Stat Assistant Professor

Agreed with @StatsG0d. It would be very beneficial to take upper division, proof-based mathematics courses, especially Real Analysis and maybe upper division proof-based linear algebra. The first two years of coursework in any Statistics PhD program will be quite theoretical, and you will be required to take classes like measure-theoretic probability, linear models, advanced statistical inference, large sample theory, etc. At most Statistics PhD programs, it is possible to do a more "applied" dissertation with little/no theory -- this is true even at heavily theoretical programs like UPenn Wharton or Stanford (usually there will be at least a few faculty members who are more applied and don't really do theory, and you can ask them to supervise you). But to even get to the research stage, you need to get through some pretty theoretical classes and pass qualifying exams on some of this material. The qualifying exams are also challenging, and there are always a few students who do not get a PhD pass on this exam and have to retake it or leave with a Masters. It doesn't really matter what the rank of the program is; the vast majority of ranked Statistics PhD program will be this way. More advanced math classes -- *not* more undergrad statistics classes -- would be the best preparation for a Statistics PhD program.

@Bumblebea This is all very fascinating to me. I read your post with great interest (although I am not in the humanities, it is always interesting to read others' accounts who have also braved the academic job market). I definitely agree with you that aspiring faculty should give themselves a time limit for amount of time they are willing to be on the academic job market. In my field, it is commonly the case that PhD graduates need to do a postdoc or two in order to land a TT position. In my specific field, I would strongly encourage PhD alumni to weigh their other options if they come up empty-handed on the academic job market after a second postdoc (~4-5 years out of the PhD). This is not to say that my field is comparable to English lit, but in any field, there does seem to be a time limit before a PhD becomes too "stale," and it's best to cut your losses. I also agree with you that it may be worthwhile to pursue a PhD even with all of the sunk costs, but it is also important to have realistic expectations re: job market and not to romanticize the professorship (or any job). There can certainly be a lot of tedium in academia, just like any other job, and I often have to devote a substantial amount of my day-to-day doing paperwork and working on things unrelated to research/teaching. And you do have to deal with constant rejections (rejections for articles, grants, book contracts, job applications, etc.). I think this can be tough for some people's egos, if they are used to excelling in school. I've gotten used to it by now, but it was really demoralizing for me at first. Anyway, I don't want to venture "out of my lane" too much. But I wanted to say that I really appreciated your detailed post, and a lot of the things you said ring true universally in academia, not just in English lit!

I don't have any first-hand experience working in big pharma, but I'm wondering if there is a difference between the jobs that are available for Biostat Masters holders vs. Biostat PhD holders? At your current workplace, do the employees with Biostatistics PhDs get to do more "stimulating," methodological/statistics work than regulatory writing? I know some Biostat PhD graduates who work as research scientists at Eli Lilly and Company. Is the type of work they do materially different from the work that gets done by Biostat MS holders? Also, I have definitely seen Biostatistics PhDs can go into data science, some at companies like Google, Amazon, etc. So I would think that Biostatistics graduates are not constrained to only working in biostatistician fields.

Besides the required classes for your Masters program (I'm assuming that this entails 2 semesters of Casella & Berger-level mathematical statistics, 2 semesters of applied statistics, and a semester Linear Models), I recommend that you take: 2 semesters of real analysis proof-based upper-division linear algebra one other advanced math class of your choosing, if you can fit it in your schedule (if not, then the 3 math classes above should suffice). If you can afford it, it may be worthwhile to enroll in a math class for credit at an accredited school over the summer *before* starting your Masters program. That way, you can knock out one of the classes above over the summer and you can take other classes that interest you in your Masters program. Instead of taking PhD-level Statistics classes (i.e. those classes *beyond* the Masters-level, e.g. Advanced Inference or Measure-Theoretic Probability), I agree with previous posters it is a better idea to take statistics electives that interest you and/or more math classes. I think if you have at least 2 semesters of Analysis and a semester of proof-based linear algebra, you should see good results for PhD admissions in Statistics. A 3.84 GPA in Physics (a tough subject!) from an Ivy is definitely nothing to sneeze at. Those factors --plus strong performance in a Masters program and A's in two semesters of Analysis -- should put you in great shape to get into a good Statistics PhD program. I'm not sure if it will guarantee admission to Columbia, Harvard, or Stanford, but I wouldn't be surprised if you get into say, University of Washington, University of Michigan, or Duke (assuming that you perform well in your Masters program and take the classes above).

