another thread about choosing a statistics program

[wine in coffee cups]

Hi everyone, been going on visits and doing a lot of thinking but having a hard time converging on a single statistics PhD program. I've narrowed it down to Berkeley, Duke, and the University of Washington, all of which I visited and really liked. These are all great programs so in most senses I can't go wrong, but I can't enroll in all of them and need to make a decision soon under some basis. My leanings keep drifting each day and my gut instincts aren't kicking in, so any comments or advice to spark some new lines of thinking would be much appreciated! Rambling thoughts below, which sadly are condensed versions of what's been turning over in my head and in conversations with my beleaguered friends and colleagues.

Berkeley: Fantastic location, top notch reputation, and everyone I talk to who isn't in statistics thinks this is the glaringly obvious choice. My research interests are in applications of statistics to the social sciences, though, and there are only a couple of faculty who would make sense as potential advisors within the department, both of whom I liked but whose research is a little removed from my own interests (still curious to learn about, it's just a matter of degree). I can have an advisor outside the department to give me more options, but that's more work on my part to make those connections happen. I'm thinking the program might still just be too theoretical for me but it's hard to say, they don't think so, and I don't want to rule it out on this basis prematurely. A current student there finishing up who has a similar applied orientation and worked with both of the faculty I'm interested in gave me a lot of reassurance that this can work for me, but it still seems like more of a stretch than the other two departments. I made the mistake of looking up the webpages of current students and am cowed by how accomplished they were upon entering the program -- I get the sense that I'd be an intellectual runt and am concerned about potentially not being able to complete the degree in a timely fashion since I am not coming in with a master's (NRC data suggest as much but I'm not sure how relevant those dated numbers are). If I did make it out unscathed, though, I'd have solid placement potential. I've been offered a VIGRE of $27K for the first 12 months with some light obligations, but don't have a great idea of what my package would look like in subsequent years (certainly a mix of GSI and GSR, just not sure if the dollar amount would be that different -- I suspect not). High cost of living for sure, but current students sounded like they were getting by okay on their packages. I have a few friends in the area but no other strong ties.

Duke: My impression when I visited was that it was cultish, for lack of a better term (very into themselves and unabashedly Bayesian), but in a way I was okay with. I appreciate the efficiency of the department in getting students involved in research early and getting them out quickly. There are a couple of faculty there who I really liked, particularly one in a specialized area I am interested in that isn't represented in the other departments and whose students had glowing things to say about working with him. I think it's a slightly better fit than Berkeley research-wise. The various Research Triangle connections are appealing, no shortage of interesting and funded work to go around. My concern -- based mainly on the reactions of faculty and students at other departments when asked where else I'm considering -- is that this program might be viewed as a little narrow, has ascended the rankings too recently to have a reputation, or something. In any case, for reasons not known to me I'm getting puzzled looks about liking it as much as I do, so I'm wondering if I've drunk the Kool-Aid and need to step back. I'm not connected enough to the academic statistics community to be able to figure out what the perceptions are and am having a hard time getting an informed sense of how graduates of this program are viewed. One major upside is they've given me an obscene financial package that includes no responsibilities outside of classes my first semester (so that I can ease back into being in school since I've been out for a while), a RA position my second semester to jump right into research, and $32K for my first 12 months (I think staying at that level every year). Given the cost of living in Durham, that's quite a chunk of change, and I really like the department so I'm loathe to turn down this generous and tailored offer lightly. I'm not sure how I feel about Durham as a town, certainly less exciting than the other two options, but it's still a decent-sized city and I think I could be happy living there, it just doesn't vibe with my crunchy urban liberal granola yogurt orientation as obviously as Berkeley and Seattle. Main downsides of living there are that I have no personal ties (don't want to feel socially isolated and making friends outside the department might be hard), the weather is too hot in the summer months, and I despise the basketball culture.

