wine in coffee cups

Did you advisors elaborate on why Cornell might be a better choice overall than CMU?

I think the placements coming out of Michigan and CMU look equivalent, though I only Googled the names of students from the past couple years of Michigan grads. I see lots of data scientists in the past couple of years from Michigan. CMU has some recent placements suggestive of a two body solution. I don't have the impression that one school is likely to set you up for better opportunities than the other. Postdocs seem common in statistics now even for people with strong CVs coming out of a PhD. I wouldn't view those as questionable placements unless they were with unknown groups. A friend from my department last year had many interviews and received several tenure-track offers. She accepted one of those offers but deferred its start to do a postdoc at a well-known research institute with a focus on her specialization. If you're going to become faculty anywhere besides a very large department, you might not have many opportunities to collaborate locally with people who work in your area of research because you might well be the only person or one of a few people. You can use a postdoc to work with the giants in your area and build up a network of potential collaborators from other institutions before becoming more isolated in a faculty job. Postdocs are also a final opportunity to live somewhere interesting for a year or two before being more at the mercy of the academic job market in terms of location. If you are being rather picky about where you apply for faculty jobs, strengthening your research with a postdoc while waiting a year or two for the right openings makes some sense. As for choosing: first, remember that both are good options and there are no real mistakes to be made here. At this level of program, placement is not really a deciding factor. I think if you are very interested in machine learning AND you can work on ML research without being in the joint program (I don't know if this is the case or not), the CMU makes more sense. If you are generally undecided, I think Michigan might be better as it is an overall more well-rounded department than CMU, quite large, and has a great biostatistics department too. I really hope you visit both schools and talk to as many students and faculty as you can, though. You want a place where you can fit in happily for 5 years, and that is basically impossible to judge without being on site. Some departments are more tightly knit than others, and it's nice to be at a supportive department with students/faculty who generally like being around each other. I would also think about whether you'd rather live in Pittsburgh or Ann Arbor, obviously.

I recommend taking a break from school. It sounds like you might be a little burnt out and insecure. Grad school will not improve the feeling. The problem is not that we compete with each other, but there's a chronic sense of guilt and disappointment with yourself for not working harder, lots of angst about missed opportunities and your future in general. It's nice to put that aside for a few years, work on something else altogether, do a good enough job and feel satisfied that you don't need to be doing more. You can pursue new hobbies, meet other kinds of people, figure out what matters to you, generally become slightly more of a developed adult human being. And the feeling of guilt-free evenings and weekends is something to appreciate while you can! You're well-situated to do something interesting after college with your math and programming skills. I'd guess that a majority of Americans take at least one year before enrolling in a stats PhD program post-undergrad, not at all unusual for people to delay grad school by a few years in this field. You should do some online CV/LinkedIn stalking of current statistics PhD students to get an idea for what they did. Some examples from people I've encountered in my program and others, all of whom were math, stat, CS, or econ majors: teaching math (e.g. through Teach for America), teaching English abroad, many many 1-2 year post-bac research assistant jobs helping with the kinds of data collection and analysis professors don't have grad RAs working on (e.g. econ professors, law professors, neuroscience labs, biostats stuff for hospital research units), many consultant flavors (management consulting, litigation consulting, IT consulting, healthcare consulting, media research), research at the Federal Reserve, journalism, actuarial work, programming at tech companies big and small, freelance web development, Peace Corps, fancy foreign fellowships.

