Frequently Asked Questions for Prospective Students

SGSA President Emeritus Victor Panaretos provides the following list of questions frequently asked by prospective students and also his own answers to them.

Program, application process

How difficult is it to be admitted to the Ph.D. program?

Well, let’s just say it’s most certainly not easy. As with any top school, there is strong competition, and there will be many brilliant students competing for a spot. In a typical year there could be anything from 100-200 applications, and 10-15 offers are made. It would generally be wise to apply to other schools also.

How important is the statement of purpose?

I think it is definitely important. It’s where you give a taste of your personality. Most of your application is numbers and ratings. And believe me, most of the people that apply here will be as good. This is where you have the opportunity to stand out.

Ani writes: "We're the specialists at getting information out of numbers!" Your statement of purpose is important, but so are your transcript, rec letters etc. --ed.

How important are the GRE results?

I would think that the GRE results are a necessary but not sufficient condition for admittance. A top grade is expected in maths, and a good grade in the verbal test, although emphasis would probably be given to the former rather than the latter.

I’m not sure about which area I would like to specialize in. Is that a disadvantage when I apply?

On the contrary, you are not expected to know precisely what you want to do. The first year involves some coursework that will get you to see different areas (along with seminars and conversations with faculty) so that you make a decision sometime in your second year. In fact, Berkeley’s faculty has quite diverse research interests, so that if you’re not sure what you want to do, Berkeley offers many choices.

How challenging the program is compared to others?

Well, I haven't been through any of the other programs, but my colleagues and myself have found it challenging. I guess that this would also be true for the other top class programs -that's what makes them top-class programs in the first place!

A comment on the website from a former student said that when he started, he discovered that of 20 incoming students, only he and one other did not already have a masters in stats, and he spent 2 years just trying to keep his head above water. Is it really that tough?

Well, I did not hold a masters degree when I arrived here. The first year is indeed stressful, as there are coursework requirements that take up practically all of your time. It used to be especially stressful when the so-called preliminary exams existed - I was part of the last group of graduate students to take them. But these were done away with as of this year (2004-2005), so I think that life should feel better for the first year students now. I think that the difficult time is during the first year (at least that's how it was with me). It always helps if one has a statistics background, but if one does not, they can make it – I suppose this should be true for all graduate statistics programs.

That comment was on the old website. Don't trust anything on the old website; it's old. (You'll know something is from the old website if it hurts your eyes to read it.) Nowadays, most incoming students don't have a Masters; many don't have any degree in statistics. --ed.

How interdisciplinary is the program?

The program is very strong in this respect. It has very close ties with computer science, biology (genetics, biostatistics, bioinformatics) and communications (and neuroscience and finance --ed.) In fact there exist "designated emphasis" programs for bioinformatics and communications. In addition, collaboration with scientists from other fields comes individually, through your advisor. For example, I am working in a problem arising from biophysics. I know that people work on topics from various different disciplines, including the above, but also the social sciences (e.g. political science, law). This is a very strong point for Berkeley: almost all of its graduate programs are ranked among the top ten in the country, so that one really works with the cream when collaborating with other departments.

Is Berkeley more on the theoretical side or on the applied side?

As you might probably know, Berkeley has been a powerhouse of mathematical/theoretical statistics and has a very strong tradition there. However it is not the case that Berkeley is "all" theory: Berkeley does “everything”, but expects you to know your theory well. As my advisor (David Brillinger) says: theory becomes practice.

How many years does it take to complete the Ph.D. program?

In theory the course is a 4-year course, although most students take 5 years. This will certainly depend on how soon one decides their preferred area for specialization and accordingly choose an advisor. It is certainly realistic to aim for four years, and in certain cases some students have completed the program in 3 years.

Victor is finishing in four years. I'm jealous. --ed.

Is it true that Berkeley Statistics is one of the best in the US?

In the last National Research Council assessment exercise, Berkeley Statistics was ranked first in the US (tied with Stanford). It is one of the most historic departments in the country with a long tradition of excellence, and considered an elite center for probability and statistics. Of course it has some special characteristics, placing considerable emphasis on probability theory and theoretical statistics, consequently expecting a strong mathematical background from prospective students. Finally, Berkeley is probably not the place if one is interested in a strictly “Bayesian education”.

Which is not to say that some people, e.g. Mike Jordan, don't do excellent Bayesian stuff. Also, my Math background was rubbish, and I'm still here. --ed.

Could you share with me observations on the culture of the program, the interaction and relations between faculty and students, and the level of support from faculty?