UW: My main research interest is in applications of statistics to the social sciences, and they have a whole interdepartmental center working on these very topics! So in terms of research fit it's a slam dunk, there are many great potential advisors and I really liked the faculty and students I met. I also like the integration with biostats for the first year coursework for social reasons -- makes a relatively small program feel a lot bigger. My impression is that it's well regarded in the statistics world and that I should be able to place decently coming out of it (even if I bristle at what a Berkeley prof told me, that UW is good but Berkeley is "a tier up"). Unfortunately, this is a cash-strapped public university, and while I am still adequately funded, it's worse than what I'm being offered at the other two departments (well, maybe comparable with Berkeley once cost of living is accounted for). This includes a couple of supplemental fellowships as it is, so I doubt any additional funding is possible, especially given the department's financial troubles a few years ago. (I will say that current students seemed surprisingly unconcerned about this.) My first year package is about $22K for 9 months, but I believe that will drop to around $16K in subsequent years if I don't find outside funding and will have to do some legwork to figure out summers. This also comes with the most onerous TA time demands. I'm in a good situation to deal with this because I have a lot of savings and don't mind being kept honest with some TA work, but all things considered, the other offers would make for an easier transition back into the academic world. I have strong family/friend connections to the Seattle area, particularly love the music scene and having access to the mountains, and know I would be happy living there again.

So: three great options with relatively minor downsides in the grand scheme of things and I'm dreading having to turn two of these down. The other relevant dimension to this is the two-body problem. Partner of many years is finishing up his PhD in pure math next spring. He's at an okay pure math department (top 50? 40 maybe?) but he's at least one of their star students and his advisor is well regarded in his subfield. He will be trying to do an academic job search this fall targeting the geographic area I choose. For getting through that first year apart, Duke is the winner because of shorter flights, whereas one or both of us will be in a jet-lagged fog and burn through a lot of money if we try to do the cross-country haul to/from Seattle or Berkeley more often than every couple of months. But, it's just a year apart, so it's an important consideration but I'm trying to think in the short, medium, and long terms. Job opportunities aren't great for his field, but picking the least overtly sucky of these areas would be nice. Duke has the other triangle institutions nearby so there would potentially be a decent number teaching positions, but his particular subfield doesn't seem to have as much activity at any of the three universities (so postdocs are not too likely) and it's not clear how competitive the teaching positions are or what other opportunities are available outside of those schools. Berkeley and Stanford being the main players in the Bay Area makes me think the competition is going to be too intense to be viable. Seattle has the UW but not much else aside from some smaller private schools, I think the market isn't great but there is some potential for postdocs and teaching positions. I'm wishing now that he had a more applied orientation because that would open up more doors and make this easier, but you can't change the one you love, etc.

If you made it this far, thoughts?

[Biostat_Assistant_Prof]

Hmm, I don't think I could point out anything that you don't already know. I really think you should go with your gut.

Personally, I would be torn between Duke and Berkeley. Berkeley because of the location/weather and Duke because of the financial package.... $32k will get you more in North Carolina than it will in the Bay Area or Seattle; you could live a very decent life there with that financial help,

[sparkplug]

Obviously I can't give you any insight into the academics and research fit for these schools you didn't already know. So I'll try to throw out my two cents based on some of the other things you mentioned,

Research fit considerations aside I think I'd have a harder time picking between Berkeley and Seattle. I have a similar "crunchy urban liberal granola yogurt orientation," and from having been to all 3 of these areas before I definitely felt that Duke was the odd one out. Basketball culture, as well as all the other cultural/social associations that come with that, can definitely be annoying/abrasive. I now live in Madison WI (similar to Durham in size and it's love for university sports) after coming from Berkeley, and I've found the vibe "off" enough where I don't think I can be here for 6 years, let along work in the area afterwards. Obviously the first priority is to pick the school that will be the best for you academically/professionally, and the one where you will be well supported by faculty and a good financial package, but you would probably fit in better near a major metropolitan area during school and even after your PhD. Personally (and based on my experiences this past year), I don't think I would be able to live in an area I'm not crazy about, no matter how great the school is. But I'm probably a lot more socially particular than others so I don't know how much my picky preference is worth.