Yes, you could do that. I think you might have the most luck at schools ranked a bit lower, e.g. those ranked 20-50 in USNWR. However, there is quite a bit of noise in PhD admissions, so you never know. Despite the noisiness in admissions, I *would* say that the very 'elite' programs will most likely favor international applicants who have degrees in Mathematics or Statistics from elite schools like Peking, Tsinghua, USTC, SNU, ISI, etc. and who are usually at the very top of their class. Usually the successful applicants will also have relevant research experience (I saw that Stanford admitted a student one year who had a publication accepted in Annals of Statistics in their first year -- goodness!). So I am not sure that you would be able to compete against such applicants. Applying to a few Masters programs is a good idea too -- you could even do a Masters in Statistics/Mathematics at a reputable school in your home country, and that might help too.

It does seem as though the very top-tier schools (Stanford, Harvard, UChicago, Berkeley) do accept mainly international students who majored in mathematics or statistics, but I have occasionally seen some economics majors do okay in Statistics PhD admissions in the tier of schools below that (e.g. I know of some people who went to Duke for their PhD who have undergrad degrees in Economics). If you major is not Math or Statistics, then I think it is very important to demonstrate strong evidence of mathematical maturity in your application -- e.g. two semesters of real analysis and other evidence of ability to do advanced proof-based math would help. Is there any way that you could pick up a double major in Mathematics? That would probably be your best bet if you're aiming for "top" schools. Internships don't really matter much for Stats PhD programs.

No, it shouldn't put you at any disadvantage for industry. In fact, PhD grads from programs ranked 60-80 also can (and do) get jobs at Facebook, Google, and other "prestigious" companies. Industry as a whole is less "prestige"-driven than academia. Relevant work experience (e.g. an internship) and hacking skills are also more essential to getting most of these jobs than publications. I personally know Statistics PhD graduates who got jobs at Google and the like who had no publications. FWIW, schools ranked 30-40 (and lower) are also good enough to land jobs at R1 universities and prestigious liberal arts colleges. You might not be able to get a job at Stanford, UC Berkeley, or an Ivy, but if you have a strong research record, you can still land a job at a solid research university. For example, last year, UNC-Chapel Hill STOR (a very solid program) hired somebody who has a PhD in Statistics from Florida State University. And North Carolina State University Statistics also hired somebody whose PhD is from University of Illinois-Urbana Champaign.

Your list looks very reasonable to apply to with your profile. I think BU and Pitt are relatively safe choices for your profile, and it wouldn't hurt to add another school like MD Anderson or University of Minnesota Biostatistics (i.e. in roughly the same tier as Columbia and UPenn Perelman).

Even if you were to do well in your Masters program, I would consider UC Berkeley, Columbia, and Yale to be completely unrealistic. You have too many B's, and the competition for these schools is very stiff. Some of these schools only accept very few domestic applicants to begin with, and I'm afraid you won't be able to compete against applicants from Ivy schools, Stanford, UChicago, MIT, etc. with higher GPAs and possibly some solid research experience. UC-Davis, UCLA, UC Irvine, and UT-Austin are reaches as well, IMO. It seems as though the UC schools are all very competitive, regardless of their rank (except for maybe UC-Riverside), because of their desirable locations. But I'm not sure how open UC-Davis would be to accepting their own Masters students as long as you perform well in their program, though -- that might be something to look into. I would say that in order to be competitive for PhD programs, you have to get all A's in your Masters program, especially since you got a B in a graduate Statistics course. Definitely also take a full year of analysis and get A's to make up for your B in undergrad and possibly one or two other advanced math class (e.g. proof-based linear algebra) to show that you can succeed in math-heavy courses. The second year of a Statistics PhD program is pretty theoretical for the most part. If you do well in your Masters, you might be able to get into a program like TAMU or Iowa State. However, the most realistic schools would probably be those in the range of 37-80 of the USNWR rankings (i.e. those ranked below Yale). For your profile, I would consider TAMU and ISU to be the upper end of the schools you should be applying to for Statistics PhDs.

