Chicago PhD vs Duke PhD vs UPenn PhD

[Rob1234]

I have got an offer from University of Chicago, llinois Statistics department and have also had encouraging conversations through Skype about a possible offer in near future from Duke University and University of Pennsylvania. The suggested financial support for Upenn is  more than both Chicago and Duke, which are around the same. I would like to do my PhD in Theoretical Statistics but I am in a fix which offer should I choose. What do you guys suggest and how would you rank these three Universities??

[DMX]

congrats!

given your theoretical bend I think UChicago might be a good fit (I am also selfishly motivated since I am waiting to hear back from Wharton (my first choice) and Duke)

Edited January 31, 2013 by DMX

[crazystat]

If you want to do the theoretical statistics or probability, Univ of Chicago is the best choice. 

[crazystat]

Actually, for theoretical Stat, Upenn sucks.

 

If you want to do Bayesian related work, you may consider Duke actually 

[crazystat]

Upenn is best for industry track. Univ of Chicago has very strong Statistics program. As far as I know, their program is pretty hard. But lots of students have good publication with advisor during their PhD study. 

[creed_the_third]

Actually, for theoretical Stat, Upenn sucks.

Really? I read in some other thread here that penn was a pretty theoretical place.

Edited February 1, 2013 by creed_the_third

[crazystat]

Really? I read in some other thread here that penn was a pretty theoretical place.

 

Some professors are pretty excellent, for example, Tony Cai. 

 

But I don't think their program is comparable with Chicago's stat. 

[efh0888]

I mean Upenn's stat dept. is housed in the business school, so it makes sense...

[JZappa]

I'd definitely say UChicago. Duke is nice but I'd be wary of the fact that they are practically 100% Bayesian. Unless you know absolutely that's what you're interested in. Also I would say UChicago is a better school for theory than Duke.

[outofstate]

All three programs are very strong and good for future academic placements. The choice should really be based on your statistical interests. Historically Chicago was a very strong theoretical department when Bahadur was there but that is now a long time ago. Now Chicago is very strong in terms of applications to say genetics or finance. The theoretical focus is in probability rather than statistics. Duke is clearly the place if you want to focus on the Bayesian approach. Upenn is not a good choice for

industry as it is focused much more on theoretical statistics as well as areas such as causal inference and in both these areas it is very strong. It is strange that it is in the business school as few of the faculty are interested in business.