bayessays

They are not technical interviews - generally, they'll ask you to talk about why you want a PhD in statistics, ask you about your research projects, etc. - you should be able to talk about these things. You might want to have some questions prepared for them about their program. Definitely not going to be quizzed on material in your classes, and if you are, I'd take that as a massive red flag of a dysfunctional department.

I would say 167+, do not retake. A 170 looks nice, but I don't think people are going to draw much distinction above 167. 165 is borderline, and 164 and below is where I think people should definitely consider retaking for top programs.

Hard to say since you didn't list your specific grades in undergraduate math courses. If you got Bs in your undergraduate math classes, no, I don't think your graduate GPA makes up for it because usually grad school grades are inflated so 3.88 is average.

I think most of your reach list is not worth applying to - I don't think you have any chance of being admitted to Columbia or Cornell stats PhDs. The realistic/easy choices are things I would consider "matches" for your profile but I don't think places like ND/BU/UMass/UVA are safe guaranteed admissions. Your results will probably be all over the place considering your undergraduate GPA.

I'd say you have a decent shot, yes. They're all prestigious schools, so I don't think any of them is a guarantee, but I think the types of schools you're applying to are appropriate for your profile and I don't think you should go much lower for biostat. I feel like switching Minnesota from stat to biostat would make your list feel much safer to me. (But they're incredibly different programs, so this is up to you and your preferences for research) As for being an east Asian male, I can only say that each person on an admissions committee is just a person making a somewhat arbitrary decision from a pile of great candidates, and their own personal views may consciously or unconsciously affect their choices. Overall, I'd say you are basically in the same position as domestic white male applicants.

Honestly I think that's a great list for you - I think about half of those schools are strong matches and then a lot of (realistic) reaches that are reaches for almost everyone. As for digging into the stat vs biostat choice, I sort of wonder about your choices for Duke, UNC, and Yale. For Duke, their biostat department is relatively new (I think this is one of the safer options on your list along with Emory and Iowa State) whereas their statistics department is elite. But for Yale, you made the opposite choice of Duke - choosing the elite stats department over the relatively middle-of-the-road biostat program. For UNC, I don't think you can go wrong with either department, but UNC's stats department is sort of unique in having a big probability focus. This might be a pro or con depending on your interests. I don't think you can go wrong either way, really.

I think you'll be in good shape. Focus on getting strong letters from your research advisors. 5-15 on US news is doable, but I'd apply to plenty of programs in the 15-30 range as well. Your questions: 1. First, I don't really think they are essentially equivalent for everyone - they are for most students though. If you want to study probability theory or extremely theoretical statistics, you will want to go to a statistics department. If you only want to do applied research (not be a professor) and don't really like hard math, you will want to go to a biostatistics department. If you're in between, it probably doesn't matter much. Your profile will be very good for either type of program, but I certainly think you have a much better shot at Harvard/Washington/Michigan if you apply to their biostat programs rather than their stat programs. 2. Most statistics students do not do research their first year, so I wouldn't bother. This may vary for certain programs, but most students will probably not be doing this and it is not expected. 3. I would not worry about this at all. Differential equations is irrelevant, and one semester of linear algebra is fine. Your math background is sufficient. The only thing that could improve it really is to take lots of difficult proof-based classes/graduate classes to boost your profile for top 10 stat programs, and this is definitely not worth the effort.

Start with US News, I'd focus mostly on programs in the 10-60 range. See what locations you like and look at those department pages - most programs have a faculty page that will list research interests. Most programs will have one or two faculty that focus on Bayesian stuff - programs that focus on Bayesian stats include Duke, Texas, Missouri, Iowa. You'll have to dig a little harder depending on what exactly you mean by "machine learning." Also look at biostatistics programs in the top 20, as your profile will play well there. I'd talk to your professors, and focus on securing good letters of recommendation from research advisors - you go to a top 15 stats program, so try to take advantage of the vast connections of your department if possible.

