaching portfolio/teaching evaluations as part of the application). I don't have firsthand experience with Biostat PhD programs -- my degree is from a Stat department, so I am not sure if Biostat PhD students are able to get this kind of teaching experience... based on my postdoctoral experience in a Biostat department, the PhD students were overwhelmingly supported through research assistantships. But Stat PhD students routinely have the opportunity to teach undergrad stat courses. (Addendum: if any prospective Biostatistics PhD students are interested in getting jobs at teaching institutions, they may be able to do so if they do a visiting assistant professorship where they can shore up some teaching experience)

1. Recommendation letters from those professors should be fine. Maybe ask them to emphasize that you're also a strong math student who has taken some advanced math like Analysis I, complex analysis, and optimization. 2. I think admissions committees will view someone with a physics BS from a top 10 undergrad and a ~3.9 GPA very favorably actually. 3. In your case, I don't think you need to say much about transferring PhD programs. It would be different if you were transferring from one Statistics PhD program to another, or if you had already completed 3 years in a PhD program. Since you completed only one year, you can just say, "I will be completing my Masters in computer science in May 2021 and am looking to further my education in Statistics" and then talk a little bit about what research you could potentially see yourself doing. 4. Of the schools on your list, I would say your chances at Duke, UNC, NC State, Wisconsin, UM and TAMU are all very good for your profile, since you have the math background, a degree in physics (and soon degree in MS), and a strong pedigree. I would say your chances are well above average at all those schools. Harvard is a possibility too, but they (like many of the Ivies and UChicago) seem to admit very few domestic students, so it's not a sure thing. But you should definitely apply there.