Budapest is great, enthusiastically recommend doing the BSM program if you can. I went, as did several other people in my statistics PhD program. I'm not sure if BSM helps directly with admissions (it definitely doesn't hurt), but it's a good opportunity to take classes your home university doesn't offer, particularly in combinatorics or graduate level analysis. Budapest is a really fun city and inexpensive too. You also say you're worried about being younger than most graduates (did you skip some grades or have an accelerated undergrad?), so you might consider waiting to apply until after you graduate. Get your study abroad experience in there, get A's in more advanced classes on your transcript to offset those calculus B's, get a little more research or work experience. Besides, it super sucks to be under 21 and unable to participate in an admitted students event at a bar if you're 20 when you go on visits. Can you elaborate on what you mean by having "done a pretty mediocre job at what I perceive to be a mediocre school"? Your profile doesn't sound mediocre so maybe you're just being excessively pessimistic, but what probably matters most is how your faculty recommenders would describe your performance. A compelling recommendation would compare you favorably with other graduates from Clarkson who have gone on to succeed in math/stat PhD programs and have your professors gushing about your abilities, intellectual curiosity, potential for research, blah blah. From a small department, they should be able to say that you're one of the strongest students they've had in recent years, maybe the strongest in your year. Someone from Duke should be able to compare you favorably to their students, too, which are more of a known entity in the statistics world. If you're in a position now where you think they would be able to say things along those lines, then I wouldn't worry too much about being able to get into a decent stats PhD program.

From what I know about UCSC, they are very Bayesian and have some strong faculty and alumni particularly in Bayesian nonparametrics. That's a hot area right now, so if that's your jam, Santa Cruz would be a no brainer. I don't know what it's like to be a student there but it certainly seems like it's an up-and-coming department. (I don't know anything about Irvine.)

Nobody wants to pay for a degree if they don't have to. But people who are admitted to funded PhD programs have demonstrated that they have the potential to succeed in graduate coursework and research: they have high grades from undergrad, they have taken a lot of preparatory classes and are ready for grad-level courses, they have recommendations from professors who had very positive things to say about their abilities. (hapa, that's something you haven't mentioned yet: who do you plan to get references from and what will they say?) If you are weak on these things, then you shouldn't expect to be able to get into a funded program from the get-go. You have to prove yourself and you don't have the privilege of having that experience on someone else's dime (at least for the first year or two). I think you need to swallow your pride and consider that your most likely path to a statistics PhD will involve an unfunded MS to start with. Finally, regarding this, just want to make sure your expectations of what your work experience will and won't do for you are reasonable: I started my program at 27, before that had a job where I did a ton of programming/database manipulation/basic statistics, lots of thinking carefully about how to get useful results from whatever data we had. I would say the degree to which my work experience has helped me in my program has been lower than I hoped. My research interests came out of the domain of my previous work and got me my RA project (and eventual dissertation hopefully?), and knowing how to deal with annoying coding/data problems has certainly saved me some time solving those issues relative to someone who hasn't experienced them before. However, none of this was preparatory for the coursework (years away from doing long problem sets and taking timed exams have been a detriment, actually) or in the main work of pulling the statistical theory together to publish papers and write my dissertation. It might become more helpful when I do the consulting requirement, but overall, I'd say the kind of work experience we have had provides only a mild benefit in statistics PhD programs.

If you are targeting one specific school, you should probably be in touch with their admissions coordinator (at least after the current application season dies down). That GPA might be below minimum grad school requirements, and you should find out whether this is the case or not and if so, if additional math coursework could mitigate that so that your application could be considered. You could also ask about MS admissions and if one could enroll as an MS student and be admitted to the PhD program later (I think you could have more of a chance with this route rather than direct to PhD since your math/stats background will probably be weak relative to other applicants).

First of all, you should be looking at what kind of research takes place in these departments. Any program that is small and not well-known probably specializes in just a few areas, or even is mostly a service department to the university (teaching stats courses and assisting researchers from other departments, little methodological research). Your fit within the program is very important for making sure you can actually get a dissertation out of it. Look over the faculty who still publish and make sure there are at least a couple you would plausibly be happy with as your advisor. With smaller lesser-known programs, my suspicion is that graduates tend to go into a couple of industries that are concentrated in the region of the school. Completely guessing here, but I would not be surprised if UVA/George Washington/George Mason/Virginia Tech have most of their alumni in government-related positions in the DC area. Pitt and Temple might have a tendency towards pharma. UCSB or UCR might have more of a tech presence. Not sure how geographically dispersed the alumni are from any of these. It would be a good idea to think a little more about what kind of a non-academic job you want and where you might want to live post-PhD when looking at your funded offers, as some of these programs might be better at placing graduates in a broad array of industries and locations than others.