Anyways, you probably already knew I'd be rooting for the west coast. Good luck to you in trying to decide between some phenomenal programs!

[mathgirl28]

wow! you have thought quite deeply about this (as you should). I dont have any concrete suggestions (well maybe some), but more impressions:

1) I dont think you should be intimidated by Berkeley. There are no qualifying exams, and the students seem to band together to finish problem sets etc. (i.e. while you will be challenge, I do not think there is a chance of you falling behind. It seemed more like a learning together atmosphere). That challenge might be the best thing that ever happened to you (who knows) =). Also, the time in which you finish your phD (im willing to bet) will be the same at Berkeley as it is at other schools. The reason i say this is because if you're lucky, your research area/topic will be one of deep interest to you as well as one you are or become strong in. That process will be the same no matter where you go. I personally think that an atmosphere where you feel like everyone is a rockstar is a good thing because it pushes you (and if your going to academia, you need that).

2) In terms of jobs, there are other schools around California that will be doable in terms of making trips, but probably not pleasant (UCSF, UCSD, UCLA etc.). I have a friend who makes the UCLA-Bay area relationship work, so its doable. I really think you should tackle that problem as it comes. That said however, Duke would probably be a safer bet on that front since it is in New England and while he might not get something that close by (or even in the same state) there are more schools in New England to choose from.

3) It seems like your decision is more between Berkeley and Duke. It also seems like you want to go into teaching, so you ought to weigh the strength of the advisors you want to make the decision and not necessarily school reputation (although this is still very important). Start email correspondences with the professor(s) you want to work with (that is what i have been doing at least, and they have been very receptive). Further engage them about projects you think you might want to work on and even let them know how you are thinking about the decision.

Ultimately, i think if you can find a professor at Berkeley who wants to work with you (and on what you want to work on) you should go for it. Its one of the strongest programs and i am willing to bet you will succeed there. If you find that person(s) at Duke, and you click with them better, go for Duke and don't look back. I think you will be fine coming out of either one.

Also, I really wouldn't worry about financing at either one. I think that should be one of the last things you look at as the stipends at all places will be livable. (i.e I wouldn't pick Duke because i could be living in a better apartment...you could end up receiving a fellowship down the line (you should apply for some in your first year) and you wont be living badly even if that doesn't happen...)

This is just my two cents...

Edited March 28, 2012 by mathgirl28

[Statistique]

Wow, I'm surprised everyone is jumping on the Berkeley/Duke bandwagon, because my first instinct was "(s)he should totally go to UW!"

Just to pluck out a few points where I have opinions:

Point: Every single person I speak to cautions me against deciding based on money. Your offers are all very substantial - more than enough to live on. It's certainly tempting to get SO much more money per year, but it'll be a drop in the bucket compared to differences in job offers, so you ought to choose the place that will get you a good job that you'd be happy doing.

Counterpoints: If you have debts to pay or partners/children to support, that might negate what I said. Also, while money amount may not be a consideration, what you have to do to earn that money is. Onerous TAships are a big negative for UW, especially if that expectation continues into your research years and chews up all your time. Also, while your stipend may not be a factor, a "cash-strapped" place might have issues with chasing away professors by lowering salaries, not having computing resources, etc. Something to ask about.

Point: Having top professors who align with your research interests is absolutely the most important thing. Your job opportunities (if you go into academia) will come from rec letters from well-known profs, not from the ranking of your school. This will also impact your quality of life hugely; presumably you are getting a PhD to do what you love, and deviating from that totally defeats the purpose.

Counterpoint: I don't have one. Seriously, if it were me, the research match at UW would totally trump everything else. (Unless that match is not as perfect as your post seemed to imply.)

Point: Don't sweat it about being an "intellectual runt". You got into three amazing schools, with absurdly generous offers at all of them. Obviously, you are awesome.