If anything, I might just switch a few schools from their stat to biostat departments - ie, Michigan/Minnesota/Columbia/Duke are reaches for stats, but their biostat programs are easier to get into. It really depends on if you're more interested in doing math or doing applied work though - I don't think the same person would be happy at both Columbia stat and Columbia biostat.

I think your classification of schools as reaches/matches is basically correct. Your profile is basically ideal, except for the lack of real analysis being a detriment at top stats departments. I think you'll have a lot of success.

From their website, it seems like Math PhD students on the stats track would only have to take exams in Mathematical Statistics and Real Analysis, but do you really want to be taking a real analysis qualifying exam at a top 20 math department? Do you want to be wasting your time studying for PhD Math classes? Personally, given my preferences, it would be a no-brainer to choose the data science program because I don't want to do that much math. If you don't want to be a professor, I don't see many advantages unless you feel like you would get a lot of personal enjoyment and satisfaction from completing intense math classes. I'd rather spend that time taking stats/data science electives and getting research experience, so I'd go for the data science PhD. (There may be some areas/professions where the Math PhD will help you a lot in getting a position - for instance, maybe if you wanted to go into some type of quantitative finance position. But for biostatistics or technology jobs, I don't think the difference between these programs would matter at all, and I think you could get a lot more research done/learn more practical stuff without the distraction of hard math classes that have nothing to do with statistics)

Since cost is not a factor, I think location is really the biggest consideration. Maybe people who grew up in California have more associations with these schools, but you're not choosing between schools that are at vastly different levels of prestige like Stanford vs CSU East Bay. I don't think the availability of slightly more electives is a reason to uproot your life if you don't really want to. CSU East Bay would be my choice for this reason.

Where's the list?

Probably top 5, almost certainly top 10-12. If you're working with a prof on a deep learning project at one of those schools and can get a good letter from him, you're pretty set with your grades, extensive math coursework, and your school pedigree. Statistics is not like CS where you need publications before applying.

"Advanced calculus" and "real analysis" are the same thing to most people - if you have done epsilon-delta proofs extensively, that's basically what they want. I don't think it's worth applying to the PhD programs at Stanford, Berkeley, Michigan, and Washington (these would be huge reaches with your profile). I think the other schools are worth a shot but not guaranteed. I think a master's could improve your profile for sure, but I think it's definitely reasonable to apply to some PhD programs to see if you can get in and take the more direct path - a lot of schools will offer you master's admission if you don't get into the PhD program.

I think most reasonable people would not care much about this, but every person on an admissions committee is just a person with their own opinions, and I have experienced professors who think non-research industry jobs are beneath their PhD students. You probably don't really want to attend a place where you have to lie about your goals for years anyways, and in a discipline like statistics, you'll definitely find departments and advisors that are supportive of this.

It probably doesn't matter too much. The bigger problem people have is being too tied down to a specific area/professor and not being flexible. If they require you to put some names in a certain box on the application, I'd just choose some people who look interesting and are still actively doing research that you'll be qualified to do (ie, don't put a probabilist if you haven't taken real analysis, etc)

Check the programs individually, but basically any program will let you master out, and many people transfer out after the master's and go to a new PhD program (this may be a headache and some people feel guilty about it, but it happens all the time).

Your profile is obviously great. The only small criticism might be that your math background could always be deeper when you apply to a place like Stanford, but it's more than sufficient. What's the reason behind only applying to UCLA/UCDavis's MS programs? UCLA and UCD are great PhD programs (much better than UCSB). I'd apply to Stanford, UW, Berkeley, UCLA, UCD, UCSD, UCI, UCSC and UCSB's PhD programs and I think you'd probably have some choices.

It's #13 on the US News rankings and based on the schools ranked similarly, I think that's pretty appropriate. Not really comparable to the biostatistics program - the statistics program has top people in high dimensional stats, MCMC - much more mathematical. Have seen UMN Stats alumni land very good positions. I interviewed there 10 years ago and had a wonderful impression of the program, but the stipend back then was not enough to live on - hopefully they've fixed that.