The culture of the program is, I think, that you basically get a firm background during the first one to two years (during which the program is more structured) and then you get into research and things become much looser. The faculty interacts with the students through seminars and classes (this is the "organized interaction"), although I have found the "unorganized interaction" to be much more important: whatever door you knock, the professor will always be willing to hear about your problem - whatever that may be, personal or scientific- and help as much as they can. This is especially important when you do research, since you have access to people with a very wide variety of research interests.

I have an offer from Berkeley and from other places too. How can I decide where to go?

I think its best to visit the different places you have offers from. This will give you a feeling of the atmosphere there. My visit to Berkeley was crucial for my decision to join the program.

Faculty

What are the faculty research interests? How likely it is that I will find something interesting to do research wise, if I come to Berkeley?

An important thing about the department is the wide range of faculty research interests. This is naturally reflected in the students, who work on a wide spectrum of topics, ranging from probability as pure math to computational statistics and machine learning. There is also a lot of interdisciplinary activity going on (I am an example of it) especially in the areas of biology (here I include biostatistics, statistical genetics, bioinformatics) and electrical engineering/computer science. I think the department is trying to promote collaboration with people form other departments, and it is actually possible to be co-advised from someone in another department.

How much help do you receive from the faculty for your Ph.D.?

My impression is that the atmosphere in the department is very informal and quite friendly. From my experience all faculty members are happy to discuss with graduate students and give advice. As for how much help your advisor provides, that depends on the individual. There are professors that will monitor your work closely, and others that will expect you to be more independent. Whatever your preference may be, I expect that there are plenty of people on both sides!

Students

Do all Ph.D. students get funding?

Yes, any student admitted to the Ph.D. program receives funding at a level comparable to the rest of the top ranked programs. Students typically do a combination of teaching and research assistantships in return.

How “close” are the students among themselves?

There is no general rule here, in that it depends on the “year”! Certain times the incoming students where very social, other times not so much. The students are not very coherent as a group (since there’s around 50 of us it could not be otherwise), and you will find that they are connected usually more with people from their own year (maybe plus or minus one) or with people with similar research interests. This has also been my experience.

Actually the closest among us are the Wii players. --ed.

Life in Berkeley

How is life in Berkeley?

I find it very pleasant. People are easy-going, it is relatively safe, and the weather is mild throughout the year. San Francisco is only half an hour away (and you get a free bus ride back and forth with your student ID!!).

Take the BART, it's quicker. --ed.

Where do you see a (major) difference between UK Masters and Ph.D.s in Mathematics and/or Statistics and US ones?

I don't know much about the masters programs, because I did not do one myself. I should think that the differences are minor academically, although you should note that the masters degree in the US usually is a year and a half to two years long (while in the UK it's a one-year program). The Ph.D. programs are quite different though. In the UK you start straight away doing research, usually having selected an advisor when applying. The emphasis is completely on research. This is excellent when you know what you want to do, and have a clear idea about whom you want to work with. In the US there usually is a first year of intense graduate coursework to build a stronger background. At the end of that year there is usually an exam you have to pass, and after that you start looking for an advisor (at Berkeley they did away with this exam in 2005, and instead progress is monitored through the coursework). Sometimes you have to take courses even in your second year (although in a much more relaxed way). The first year tends to be very intense. One thing to take in mind is that in the UK it is harder to go straight into a Ph.D. program without a masters degree, although in the US this is more common.

Is there a different attitude towards statistics in the US than elsewhere?

My impression is that in general in the US, statistics departments tend to be more mathematical. In several places statistics is considered part of mathematics, and so there is quite some emphasis placed on the mathematics of it.

In many, though far from all, statistics departments outside the US, the math is less rigorous so that other aspects of the subject can be focused on. This generally doesn't happen in the US, and it especially doesn't happen in Berkeley. --ed.

Is it harder to accept international students for the Masters program than the Ph.D. program?

To be frank I don’t know that much about the masters program. However this opinion may have some merit, in the sense that the course Ph.D. in the US involves a lot of coursework anyway, so that a masters degree may have a more professional profile to it, hence larger demand from within the US.

Ani writes: "For foreign students it's particularly hard to be admitted, whether for Masters or for PhD^?. It's not true that there's more US demand for the Master's degree. Most of the applications are non-US. The acceptance rate is about 1 in 10 for both programs."

Is it easier to be accepted for a Ph.D. or for a Masters?

There is no general rule. Obviously this depends on the applicant. What I do know is that at some point in the application for the Ph.D. degree, you can state that you would like to be considered for the M.A. in case they cannot offer you a place in the Ph.D. program.