JHU PhD Biostatistics and Columbia PhD Statistics

[Gill1990]

Hello guys, I was fortunate to receive offers from Johns Hopkins PhD in Biostatistics and Columbia PhD in Statistics. May I have some sincere opinions on school choices? 

At this stage, I'm open to both industry and academia opportunities, maybe leaning towards academia.

Regarding research interests, again I'm open. My slight concern is that biostatistics may be a bit narrow to land a job? I'm really not sure about this. 

 

Thank you very much! 

[Robatum1030]

Mega-congratulations on those two acceptances! Hopkins is consistently ranked the #1 public health school in the country and their biostatistics program is definitely up there, too. However, Hopkins is in a terrible location to me. I currently live in Baltimore, and it just is not a comfortable place to live in (for me, anyway). This fall, I will be starting a biostatistics PhD program at Columbia (that was my dream school, more so than Hopkins). Columbia's statistics program is in the top 15, which is not at all bad. The school still comes with great name recognition. It really depends on the location you want to live in and your research interests (look at faculty research in both departments and see which stands out most to you). I chose Columbia because I am really interested in entering the pharmaceutical industry after grad school, and they have a whole bunch of pharmaceutical stats courses and professors who do drug-related research that I think will best prepare me this type of career. Columbia is also pretty close to home and friends for me, which played a role in my decision. I honestly think both schools will get you very far in either academia or industry. Also consider which financial aid package is better. Living in Manhattan is more expensive than living in Baltimore. Regarding the "narrowness" of biostatistics, I think if you want to do more applied stats work after grad school, biostatistics would be a better fit. Most stats PhD programs are very theory-heavy, which is why I chose biostatistics over stats. In the job market, I don't think it makes a huge difference in industry, but it might if you want to enter academia. Biostatistics and stats professor jobs can be different depending on the school's emphasis on theory vs. application, but either school is a fantastic option!

[Geococcyx]

I'm just another prospective student, so keep that in mind, but if it's opinions you're looking for, I can certainly provide that.

They're both really good schools, as you know.  You are narrowing down what you'll be doing during your PhD program a little by going biostat instead of stat, but Johns Hopkins is good enough that I wonder how much you'd be narrowing your options beyond your PhD program.  For instance, one of Duke's statistics professors got her PhD from Johns Hopkins.  You might want to push for some measure theory background at Johns Hopkins, but I bet you'd do fine on the academic job market anyways, although you might skew more towards looking at biostat jobs instead of statistics jobs coming out of Johns Hopkins instead of Columbia.  Your interest in stat vs. biostat faculty positions might inform your choice a bit -- I get the sense that biostat professors tend to be co-investigators on lots of grants, and support a lot of their salary through that sort of research (or at least, they do at Hopkins -- feel free to check out Roger Peng's podcasts, his perspective may be helpful -- since this is mostly about academia, I'm guessing The Effort Report would be more helpful).  Meanwhile, it sounds like stat professors have more hard money, and have more teaching responsibilities and fewer collaborative research requirements.  

Being super open to research topics is a good and realistic position, but it does make it a little difficult to provide too many opinions.  For instance, if you were interested in imaging research, I'd imagine it would be pretty easy to pick Johns Hopkins (along with many other research areas).  I'm less clear on Columbia's research strengths, but I'm guessing from Gelman that they're strong in social sciences and Bayesian research amongst other strengths.  

Personally, I'd choose Johns Hopkins, but I'm particularly interested in neuroimaging and health applications, which makes it an easy choice for me.  Without knowing more about how applied vs. theoretical you research interests are (or any areas of particular interest), I don't know that I can really prescribe either one as being better for you.

[blehperson]

I don't know much about JHU, but I chose columbia's stat phd over my other options (harvard and UW). I think they are stronger in applied research than many other top stat phd programs, and have a broader definition of statistics research than many other schools. A handful of faculty work in computational biology and neuroscience, and they have a strong machine learning community that is integrated with the stat dept. At many other schools (such as UW) machine learning is dominated by CS, so this was a major factor for me. 

[bayessays]

If you're interested in the biostatistics research enough to write a dissertation on it, I don't think biostatistics will hold you back. JHU has great placements.  I don't think you can go wrong.  I'd consider the environment you would enjoy more (public health school vs. liberal arts) and which city you'd rather live in for 5 years.

[phdinstat]

Congratulations! You can't go wrong with both choices, but if you are open to the biostatistics research, JHU would be a great place to attend regarding its reputation in public health and biostatistics. The professors in JHU are worldly renowned for their research on various topics.

I'd choose JHU since their placements seem slightly better than Columbia.