Top 3 Biostatistics vs top 10 Statistics Ph.D.

[gc2012]

I think it's valuable to think about your long term goals here. Keep in mind, your institution and degree program (statistics/biostatistics) will be at the top of your CV for the rest of your life. It will be the first thing every hiring committee sees. That aside, I don't think it's unreasonable to assume that going to a biostatistics program puts you on a track to become faculty in a biostatistics department, and a statistics program points you to positions in statistics departments. Obviously that is not true in every case, but it's a reasonable rule of thumb. I do agree that its easier to go from statistics to biostatistics than the other way around. 

The other key thing to consider, at least about biostatistics, is that it is very applied in practice. To be sure, there are very high profile academic biostatisticians doing fairly theoretical work, but the vast majority of practicing biostatisticians spend a large amount of time providing statistical support to biomedical researchers. Moreover, methods development in biostatistics will require learning a lot about the underlying science. For instance, if you work in statistical genetics, you need to know a lot about genetics. I don't know enough about regular statistics professors to comment on whether they have similar experiences, but my suspicion is there is less of an applied aspect. 

As to the salary issue, let's be precise. Amstat news regularly produces a salary survey and biostat professors report making about $20,000 more than stat professors of equal experience. It is true that the salaries for statistics faculty are for 9 month contracts, compared to 12 months for biostatistics. To be clear, though, a 9 month salary means that stat professors are only contractually obligated to work for 9 months,  but they also aren't paid at all for the remaining 3 months of the year. So at the end of the day, that 9 month salary is what they get in institutional support over 12 months. So the distinction here is that biostatistics faculty are required to work more, but they do in fact get paid more on an annual basis. 

With regard to biostatisticians having to get most of their salary support from grants, that is true, but with some important caveats. Namely, it's easier for biostatisticians to get grants than most other scientists. Basically, the way it works for most biostat faculty is they spend a fairly large amount of time writing the statistics sections of grant applications for biomedical researchers. They get in on enough of these grants that when even a few of them go through, they have salary support. So if working on grant applications does not appeal to you, think very hard about going into biostatistics at an academic institution. However, for those willing, obtaining salary support via grants is not so scary. 

[Stat Assistant Professor]

Excellent points by gc2012. I would make the decision based more on research fit. Both Statistics and Biostatistics programs at top schools will be mathematically rigorous and provide you with training in both statistics theory and applied statistics (computing, regression, etc.), so the training will be excellent regardless. And both are likely to lead to careers where the compensation will enable you to live comfortably, so I wouldn't make the decision based on money. Above all, it is most important that you are personally fulfilled by what you're doing/researching. 

 You would be able to develop new methodology for things like high-dimensional statistics, causal inference, etc. in either a Statistics and Biostatistics departments. However, a Statistics dept is likely to be slightly more theoretical (as in,  the mathematical properties and foundations of your methods -- e.g. convergence rates, tail/asymptotic behavior,  properties under different loss functions, etc. -- are of greater interest than their specific application... whereas in biostats, the methodological contributions and their application to stuff like genetics or electronic health records tend to be the emphasis). A dissertation coming from a Stat department is more likely to be primarily a theoretical contribution rather than an applied/methodological one. So really, it depends on what type of research you would be most fulfilled doing . It's a matter of personal taste and preference. 

Edited February 9, 2018 by Applied Math to Stat

[bayessays]

I think research fit is part of the equation and there is obviously a ton of overlap, but the environment in the two types of departments is incredibly different. Someone who wants to do math all day and doesn't care about biomedical or public health research is going to be very unhappy in a biostatistics department in my opinion.  There is a huge difference between writing a statistics paper that has some applications in genetics, and being part of the collaborative projects that will be a necessary part of being a biostatistician. 

[StatsG0d]

14 hours ago, Applied Math to Stat said:

You would be able to develop new methodology for things like high-dimensional statistics, causal inference, etc. in either a Statistics and Biostatistics departments. However, a Statistics dept is likely to be slightly more theoretical (as in,  the mathematical properties and foundations of your methods -- e.g. convergence rates, tail/asymptotic behavior,  properties under different loss functions, etc. -- are of greater interest than their specific application... whereas in biostats, the methodological contributions and their application to stuff like genetics or electronic health records tend to be the emphasis). A dissertation coming from a Stat department is more likely to be primarily a theoretical contribution rather than an applied/methodological one. So really, it depends on what type of research you would be most fulfilled doing . It's a matter of personal taste and preference. 

This seems to be the thinking from many people, and while it may be true in some programs, it's not true in the others. For example, I attend a top-5 biostats program (not UW), and our coursework covers measure theory, limit theory, decision theory, etc. just like a traditional statistics program would.

