localfdr

Indeed, at most universities you will receive the same funding as all American students, should you get accepted. At some departments there might less available funded positions for internationals though. For example, in some biostatistics programs the domestic candidates are funded by the NIH, while the graduate school pays for the internationals.

So if I understand it correctly, you want rigorous math, statistics and machine learning (and AI etc), but come from a biological background. Well actually it really depends on the University. For example, for UW I think you could apply to their biostatistics department, because they also have rigorous math courses and also machine learning (which is what you want I think). In contrast, other biostats programs place a lot more emphasis on computational biology/public health/epidemiology applications rather than on rigorous math (e.g. Harvard biostats), so these would not be a good fit for you. On the other hand, you will get rigorous math and a good dose of machine learning (if you are interested in that), in pretty much any statistics department, which is why I recommended it. Plus you can always do biological applications in stats departments. Of course, indeed your mostly biological background means that you will have a better shot at getting admitted at biostats programs. But still, I think you have covered almost all math courses that are important for a stats PhD program and some stats departments look for people with an applied rather than math-y background. Depending on your research experience/GPA/LORs I definitely think you could have a shot at getting admitted to stats programs. Very mathematical programs (usually the ones which also have probability theorists in the stats department, such as Berkeley) might be a longer shot, but maybe still doable with a very strong GRE math subject test. In conclusion, my advice would be to apply to some biostats departments which are similar to UW plus to a few stats departments.

I really think that what you mean by pure/theoretical math is what most people call applied math. Also your post is a bit confusing... But my general feeling is that a Statistics department would be a great choice for you! You can do biostat applications in most departments (including neuroscience which is not really traditional biostat), take quite theoretical math classes (measure theory), computational (programming, optimization) and statistical machine learning classes.

Well, they sometimes are. Based on my great sample size of 1+1 offers from these areas, public southern schools might offer something like 19k per year, while private California universities seem to double it.

Congratulations! I guess you cannot go wrong with either choice. From a purely academic point of view, I think most people would consider Chicago to have the stronger and more diverse Statistics department and for a PhD I would definitely go with Chicago (unless there was a particular faculty member I would like to work with at Harvard - as always). For a Master's degree, things are more subtle: Again from a purely academic point of view, I would choose Chicago, but connections and the name also matter a lot. I don't have any information regarding that and it might be advisable to ask people in the field and part of the world which you want to join later.

@cyberwulf Ok, thank you!

@cyberwulf is the dress code for recruiting visits also the same as for interview visits? Or is it more casual?

First of all, congrats on getting into two of the best biostats programs! I would make this decision dependent on whether you already have clearly defined research interests and whether there is a specific faculty member at Berkeley with whom you want to work with and who is also interested in taking you on as her/his PhD student. For example, I could imagine that working with Lior Pachter or Bin Yu could lead to a very productive PhD, assuming research interests (and personalities) match. But if you are not sure about a particular Prof. at UCB with whom you want to work with, then I would go to UW because it leaves you so many more choices. For example, I did not even apply to UC Berkeley, because I felt that it would not be a good fit for me. UW biostats on the other hand accommodates pretty much everyone interested in anything biostats/stats related!

Based on the titles of the thread, I would have said go for nonparametrics, no doubt (see next paragraph as well). But based on the description you posted, the nonparametrics course seems lightweight content-wise, especially since I guess it will not be a proof based class and does not cover cornerstones like kernel density estimation and nonparametric regression. On the other hand, the time series course covers more ground, so I would be inclined to choose that. Now regarding biostats: I think both courses could be useful. Depending on your biostats interests though, you might not come across time series material (really time series would be more important if you were looking into econometrics or mathematical finance or something like that), while the nonparametric stuff you will see everywhere. But I'd still go for the time series class, if only because it will increase your statistical maturity more. Plus you might still need it for biostats!

Me too, just got an offer by Stanford (guess where local fdrs were first proposed :D)!!!!!!! Wooohooo

Also let's see when CMU & Stanford will start answering..

As another case in point, Harvard's biostatistics department does *not* invite international applicant to the recruitment weekend; only domestic ones. In other words, if you are an international, depending on the school you are applying to, you should not worry at all about not getting invited!

Yes, to both. Though for schools which have both biostats and stats departments I applied only to biostats. Are you applying only to stats?