Yes, Ohio State had a very strong spatial statistics group at one point including Noel Cressie and Tao Shi, who had great PhD placements (e.g. Jonathan Bradley at FSU was supervised by these guys). However, it seems as though the group has reduced in size. Both Cressie and Shi have left OSU. I would keep an open mind about research areas -- you may ultimately change your mind about what you're interested in, and nowadays, it is common for a lot of people who go into academia to change their research focus during/after their postdoc. This was probably less common many years ago (so you could spend your whole career studying the same topics), but nowadays, it's common to switch research focus. If you are considering academia, you may wish to consider academic placements of these departments, *in addition* to who the PhD advisors were for the graduates who went into academia. It seems as though OSU placed people in the past who worked on spatial statistics, but as you know, some of their more prominent spatial faculty have left. So it's good to look at historical data, in addition to whether or not those faculty are still at the school or not.

Gradient boosting is essentially an additive model tailored to decision trees, and the concept of additive models was first developed by Friedman and Stuetzle at Stanford. It is possible that somebody else suggested the idea of boosting for tree-based models, but the gradient boosting machine (GBM) paper that gets cited the most often was written by Friedman when he was at Stanford. The generalized additive model (GAM), an important development in nonparametric regression, was also developed by two Stanford statistics faculty, Hastie and Tibshirani (albeit this was before they joined the Stanford faculty). Nobody is claiming that there are *no* people who have revolutionized the field at other programs. Obviously, at UC Berkeley, you have Michael Jordan and Martin Wainwright, and Lucien Le Cam also spent his career at UCB. But there is certainly a higher concentration of such "field-changing" folks at Stanford than at any other school. Bootstrapping, compressed sensing, lasso/L1 regularization methods, additive models -- all very "revolutionary" developments in Stats -- came from people who are at Stanford.

Yeah, Stanford truly is in a class of its own. No other program has as many faculty who revolutionized the field -- e.g. Tibshirani, Hastie, Johnstone, Candes, Efron, Friedman, etc. Lasso, bootstrapping methods, compressed sensing, additive models (including gradient boosting), etc. were basically invented at Stanford. I would note also that Stanford is also very strong in applied statistics. Efron is the founding editor of Annals of Applied Statistics, and Stanford also has people like Susan Holmes, Trevor Hastie, and Robert Tibshirani. Some of these folks have also worked on statistical theory too, but they also work a lot on computational statistics and on applications in biostatistics, bioinformatics, and health policy.

Yeah, in terms of academic placements, Yale has done exceptionally well. This leads me to believe that Yale S&DS must be viewed favorably among many departments. Of course, it's really the responsibility of the applicants themselves to make themselves 'stand out' (you can't just rely on the brand name of your alma mater). But the fact that Yale has produced so many of these outstanding job market candidates who got tenure-track jobs at University of Chicago Statistics, Columbia Statistics, UPenn Wharton Statistics, and Princeton OFRE speaks to the department's strengths. I am not sure about Yale's industry placements, but I can't imagine it being *much* different than the industry placements for UChicago.

I am not inclined to answer this on a public forum. But to make my own personal assessment, I look at student outcomes (i.e. where the PhD graduates have placed), the caliber of the faculty (i.e. current big names and "rising stars"), and the research output (i.e. how actively are the faculty publishing? And which venues are they publishing in?). Note that to assess how "strong" the research output is, it may be a bit different for Statistics vs. Biostatistics. For example, Biostat departments might possibly put heavier weight on publications in journals like Biostatistics, Biometrics Practice, and JRSS: Series A, as well as publications in top specialized journals (like BMC Bioinformatics, Nature Genetics, and American Journal of Human Genetics if the faculty member works in statistical genetics). These venues enjoy wide prestige in the Biostatistics community, but you might not see as many Statistics faculty members publishing in them. Based on the criteria I mentioned, I think that Yale S&DS is quite under-ranked, given their PhD placements, caliber of faculty, and research output.

Congrats on your outstanding acceptances! IMO, Yale is a very good program with some excellent professor (e.g. John Lafferty and Harrison Zhou) who are very renowned and well-known in the field of statistics. Yale's academic placements are also fairly impressive -- they've placed PhD alumni at UC Berkeley, University of Chicago, UPenn Wharton, and Princeton ORFE to name a few. This is an excellent placement record, which leads me to believe that many top schools view the Yale S&DS program favorably. If you are leaning towards industry, then I think that either Chicago or Yale would be fine. It seems like you have especially good vibes from Yale though, so I would definitely not think it's insane to pick Yale over a "higher ranked" school. I am not really sure how their USNWR rank was determined but I truly believe that Yale is under-ranked relative to the department's strength (in terms of its faculty and student outcomes). Rankings are based on popular perceptions (surveys sent out to academic statisticians), which might be slow to change. I think the rankings are still fairly good for the most part, but there are a few that I think are over-ranked or under-ranked relative to their current strength -- Yale is one of those (I believe it's under-ranked).