Counterpoint: One thing to keep in mind, though, is competition for advisors. This seems to be a concern at some places and not others. You're better off being advised (or co-advised) by a top professor than going to a school with great faculty but working only with unestablished people. So, while I'm sure you would excel at any of these places, ask around about this issue. Especially for Berkeley, where there seems to only be one person who fits your research well.

Point: UW may be a "tier down" from Berkeley...

Counterpoint: ... but that's because Berkeley, Stanford, Harvard, and UChicago sort of occupy their own subset of the top tier. And I've often seen UW thrown into that ring as well. If it's a step down, it's not much of one at all.

Point: Your non-stats friends are reacting to the undergraduate reputation of the schools, not their stats dept ranking. Ignore them.

Counterpoint: If you intend to go into industry, etc. rather than academia, namebrand could matter.

As for the rest of it, I can't give you much advice... in fact, I'm struggling with a similar issues (Chapel Hill vs. Bay Area, jobs for Boyfriend vs. my career). In the end, only you can decide where you'd be happiest and only you (well, you and your partner) can decide how much he factors in to the decision.

OH and: your estimation of Bay Area schools is wrong, I think. UC Davis and maybe UCSC are both close enough for you to see each other (though maybe not live together), and there are a ton of state schools in the area (SF, SJ, and Sac) that may or may not have jobs. I can't speak for the other areas, but take a closer look at Bay Area schools, it's not just Berkeley and Stanford.

Good luck, friend! Remember that with options like yours, you'll have a great time wherever you end up!

[danthechub]

This is like when people complain that they can't close their wallet because there's too many bills in there. You have a real problem but it's not a bad problem. I say Berkeley or Duke-- either for Prestige/Location or Research fit/Non-intimidating/Partner consideration. UW is obviously also great, but even though that center for interdisciplinary work sounds good, it's a small program so you never know and I hear they have funding issues too. Just a brief 2cents but you really can't go wrong-- solid offers all around!

[wine in coffee cups]

Really helpful responses, thanks to everyone so far for your thoughtful input! I can't say I'm closer to a decision yet but hopefully I can get my act together soon. And apologies for being that annoying person complaining that my life is toooooo awesome.

[zcyharrypotter]

I see you mentioned UC Davis and I want to know that could you offer more information about the MS in stat program of this school especially in the graduate placement part. Thanks a lot!

Wow, I'm surprised everyone is jumping on the Berkeley/Duke bandwagon, because my first instinct was "(s)he should totally go to UW!"

Just to pluck out a few points where I have opinions:

Point: Every single person I speak to cautions me against deciding based on money. Your offers are all very substantial - more than enough to live on. It's certainly tempting to get SO much more money per year, but it'll be a drop in the bucket compared to differences in job offers, so you ought to choose the place that will get you a good job that you'd be happy doing.
Counterpoints: If you have debts to pay or partners/children to support, that might negate what I said. Also, while money amount may not be a consideration, what you have to do to earn that money is. Onerous TAships are a big negative for UW, especially if that expectation continues into your research years and chews up all your time. Also, while your stipend may not be a factor, a "cash-strapped" place might have issues with chasing away professors by lowering salaries, not having computing resources, etc. Something to ask about.

Point: Having top professors who align with your research interests is absolutely the most important thing. Your job opportunities (if you go into academia) will come from rec letters from well-known profs, not from the ranking of your school. This will also impact your quality of life hugely; presumably you are getting a PhD to do what you love, and deviating from that totally defeats the purpose.
Counterpoint: I don't have one. Seriously, if it were me, the research match at UW would totally trump everything else. (Unless that match is not as perfect as your post seemed to imply.)

Point: Don't sweat it about being an "intellectual runt". You got into three amazing schools, with absurdly generous offers at all of them. Obviously, you are awesome.
Counterpoint: One thing to keep in mind, though, is competition for advisors. This seems to be a concern at some places and not others. You're better off being advised (or co-advised) by a top professor than going to a school with great faculty but working only with unestablished people. So, while I'm sure you would excel at any of these places, ask around about this issue. Especially for Berkeley, where there seems to only be one person who fits your research well.