I just looked up UConn on gradcafe's results page and it still seems like they give out unfunded PhD offers to international students (FSU and Georgia are also famous for doing this). It might not affect you though as a domestic student, but it just gives bad vibes.

I am probably being too harsh on UConn and Temple - I looked into Temple in previous application cycles and their faculty is pretty impressive, and UConn is strong too and has a long history (although in a similar way to Iowa State, they seem to have been in decline for some time). I seem to remember 10 years ago not applying to UConn because their department had funding issues, but I could simply be mistaken, and of course things could have changed in the meantime. (These programs also seem to have risen in the rankings in recent years) Sometimes it's not even really about the quality of the professors, but just the presence of their graduates - I don't see UConn and Temple graduates with the same frequency I see graduates from a school like UIUC, even if there maybe isn't a huge difference in the faculty/research quality. Part of this is reflective of the student population, and not the faculty - some programs in less desirable locations have overwhelmingly international student populations (Mizzou and Iowa are extreme examples that come to my mind, but I suspect UConn could be similar) and thus their graduates are not populating as many jobs in the US relative to the quality of the faculty. This could affect your social life/cohort experience and really depends on your learning style and personal preferences - I couldn't imagine my master's program without having had the large cohort of domestic students to work together with, but when I went back to school later and didn't care about that aspect, it wasn't a big deal. I think you can do good research pretty much anywhere. I went to a program much less known than uconn/temple, but they had strong faculty. I never mentioned bad grades in my SOP, and I've never heard anyone recommend it unless it was both recent and for a really good reason (personal illness/death in family/etc). I think freshman year is self explanatory (I got a B in calculus and B- in linear algebra and it seemed to barely affect my applications despite all my anxiety about it at the time) Also, you can't discount the wildcard of the good recommendation letters from a professor at a good program - if they have connections on admissions committees and people respect their opinions, that can change anything.

Your profile is very similar to mine (liberal arts college, bad math grades early on, high GRE). Frankly, I think your list is very bizarre, that there is a HUGE gap between your "mid" and "safeties." I don't think you have any realistic chance at the hail marys. I'd classify UW as a hail mary too and would switch UW/Penn to their biostat programs for a more realistic shot. I think UC Davis, Purdue, and Iowa State is a good range to aspire to but I definitely wouldn't consider those safe by any means (except perhaps for your Iowa State connection). I never had any luck applying to Purdue, so they seemed to judge my early math grades pretty harshly. As for UConn and Temple, I definitely think you can do better than those, but they wouldn't be bad safeties if the location works I know you say you are restricted in your location, so here are some schools near (within an hour or two) of some of the ones you mentioned that might be good matches: Near Stanford/Berkley: UCSC (almost 2 hours away) Near Carnegie Mellon: Pitt Stat or Pitt Biostat Near Purdue: Notre Dame ACMS (New up and coming program with really good faculty)

If the Bs are in core math classes (calc, linear algebra, or real analysis), I would definitely lower my expectations. I was thinking the low GPA was solely from some irrelevant grad CS class (something like operating systems or compilers). Given the new info, I'd consider the list to be too reach-heavy for comfort, and I think the top row is very unlikely.

Have you talked to any of your stats professors at UIUC? I think that would be a good target given your profile (though of course I'd understand if you want to go somewhere new). I'm going to assume you basically have straight A/A-s in your math/stats courses I think Penn and Columbia are big reaches, but if application fees aren't an issue I don't think your list is crazy - I think the first row of schools (up to UW Madison) are probably "reaches" and the bottom rows are "matches" but you don't have any really safe options, if there even exists such a thing in PhD admissions. I don't think you need to settle though, so if I were you I'd probably just apply to more schools in the same league as the bottom row. For the top schools, I think the biggest issue is that you don't really have any statistics experience that stands out and people might wonder why you didn't get involved in research during your master's. Combined with going to a good but not elite school, there's not a lot that makes you stand out.