Yeah Chicago definitely does it in batches (i.e. I have not lost hope yet!) , also to quote from their FAQ: "We consider applications in batches. For example, one week we may consider all the applicants who are interested in a particular specialty or who are from a particular part of the world. We release decisions throughout February and March. If you have not received a decision from us, we are still considering your application." I do not really know about Harvard Stats since I have not applied there. At least it seems like they send a big batch of their admission letters simultaneously.

Yeah exactly! But again this is just my opinion, there are other people on this forum who are more qualified than I am to answer!

My feeling is that, with the exception of the Calculus issue, you should be competitive for these Master programs and that you should get into one if you apply broadly. But maybe someone else who has a better overview than I have can provide a better assessment. Now regarding your second question, this would actually worry me. As I said in my original post, not having multivariate calculus could be a major red flag to admission committees and I don't think you could alleviate this by trying to say you have this background in the SOP. To put it bluntly, based on what you wrote in your paragraph you actually *don't* have the necessary multivariate calculus background (since on the one hand you say that you know these things already by your stats courses, but simultaneously you believe you would not do well in that calculus course). So there's actually a second reason beyond satisfying the admissions committee for taking this Calculus class: It might actually be the case that even if you get accepted at say Chicago, that you will not be able to finish the courses there, because of your lack of multivariate Calculus knowledge! There is a reason the departments list this as a prerequisite! So my personal suggestion would be: Even if the calculus sequence at your department is harder than usual (which is actually good!), you should definitely take it. I see you have a very full course load, could it be possible to take less credits or do you actually need so many? Also if you think it will be too much with Calculus, you might consider replacing some of the classes you suggested for easier ones. But don't skip Calculus.

It seems like Chicago & Harvard Stats have also started sending out acceptances.

First of all I think you have a very strong application and already a lot more experience in statistical courses than most applicants! Now, it is hard to say without knowing what the syllabus for these courses is and at what universities you want to apply to. I definitely think you should replace data mining by numerical analysis. Knowing your matrix factorizations and basics about numeric stability is really important for implementing most statistical methods and demonstrates your aptitude for taking M.Sc. courses to a higher degree than data mining, which can often be very dumbed down (again depending on the department teaching it). You also already have more than enough statistics courses in your transcripts, so that actually I would also choose Real Analysis over nonparametrics. At least for the PhD programs, it is really important to have a solid background in Real Analysis. (You also have an Intro to Analysis, but still taking even more real analysis will strengthen your application. You should know the material in Rudin's Principles of Mathematical Analysis.) On another note, do you mean "multivariate calculus" when you write "multivariate"? You really need to have that! In fact, I'd try to take something as multivariate real analysis or advanced linear algebra or basic functional analysis too, if it is offered and drop another stats course. While reading through your course list, the main weakness of it seems to be that you might have only worked in 1-dimension so far (except your linear algebra course), while for a stats M.Sc. you should be comfortable working in higher dimensions. Departments probably also expect that you have taken a Calculus 1-3 sequence, so you need to demonstrate your knowledge of these subjects too (right now you only have Calculus 1 if I see it correctly).

It seems to be the general trend this year, that replies are more delayed than in the past (see also pretty much all active threads). Therefore I would not worry about it yet!

Being an applicant myself right now, I do not think I am in a position to evaluate your profile. But if you are really interested in Bayesian Statistics, I would recommend also looking at UK universities. My feeling is that most US statistics departments (with some exceptions, notably Columbia, maybe Berkeley because of Michael Jordan) are not particularly strong in Bayesian stats, while the opposite can be said about some UK departments. This is even more so the case if you are doing Bayesian Stats in machine learning with people like David Barber, Zoubin Ghahramani and David MacKay being there.

@brewing16 ah thanks, good to know! I was afraid something had gone wrong with my application :P.

@brewing16 a stupid question: I received an official admission letter but when I log into the application website, the status of my application is still "submitted" rather than "accepted" or "admitted". Is it the same for you?

Thank you!! Indeed I applied to biostats and not stats. Good luck with your results! Thanks!! Congrats to you as well!!

Yes it has been really late this year compared to previous years!! But I received an e-mail informing me of getting admitted like an hour ago (I am so ecstatic right now!) Also looking at the Gradcafe survey, I am not the only one!

The forum has been surprisingly quiet, which scares me a bit :). Last year, a thread, which accomodated all the applicants waiting for their decisions and constantly refreshing their e-mail, was already active. This thread serves the same purpose. Last year UW biostats had already sent out results for many people. I am getting impatient :(.