I had the same issue. Made my OCD spreadsheet comparing the three places I had whittled it down to on every dimension I could think of (largely the same as cyprusprior's list), came away understanding the tradeoffs better but not really much closer to a final decision. Asked friends/coworkers/family/recommenders/internet randos here, got thoughtful responses but completely conflicting recommendations. You have to send a lot of emails after you've made a decision: to the departments you haven't turned down yet, to the helpful faculty and students you were in contact with along the way, to your letter writers and everyone else who helped and supported you. A lot of these parties want to know why you made the choice that you did (you will get follow-up from most departments if you don't provide it), so you need to have a short explanation for your decision prepared. It's going to sound stupid, but sometimes you just have to write the emails to draw out of you what you really want. Sit on them for a day or two. You'll find there's probably one version of those emails you would regret sending more than the other. For me, it was observing this dissonance between cheerful email drafts telling everyone where I was going while realizing I couldn't visualize my life/work there that nudged me towards my decision.

If you're in a department that somehow does more SAS than R, you're in a pretty applied program and that will manifest itself in ways other than software orientation. For gauging career opportunities, I would look directly at where the alumni are rather than guess based on software.

I think you might have suggested this generally with "age structure of faculty", but I would add probable retirements and anticipated new positions to a list of things to think about. I really like being associated with a department that has added a lot of new faculty in the past few years but also has an established core of mid- and late-career researchers. The young faculty are great: tremendously energetic and productive, bringing in a lot of their connections as outside speakers working on cutting edge stuff, really helpful for job market advice for senior students because they've just gone through the process themselves. Might be less relevant for biostatistics, but for statistics, I would also add ease of connecting with researchers in other departments if you have particular areas of application in mind. Every place says they collaborate with other departments, but some are better at it than others (and all are much better in some specific fields than others). At my university, you'd be in a great position if you wanted to work on social networks, genetics, ecology, or demography, but in a tough spot if you wanted to get data from researchers in some area we no longer have ties to (like finance).

Is Michigan actually considered weak for clinical trials, or is it just not as distinguished in that area as it is for genetics research? I don't know much about the biostatistics side, but I see quite a few hits looking on the biostat faculty page doing the idiot thing and searching for "clinical". Clinical trials, isn't that the bread and butter of all biostatistics departments? And on the statistics side they have Susan Murphy doing some fascinating work with SMART designs, does she ever work with biostatistics students? What I'm suggesting is that Michigan is an otherwise very good department and I think you're unlikely to improve on it at least in terms of overall reputation if you reapply. (I don't know anything about BU.) You already had solid research experience applying this year, weren't the projects you were working on mentioned in your essays and rec letters? It'd be nice to have actual publications rather than papers in review if you reapply of course, but this doesn't sound like added new information IMO. Do you think your recommendation letters are likely to be much better, or more of the same? I would also weigh this against how rapidly biostatistics admissions have become competitive in the past few years. Turning down places like Michigan or BU to roll the dice again at the places you were rejected by seems questionable to me. You might contact the admissions coordinator at Rice, etc. and see if they might be willing to discuss the weak spots in your application that led to rejection or how they view re-applicants in general. You could get ignored, of course, but better to have this information before you turn down good opportunities than to find out that they don't have much interest in applicants they previously rejected.

Just a couple of additional thoughts: Guessing based on names, could be wrong, but Wisconsin has way more international students from China and South Korea than the other departments, which seem to have more of a mix of Americans and students from other countries in there. Not that students from outside North America can't land solid postdocs or tenure-track jobs (clearly some do), but I think there are marked differences in job and location preferences between international and domestic students. I think Americans tend to have better academic placements for reasons that I can only speculate. Pharma in the US is heavily Chinese and you definitely see that reflected in the Wisconsin placements, maybe particular advisors with strong industry connections feeding a steady pipeline to pharma and some homophily. (Anecdotally, a few years ago, I had to read depositions taken of statisticians working in an American office of a drug company. Even though all of them received their PhDs from mid-ranked US statistics departments and had been at the company for several years, the attorneys still needed to conduct portions of their interviews through a translator and to have emails circulated within the statistics group translated.) Two-body issues can make placements very challenging, and what might look like an odd choice of postdoc or temporary position could reflect compromises made in order to be near an academic spouse or aging parents.

