Harvard MS Data Science vs Duke PhD Statistics

[stackleberg cat]

Hey everyone,

So I've heard back from 2/3 of the schools I applied to so far and currently my top choices are Duke PhD in Statistics and Harvard MS in Data Science. (I think the only other program that hasn't gotten back to me that I would consider in this debate is Stanford MS in Statistics).

I'm currently trying to figure out whether research is something that interests me. I did a few research projects in undergrad and from that, I think I'd enjoy doing a PhD but I'm not 100% certain which is why one reason I'm a bit hesitant to jump into Duke's program. Another reason I'm hesitant is because I don't have much experience with Bayesian statistics (but I think it might be interesting). However, Duke is a great school for statistics and I could always drop out after getting a masters. Harvard obviously has a great brand name, but I'm not entirely sure if the Masters in DS is a cash grab program. Also, I'd ideally want a masters program where I can get good research exposure. Harvard does have a thesis option and a couple of research classes, but I'm not sure whether this is a good enough way to hypothesis test whether I'd want to do a PhD? I can't seem to find good information on the Harvard Masters in DS website, so if anyone has any additional information, that would be greatly appreciated

Would appreciate any thoughts/advice that people have.

Thanks!

Edited February 19, 2022 by stackleberg cat

[bayessays]

Go to Duke for the free (and top tier) education.  If you don't like research, leave with the free MS degree. 

 

[j_li]

You should go Harvard MS only if 1. you are not into Bayesian at all; 2. you want to work under some famous faculty like Susan Murphy and you belive that you can be in the Harvard PhD program after you MS degree. or 3. Harvard MS provides full fnding and you are not into PhD program right now.

Duke prgram is one of the best Bayesian programs in US, if not in the world. Maybe CMU is better if you consider ML with Bayesian. If you are not in Bayesian, I saw that you have applied couple of good statistics programs like UMich, UCLA and Rutger. Those programs are not rank as high as Duke, but you may find some very interesting areas and outstanding advisors you can work under.

 For academic career, your PhD advisor is far more important than the school brand name.

[trynagetby]

If you're worried about Duke being too Bayesian, I'd consider the fact that Duke's becoming significantly less Bayesian recently. Fan Li does causal inference research that's pretty non-bayesian, Alex Volfovsky works with Cynthia Rudin in the CS department on pretty straight up ML research, Jason Xu does mainly ML, Eric Laber (Susan Murhpy's student) does Susan Murphy type stuff.

So if you're not a hardcore Bayesian there's still a lot of flexibility.