Well, if your chances of continuing on in a PhD program at UPenn are well above average, then I would strongly consider UPenn AMCS. I know that several of the professors in the Wharton Statistics Department supervise PhD students in AMCS. I also know that some folks in the DBEI Department supervise students in other departments as well (e.g. some CS students who work on NLP are co-advised by faculty in DBEI). Some DBEI PhD students have also been advised by Wharton Statistics faculty (e.g. Edward Kennedy who is now a rising star professor at CMU), so it seems like they're pretty flexible about whom you can work with. Is the UPenn Masters funded (either fully or partially through TA or RA)?? That's also something to consider.

Before starting the PhD, I did review Calculus, probability and statistics (at the Casella/Berger level), real analysis, and linear algebra (including proofs). I think that this was helpful for the classes that I had to take in my first year, since it was fresh in my mind right before starting classes. I didn't study any measure theory, stochastic processes, functional analysis, etc. before starting the PhD. I don't think this is really necessary, but if you are very interested in it, it could potentially be useful... though I should note that by the time you start research, chances are you will forget most of this stuff. At that point, you can just relearn what you need for your research and fill in any gaps as you go. But I do recommend reviewing some of the topics in my first paragraph because you'll be able to use that stuff right away when you start taking classes (rather than possibly needing it one or two years later when you start your dissertation research).

So I do mostly research in the area of Bayesian statistics (though not exclusively), and I have done both applied and theoretical research in this area. I would say for theory: it is pretty important to know analysis and linear algebra well and to be comfortable with probability theory and stochastic processes. Unless you are doing very hardcore theoretical research (and there are some people who do that), you don't need to know measure theory that well, but you should be comfortable with it. Plus, measure theory/probability theory can be pretty useful for Bayesian nonparametrics. In Bayesian nonparametrics, you frequently replace finite-dimensional prior distributions with stochastic processes (e.g. Dirichlet process, Gaussian process, etc.), and it can be useful to know a little bit of measure theory and probability theory. For methodology/applications: obviously be familiar with the Bayesian paradigm, as well as MCMC (Gibbs sampling and Metropolis-Hastings) and maybe variational inference. Most of the time, the posteriors are intractable, so you do need to do approximate inference. It would be useful to be familiar with some of the "classical" models for linear regression, classification, and semiparametric/nonparametric methods for regression/classification. I think once you specialize in a particular research area (e.g. spatial statistics, functional analysis, topological data analysis, etc.), you can learn that stuff on your own. There's no need to study it prior to starting your research, unless you are very interested in it. For programming: Be proficient with programming in R and comfortable with using C/C++. Since R can be a bit slow and have a lot of overhead (compared to C/C++), you may prefer to code in C/C++ and integrate this code with R. R is great for creating plots and visualizations, etc., but if you are going to run MCMC (for example), you may prefer to use C/C++ and then integrate this with R, because your code will run a lot faster.

Johns Hopkins AM&S is a pretty good choice, I think. The school is certainly prestigious, and the department doesn't have bad academic placements. I saw that they recently placed PhD graduates in TT positions at University of Pittsburgh and Indiana University Statistics Departments. UPenn is obviously very prestigious too, but is the AMCS Masters funded? And is there a clear "Masters-to-PhD" route? i.e. do (m)any of the Masters students get to continue on to the PhD program at Penn as well? I would inquire about this possibility (but phrase it like, "If I perform well in the program and if I were to do research with a professor during my Masters, would it be possible to transfer to the PhD program? How many AMCS MS students can continue on to the PhD program at UPenn?") UPenn has some great people for ML/Statistics, including those working in current "hot" fields like differential privacy and deep learning.

Agreed, I would weigh a fellowship offer very heavily. I was on fellowship where I only had to teach/TA for one year. The other three years were totally free to spend on courses and research. This ultimately shaved an entire year off of the (typical) PhD completion time -- finished in 4 years instead of 5, because I only had to focus on my research after the second year, and I had no other responsibilities. It should be noted that if you are interested in academia, then finishing "faster" is *NOT* always advised. If you're waiting for an Annals of Statistics or JASA paper to be accepted, then it's better to wait another year so you can have that on your CV. This way, you will be more competitive on the job market (and you can also get other papers done/accepted/published too in that time). However, being able to finish faster is great if you want to go into industry or if you have a good postdoc lined up that can make your profile more competitive than it would be if you stayed another year in the PhD program.