^ I think you guys misinterpreted what s/he did. The 2010 rankings are based only on 2009 survey results, while the 2014 rankings are based on averaging the survey results from 2009 and 2013. The two columns are the 2010 ratings and 2014 ratings, and stats_applicant was trying to use the fact that the 2014 rankings are an average to recover the 2013 survey results without the 2009 data mixed in. Looks reasonable to me.

If you're trying to assess reputation, you look at the main signalers that the faculty US News surveyed would have likely had in mind. I think things like: how many people in the department are well-known figures in the statistics community (highly cited publications, awards, frequently invited to give lectures at conferences or other departments)? Are faculty generally publishing in reputable statistics journals and the big machine learning conferences (as opposed to unknown journals or almost entirely niche venues for their area of application)? Do a fair number of PhD graduates place into nice-sounding academic positions?

Just glancing at a few faculty CVs, it doesn't appear to have existed as its own department until 2010 (was under epidemiology before). This thread from last year might be helpful: It's probably the case that the department is majority international, but it can be really hard to tell going off of names/photos. For example, 4 out of 11 people in my cohort are US/Canadian citizens and native English speakers who happen to be of Chinese or Indian descent. I would only worry about overrepresentation of international students for a couple of reasons: (1) if strong social cliques form on national lines (UF biostat might be too small for this to be an issue) and (2) if international students almost all intend to return to their home countries after graduation (limited usefulness of future alumni network and limited signaling information on reputation of alumni when you go on the market in the USA). These are issues in some departments and not in others, but current students could tell you.

All of the programs you named (besides George Mason) are not standalone statistics programs, but rather concentrations in statistics housed inside math, applied math, or "mathematics and statistics" departments and thus ranked under math.

According to the methodology, they have averaged survey results from fall 2009 and fall 2013 to compute these scores, so these don't completely reflect current perceptions vs. the 2010 version. Eyeballing the 2014 vs. 2010 reputation scores, it looks like most top programs remained the same or were nudged up one decimal place, with the biggest movement I see being Wharton bumped from 3.9 to 4.1. That reflects what I think a few people here have suggested, that Wharton has had some gains in reputation as of late and would be ranked even higher if just using the more recent data. Also: statisticians

There is some talk in UW biostat about potentially changing requirements, so inquire when you visit. (Actually, ask about potential changes everywhere you visit.) I think most UW biostat students end up passing their quals, but it sure is a rough couple of years getting there. Most students are supported on RAships before passing quals, so quals are not a total roadblock to starting on research, but it definitely is very hard to make progress on anything interesting with the heavy homework load and frequent in-class exams in the theory and methods sequences for your first two years.

Really glad someone has done this, thanks persistent_homology. Something that would be a moderate-to-severe PITA but potentially yield interesting results is to try to match up records from the same user (as I suggested ) to examine which universities tend to accept the same sets of applicants.

FYI, this forum imposes short time limits on editing, unlike mathematicsgre.com, which is why profile posts work well there and don't work here. You might use the signature feature to track acceptances/rejections/waitlists, which can be edited at any time and would have the advantage of following you around to other threads. (There are so many people here this year whose usernames are variations on "Stat PhD Applicant" or that start with "c" that this curious onlooker can't keep anyone straight, but a signature might help disambiguate )

Came across this while looking at Twitter, thought these might be helpful to people looking to compare departments. I hadn't seen current data for the number of graduates (I think NRC uses numbers from 2005), so some of these figures from 2010-2012 were pretty surprising to me. Texas A&M is the second largest producer of stats PhDs after NC State, who knew? Columbia graduated 288 stats MA students alone in 2012 -- that is, over 100 more students than the total biostatistics PhDs awarded in 2012 across all US departments! Largest PhD programs in statistics and biostatistics Largest master's programs in statistics and biostatistics If you follow the links above to the AmStat blog posts, there are links to PDFs with a complete listing of annual number of graduates by year from each program.