I don't think you can put all biostatistics programs in a vacuum and simply say "well it's biostatistics so it's less theoretical." The same goes for statistics--some even top tier programs (e.g., NCSU), will be more applied than the more theoretical biostatistics programs. We're not talking about statistics vs biostatistics here, we're talking about UW Biostatistics vs. CMU Statistics. 

 

[Stat Assistant Professor]

1 hour ago, footballman2399 said:

This seems to be the thinking from many people, and while it may be true in some programs, it's not true in the others. For example, I attend a top-5 biostats program (not UW), and our coursework covers measure theory, limit theory, decision theory, etc. just like a traditional statistics program would.

I don't think you can put all biostatistics programs in a vacuum and simply say "well it's biostatistics so it's less theoretical." The same goes for statistics--some even top tier programs (e.g., NCSU), will be more applied than the more theoretical biostatistics programs. We're not talking about statistics vs biostatistics here, we're talking about UW Biostatistics vs. CMU Statistics. 

 

Out of curiosity: would you say that a prospective biostatistics PhD student should have at least *some* interest in research problems that are motivated by real data sets and problems from public health/medicine/biology (something which is not always the case for Statistics PhD programs)? This doesn't seem to be an issue for the OP, I am just curious.  

And I do see that many good biostatistics departments require their PhD students to take year-long sequences in some advanced topics from probability and statistical theory (e.g. UCLA has such a year-long sequence in inference, as well as a year-long sequence in linear models/generalized linear models). So mathematical rigor and training in theory are not in question. It's more the "day-to-day" research I am curious about. I actually applied to a lot of biostat postdocs and postdocs in more interdisciplinary environments (like the University of Chicago Booth School, where I could also work with econometricians), because I wanted the opportunities to work on real data and branch out beyond hardcore theory research. But for thesis work, am I wrong in thinking that a Biostatistics dissertation and "day-to-day" research would be much more likely to be motivated by methodology on real data sets and public health/medicine problems than on proving new theorems and the like? That's just the impression that I got from browsing the dissertation titles at various institutions (e.g. the ones at Harvard or at my current institution), but I could be wrong.

[bayessays]

Applied Math to Stat, in my opinion you are exactly correct. You will get a great theoretical training at the top biostatistics departments, on par with top statistics departments in some cases. And for this reason, most people on this forum give the advice that it doesn't matter which type of department you go to. This is what happened to me, and it personally ended up being completely false - if you aren't at least somewhat interested in collaborative bio/public health research, the day to day work might get to you. The fact that all of your classmates are talking about the cancer research project they're working on might bug you if you're only interested in math. The public health course requirements might grind your gears. Think very seriously about this. 

[StatsG0d]

22 hours ago, Applied Math to Stat said:

Out of curiosity: would you say that a prospective biostatistics PhD student should have at least *some* interest in research problems that are motivated by real data sets and problems from public health/medicine/biology (something which is not always the case for Statistics PhD programs)? This doesn't seem to be an issue for the OP, I am just curious.  

To this question: I would say yes--they should have at least some interest in research problems that are motivated by real data sets. I would say a big difference between statistics and biostatistics is that in biostats, you're presented with a real data problem first, and then you try to work out the theory / methods to solve such a problem. In statistics, some of the research may be motivated by real data problems, but this is not necessarily so. Some are simply just about the math (e.g. probability theory). So in this regard, I would say if the OP wants ultimate research flexibility, then maybe the statistics program is better.

 

22 hours ago, Applied Math to Stat said:

But for thesis work, am I wrong in thinking that a Biostatistics dissertation and "day-to-day" research would be much more likely to be motivated by methodology on real data sets and public health/medicine problems than on proving new theorems and the like? That's just the impression that I got from browsing the dissertation titles at various institutions (e.g. the ones at Harvard or at my current institution), but I could be wrong.

For this I would again say I think it depends heavily on the program / with whom you are working. In most (probably >90%) cases, probably yes. On the other hand, one of my professors did his PhD at UW biostatistics and his dissertation was on martingales, which granted are useful for biostatistics, but I would not say motivated by a real life data problem. He now works in precision medicine and his students try to prove various estimators are consistent and what not. Very occasionally, this may result in proving of a new theorem. I suppose this contrasts with a statistics department.

You mention Harvard, which no one will deny is a stellar program. But I think that in statistical theory, there is a *much* higher focus at UW than at Harvard.

[statcan]

How might these programs compare with Wisconsin-Madison Statistics?

[Catsharknado]

3 hours ago, statcan said:

How might these programs compare with Wisconsin-Madison Statistics?

 

They're both a little more prestigious than Wisconsin-Madison, but UWisconsin is still a very strong, competitive program. Just eyeballing the rankings, UW is a top 5 program, CMU is a top 10 program, and UWisc is a top 15 program. You can't go wrong any way. Once you're deciding between programs of that caliber, research fit, choice of mentor, and even location are more important than prestige.

Edited February 22, 2018 by Catsharknado