I've seen it happen (PhD->industry->academia), but usually, the scenario was like this. The person got their PhD and decided to go work in industry. They didn't like it and discovered that they preferred academia, so they went to do a postdoc instead after 1 year out in industry. Then after the postdoc, they got a faculty position. It's much less common for people to return to academia after a number of years. I've seen it happen but this almost always entailed taking a (non-tenure track) lecturer position somewhere, or sometimes a job at a teaching-oriented college. The reason it's harder to return after many years (for research universities anyway) is because academic hiring is based so heavily on your research output (i.e. your publication record) and the department's needs. If you've been working in industry for longer than a year, it is very difficult to keep writing papers and publishing (whereas if you've been out of your PhD for only a year, you could still have fairly recent papers from your dissertation that have been published or that you're working to get published somewhere). It's also hard to keep up with current trends in academic research if you've been out for a very long time, and those tend to change rapidly these days. If your research area isn't at least a moderately sized area of interest to the statistics/biostatistics/machine learning research communities, then it will be hard to get hired. However, I suppose you could theoretically be competitive for academic positions after years spent in industry if you were in a role where you could continue publishing. Or you could also be competitive if you were to go do a postdoc after years spent in industry and used the postdoc as a springboard for getting back in into academia. I think I may have seen a couple of people do this (do a postdoc after years spent in industry). It should be noted that these people usually did Biostatistics postdocs and then went into Biostat departments.

Wisconsin and Penn State both seem to have a good mix of theory and applied people, though it seems PSU is leaning a bit more towards the applied side (e.g. lots of faculty working on statistical genetics, spatial statistics, climate modeling, etc.). UW-Madison also has some good theory people, but good applied people as well. I have a collaborator who is an applied statistician and he just got hired as TT faculty at UW-Madison, and it also seems like UW-Madison has a bunch of applied people working in areas like astrostatistics, statistical genetics, etc. I believe that the Statistics Department at UW-Madison allows Statistics PhD students to do an emphasis in Biostatistics: https://stat.wisc.edu/graduate-studies/. Thus, there seems to be much less separation between the Statistics and the Biostatistics and Medical Informatics departments at UW-Madison (whereas at some other schools, there is hardly any interaction between the Biostat and Stat departments, because the Stat department is super theoretical).

There are a few Statistics departments that are in business schools. For example, at University of Pennsylvania, the Statistics department is in the Wharton School of Business. In spite of this, it does not seem as though most of the statistics faculty at UPenn work on any particular business/economic applications, and most of them are pretty theoretical. I think at Temple University, the Statistics department is also in the business school, and not all of their faculty work on business applications of statistics. As to how reputable the NYU program is, I really do not know much about this program. It seems very, very small (< 10 students across all years). I guess a good starting point would be to look at the publication records of Statistics faculty in the IMOS department at NYU and see where they are publishing. If you're interested in industry, it might not be so important that faculty are publishing consistently in top-tier journals -- but you do want to work with a faculty member who is reasonably productive so you can graduate!

You should just make your decision on or close to April 15 if you're waiting to hear back from other programs that you are seriously considering. I don't think a program would appreciate it if you accepted their offer only to renege on it later.

It seems like the vast majority of Rutgers PhD graduates go into industry. This could be self-selection though, where most of them have a preference towards industry to begin with (maybe because of the proximity to NYC? Not really sure.) However, Rutgers does have some very famous faculty like Cun-Hui Zhang who has done a lot of *excellent* work on high-dimensional nonconvex problems and nonparametric/semiparametric statistics. And their recently hired Assistant Professors are all very impressive (publishing a lot in Annals, JRSS-B, JASA, NeurIPS, ICML, etc.). University of Florida does decently well for academic placements. There are a few PhD grads from this program who have become TT faculty at statistics/math departments (Duke, University of Minnesota, UT Austin, University of Cincinnati, etc.) as well as Biostatistics departments (University of Wisconsin, University of Buffalo, University of Louisville). Back when there were really big hotshots in the department (Agresti, George Casella), UFL was able to place some of its graduates at places like Harvard Biostatistics and Johns Hopkins Biostatistics. That doesn't seem to be much the case anymore, but the placements are still good. Coursework at UFL can be a bit on the theoretical side, but there are a decent number of applied statistics faculty there now, so you don't have to do theoretical statistics for your dissertation research if you don't want to.