Point: UW may be a "tier down" from Berkeley...
Counterpoint: ... but that's because Berkeley, Stanford, Harvard, and UChicago sort of occupy their own subset of the top tier. And I've often seen UW thrown into that ring as well. If it's a step down, it's not much of one at all.

Point: Your non-stats friends are reacting to the undergraduate reputation of the schools, not their stats dept ranking. Ignore them.
Counterpoint: If you intend to go into industry, etc. rather than academia, namebrand could matter.


As for the rest of it, I can't give you much advice... in fact, I'm struggling with a similar issues (Chapel Hill vs. Bay Area, jobs for Boyfriend vs. my career). In the end, only you can decide where you'd be happiest and only you (well, you and your partner) can decide how much he factors in to the decision.

OH and: your estimation of Bay Area schools is wrong, I think. UC Davis and maybe UCSC are both close enough for you to see each other (though maybe not live together), and there are a ton of state schools in the area (SF, SJ, and Sac) that may or may not have jobs. I can't speak for the other areas, but take a closer look at Bay Area schools, it's not just Berkeley and Stanford.

Good luck, friend! Remember that with options like yours, you'll have a great time wherever you end up!

[Rob1234]

There have already been a number of suggestions, each point well-justified too. But I would still like to throw in my opinion mainly because, I can very much relate to your problem given that I am having to choose between Duke, Chicago and Harvard now and am finding it an increasingly difficult task. I was inclined towards Duke initially due to my Bayesian research interests but now I am having to re-think after asking people around. The Duke Stat dept doesn't seem to be a very popular choice which may be attributed strictly to the cult Bayesian following that you mentioned. However, in the last 4-5 years the academic and industry brand value of Duke has increased quite a lot but still, I get the general feeling that once I step into Duke, the opportunities will still be quite limited academic career wise (unlike in industry) and mostly centered around other Bayesian departments (from looking at the Alumni positions and asking seniors at Duke). However, 5 years down the line the academic market can also open up though there is still an element of doubt. Also I think the coursework at Duke is quite limited (unlike Berkeley or Seattle). But on the positive side, the research productivity of Duke is extremely high and they mainly work on modern research areas. Also, because of SAMSI and UNC and NCSU close by, it is quite easy to build your contacts. Another thing I've heard is that the work environment in Duke is extremely friendly and they are very flexible in terms of working in and out of the department too.

 

I think most people you ask may be a bit biased towards Berkeley and Seattle as these two are broader in terms of research interests, have better reputation and better rankings. However, Duke has many positives too and it seems you may be a little inclined towards Duke too because of the research area of interest, two body problem etc. And from what I have realized now myself, seeking suggestions can never give you a clear solution as people tend to have varied opinions and there is no uniformly best solution in your case. So, you have to fix your priorities considering all options and then choose one that you think would be the best fit (and this, from personal experience, is a pretty hard thing to do especially given your options).

[tcbh]

just a little PSA

this thread is a year old, OP has already made her choice

[ilovefish]

Given your interest and information posted here, my intuition telsl me you should go UW. I think you've already put UW in the high ranking: there are most suitable advisors in UW; the Department has tight cooperation with excellent Biostat Department. I think the financial problem in subsequent years isn't so big in UW given your strong academic background to get offer from Berkeley. There are many financial opportunities. I vote for UW! Good luck!

[Statistique]

I see you mentioned UC Davis and I want to know that could you offer more information about the MS in stat program of this school especially in the graduate placement part. Thanks a lot!

 

I think this person meant to PM me or something.  We can let this ancient thread die out now I'd say.

 

To answer the question:  Unfortunately, I didn't end up going to Davis, so I really have no idea about the MS program or about placement.  Really sorry.