I think leaving a program knowing how to read an academic paper or text that uses measure theoretic statements of integrals and densities is important. But you don't really need a whole class to learn what is meant when something is written using dP or dmu notation. Would not say I gained anything statistically useful from going through tedious proofs of the Caratheodory extension theorem and all that. That said, I enjoyed the measure theory class taught in my program. Coming from a math background, it was more familiar territory for me than the statistical theory courses, and quite a bit easier. My department has measure theory as a nominal requirement/prerequisite for the PhD theory courses, but I would not say it was necessary beyond passing familiarity with the notation, understanding that you can often ignore issues arising only on sets of measure zero, that expectation is integration with respect to a probability measure (which can be a weird mixture of continuous and discrete densities), and knowing when to invoke the big theorems and inequalities (Fatou's lemma, monotone convergence, dominated convergence, Fubini/Tonelli, Cauchy-Schwarz, Markov's inequality, etc.). I think most departments would do well to condense these ideas into a short crash course rather than use a whole quarter/semester actually proving all the results, and I suspect that ones that don't require measure theory already do exactly that as part of a required probability or theory sequence. As to how this relates to choosing a program, I don't think you need to worry about the requirement/lack thereof affecting your ability to get a job. There are so many more important considerations beyond whether everyone else is required to take a math class you are planning to take anyway that this really should be at the bottom of your priorities. For example, something you maybe should care about is not so much specific requirements like one measure theory course, but how burdensome the collection of them are and the degree to which this interferes with becoming a researcher. My program used to have heavy coursework and qualifying exam requirements, which prevented students from getting involved in research until their third (or even fourth) years. This changed right before I started, so the more recent cohorts have more students involved in research during their first and second years, which hopefully will mean shorter times to degree for many of us and stronger CVs when applying for jobs.

I mostly disagree. Based on their alumni page (http://www.stat.harvard.edu/alumni/PhD.html), I would still say that those coming out of Harvard stat who get academic jobs tend to place pretty well, comparable to the academic positions graduates get coming out of other top stats departments whose names don't rhyme with Cranford or Twerkeley. I am surprised by how many graduates go into non-academic positions, though, particularly in finance. My guess is that students who enrolled hoping to use the Harvard name are more or less getting what they want. Personally, those finance positions sound incredibly unappealing to me, but someone who is interested in Wall Street after a stats PhD (and for whatever reason isn't studying something more directly relevant) would have a good landing after Harvard. There are big drawbacks with Harvard stat that applicants aren't necessarily aware of (I wasn't), but not really (or solely) the things you've identified. It's more that (1) assistant/associate professors rarely get tenure at Harvard, so you can't count on many faculty being there for the duration of your PhD, and some of the senior faculty are approaching retirement, so the composition of the department 5 years from now is a big question mark, (2) there are some, diplomatically speaking, "personality quirks" among some of the senior faculty and strange research rivalries at play that isolate Harvard students from people who do work in the same field elsewhere, and (3) these political issues extend to working with the biostat department, so the physical and political separation of the programs means they are not nearly as integrated there as they are at other universities with great stat and biostat departments (e.g. UW, Michigan, UNC) and students lose out on broader perspectives/resources. I would generally recommend that future applicants to statistics programs who are interested in Boston-area programs make sure the biostat department isn't a better fit before sending out an applicant to Harvard stat. I am happy with how my applications went and where I ended up, but if I could change one thing, I would have applied to Harvard biostat instead of stat.

The current chair of UW stat is a prominent statistical geneticist (Thompson). You'll want to investigate these young faculty at UW stat/biostat for interesting work falling in the intersection of statgen and ML: Minin (stat), Witten (biostat), Shojaie (biostat), Simon (biostat). I think there are many others in the biostat and genome sciences departments working in statgen, but the departments are so huge and statgen not my area that I don't have a sense of all of the research going on. Suffice it to say, it's a lot! Not to mention the recent and ongoing hires in ML across departments and the $37.8 million UW was just awarded by Moore-Sloan for data science/machine learning initiatives. I personally recommend waiting to hear if you get in off the waitlist at UW: in each of the past two years, I think ~2 people have enrolled off of the waitlist (don't know how many were offered). You can always send a deposit to CMU and then forfeit it if you get into UW and decide to enroll.