UC Santa Cruz Statistics PhD Program?

[coconuts2]

Anyone familiar with UC Santa Cruz's Statistics PhD Program? The vicinity to the silicon valley seems very appealing but I'm a little bit concerned about the department leaning heavily towards Bayesian. Is UCSC phd program a promising one?

[bayessays]

They have some very good Bayesian people, although I think they recently lost Abel Rodriguez.  What exactly is your concern about the program being too Bayesian? You are correct that, like UT Austin and Duke, the department is almost exclusively Bayesian. If this is an issue for you because you are not interested in the subject, then probably do not attend because there will not be many other research options.  But if you are interested in Bayesian stats, it's a pretty decent department, although not at the level of Duke/UT.

[statenth]

Though you may not be personally interested, I think Bayesian is also a promising field given its applications on graphical models and theoretical formulations of neural networks (like AlphaNet). But I totally agree with @bayessays that it's better to apply to the places with more wide options that interest you.

Edited February 3, 2021 by statenth

[Stat Assistant Professor]

Agreed, UCSC is a very good program, and they have decent academic placements (if you're interested in that). They've placed some PhD graduates at University of Chicago (Matt Taddy), University of Florida, and other good places in the past. The Statistics Department is relatively new (it was part of the Applied Math department until around 2019), which is why UCSC may not be ranked in the USNWR.  

If you aren't interested in Bayesian statistics *at all* though, then you probably shouldn't apply there. One thing I would note though... if you decide to go into industry, I'm not sure how much of your PhD dissertation research you would really use for the "typical" jobs anyway (regardless of whether you study Bayesian or frequentist stats for your dissertation). Unless it's a very research-oriented industry job like Microsoft Research, Google Brain, or something of that sort, you probably will not use a ton of the stuff you learned in your research.

[bayessays]

Unless you have a strong aversion to something about Bayesian statistics (you hate computing, you have some strong philosophical objection, or you already have a clear and strong research interest in something frequentist), I don't think most people would suffer at all by attending a Bayesian-oriented program. You'll have all the same skills for industry jobs, plus arguably some more useful ones too. I looked at the curriculum at UCSC and you will learn everything you would in more "classical" PhD program, so the only difference would be your dissertation would likely have to be Bayesian, which you might do anyways in any other program.

[supremum]

On 2/3/2021 at 4:39 AM, bayessays said:

Unless you have a strong aversion to something about Bayesian statistics (you hate computing, you have some strong philosophical objection, or you already have a clear and strong research interest in something frequentist), I don't think most people would suffer at all by attending a Bayesian-oriented program. You'll have all the same skills for industry jobs, plus arguably some more useful ones too. I looked at the curriculum at UCSC and you will learn everything you would in more "classical" PhD program, so the only difference would be your dissertation would likely have to be Bayesian, which you might do anyways in any other program.

It essentially comes down to research options (and definitely not what you intend to do coming into the program, since that invariably changes). I find that the more Bayesian focused departments are often pretty heavily skewed (and probably in the limit will converge to a Bayesian-pure department, not unlike Duke), whereas most departments have a little bit of everything, which just gives you more choices. Why that's the case, I have my hunches, but that's besides the point.

[Muscatel Darjiling]

I'm a theory PhD at UCSC CS; stat department is good, probably will expand over the next few years. As others have noted, they're very heavily Bayesian; someone also noted that Abel Rodriguez left but he's part of a 4 institute NSF grant on data science and and moved from one of those (UCSC) to another (UW) and there's still enough chatter going on b/w these institutes. 

It also depends on what you expect from a stat PhD. If you're looking to do some analytical probability/stat mech type stuff, you probably won't get much of it at UCSC. Plus Stat dept works pretty closely with CS so there's that kind of work if you care about it. 

[bob loblaw]

I'm super interested in Bayesian Stats & this program so I have some insights. 
For full context, I have no interest in academia; I want to be an applied statistician/quantitative researcher role in government/industry.  

Point 1: UCSC is VERY Bayesian.
Like people above have said, if you know pretty strongly that you're not interested in Bayes, UCSC might not be the place for you. 
Alumni comment: "Duke, for example, is known for being Bayesian. But they have a large Bayesian wing. At SC, the whole department is Bayesian".

Point 2: I don't see any disadvantages of going to a Bayesian program for my career goals personally
Even as someone interested in Bayes, I had some hesitance of going to a super Bayesian program. 
Here's a question I asked to an alumni who works as an applied statistician at RAND. 

Q: is there a disadvantage being in a Bayesian-heavy program when you become a statistician (in a place like RAND or the federal government) because a majority of statisticians are frequentist? 
A: Good question. I would say it is not a disadvantage, as a lot of places want to hire people who know Bayesian stats or are increasingly interested in using Bayesian statistics. 
However, it is true that a lot of my work ends up being non-Bayesian, and there are courses I wish I could have taken that UCSC did not offer (perhaps they do now?).
These include causal inference and survey sampling/experimental design. These topics are very important at RAND or similar places, and were not a focus of the coursework at UCSC.  

Re: casual inference... they hired Richard Li (UW) last year who's research interest and background includes this area so perhaps they are filling this gap. 

As others have mentioned above, it may not be a disadvantage for academia either but I don't really know anything about this.

Edited March 4, 2021 by bob loblaw

[Stat Phd]

On 3/4/2021 at 3:21 PM, bob loblaw said:

I'm super interested in Bayesian Stats & this program so I have some insights. 
For full context, I have no interest in academia; I want to be an applied statistician/quantitative researcher role in government/industry.  

Point 1: UCSC is VERY Bayesian.
Like people above have said, if you know pretty strongly that you're not interested in Bayes, UCSC might not be the place for you. 
Alumni comment: "Duke, for example, is known for being Bayesian. But they have a large Bayesian wing. At SC, the whole department is Bayesian".

Point 2: I don't see any disadvantages of going to a Bayesian program for my career goals personally
Even as someone interested in Bayes, I had some hesitance of going to a super Bayesian program. 
Here's a question I asked to an alumni who works as an applied statistician at RAND. 

Q: is there a disadvantage being in a Bayesian-heavy program when you become a statistician (in a place like RAND or the federal government) because a majority of statisticians are frequentist? 
A: Good question. I would say it is not a disadvantage, as a lot of places want to hire people who know Bayesian stats or are increasingly interested in using Bayesian statistics. 
However, it is true that a lot of my work ends up being non-Bayesian, and there are courses I wish I could have taken that UCSC did not offer (perhaps they do now?).
These include causal inference and survey sampling/experimental design. These topics are very important at RAND or similar places, and were not a focus of the coursework at UCSC.  

Re: casual inference... they hired Richard Li (UW) last year who's research interest and background includes this area so perhaps they are filling this gap. 

As others have mentioned above, it may not be a disadvantage for academia either but I don't really know anything about this.

Hi did you mean Richard Guo? I didn’t find a causal person named Richard Li

[MLE]

49 minutes ago, Stat Phd said:

I didn’t find a causal person named Richard Li

Richard Li