Does it matter what degree a potential Ph.D. advisor holds?

[orchidnora]

I'm looking for potential Ph.D. advisors at various schools. However, some schools (like Duke) have many professors with joint appointments. These people are listed as professors of Statistics, but when I go to their websites, I discover that the Ph.D. was in a different field, such as Electrical Engineering, Machine Learning, or CS. Other than the actual degree, they appear to do statistical research that I'm interested in. Does it matter for statistics if I am advised by somebody with a Ph.D. that's technically in a different field? I would love to hear anybody's thoughts on this matter. Thanks!

[bayessays]

If they are mainly in the non-statistics department and have an honorary connection to the statistics department, they're probably not the best choice.  I'd look at what journals they're publishing in.  For example, a professor who does non-statistics related computer vision research and has a courtesy appointment in stats is not a good choice if you want a job in a statistics department.  If you work with Kosuke Imai at Harvard, who has a PhD in political science but is a king of causal inference, you'd probably have better luck.   It's not about their degree - it's more about how likely they are to be doing the research that gets you a tenure-track job in a stats department.

[insert_name_here]

I don't think it matters what their degree is in. Especially if you're doing ML type work, practically the gap between a CS/applied stat/ML PhD can be very small/non-existent.

[Stat Assistant Professor]

It matters most what journals or conferences they are publishing in and what their reputation in the statistics community is, not what their degree is in. For example, Michael Jordan from UC Berkeley is a very well-renowned researcher in both Statistics and Computer Science, even though his PhD is in Psychology/Cognitive Science -- Jordan later switched his research field from psych to mathematical statistics and machine learning... kind of random, but he was a smart guy, and I guess he taught himself statistics, because he pioneered some fields of statistics that are very active research areas (namely, in topic modeling, variational inference, and Bayesian nonparametrics). Michael Jordan's lab and his PhD students/postdocs consistently publish in top statistics journals like JASA, Annals of Statistics, etc. as well as a top CS conferences like NIPS, so working with Dr. Jordan would absolutely be very good for someone's job prospects (I know some of his former PhD students have gone straight from PhD to TT faculty at places like MIT).

I would take a look at what venues these faculty members are publishing in and what the job placements of their former PhD students are.

Edited February 6, 2019 by Stat PhD Now Postdoc

[Gauss2017]

I agree generally with what Stat now has to say.  Sometimes the analysis of the right fit needs to go a little deeper.  Michael Jordan is incredibly brilliant and has lots of students. He is also very busy.  It is therefore difficult to get facetime with him if he is your advisor.  Sometimes it is better to take an advisor who can spend more time with you.

[orchidnora]

Thanks to all for the insights! I never would have guessed that Michael Jordan DOESN'T come from a math/stats background. That really puts it into perspective for me. Since I am mostly interested in computational stats/ML, it makes sense more sense now that I've been drawn to these professors with strong ties to EE/CS. I've been looking at the journals potential advisors publish in, as well as the student job placement, and now I feel more confident about my choice of advisors. 

[Stat Assistant Professor]

Yes, if you are very motivated and driven, you can teach yourself. That's what a PhD program and postdoctoral training are, for the most part... learning how to teach yourself new things. The coursework is useful for gaining a basic foundation, but after you're done with classes, it's all about teaching yourself (with guidance from your PhD advisor, of course).