As a PhD graduate student in statistics here at UCLA, I’d like to talk a bit about our graduate program to give you an idea of what to expect if you’re interested in UCLA or statistics graduate school in general. First, it’s important to note that UCLA does not have a specific “data science” degree program at this point; the department offers a PhD and a Master’s degree, both simply labeled “statistics.” Some statistics departments have made the decision to change their name to “data science,” but that is often just a marketing decision. The UCLA department is pretty forward-thinking and has a focus on computation, so it’s fairly “data science”-y without having that exact name.

Let’s start with a general outline of graduate school: usually, the first component is coursework, and the second is writing your thesis or dissertation. For those pursuing a Master’s degree, the project is typically called a thesis, and for PhD students we call it a dissertation. These written projects differ in complexity and length, with dissertations being longer, more technical, and generally more focused on theory than applications. As a result, students chasing a PhD usually take quite a bit longer to complete their degrees than those getting a Masters.

I just completed my fourth year of graduate school and am planning to defend my dissertation next June, making it an even 5 years from the start of my graduate career. This is pretty standard length for a PhD from our department, although there’s bound to be some variation (+/- 2 years).

# Coursework

The first year of graduate school is typically focused entirely upon coursework, and UCLA is no exception. For students pursuing the PhD track, there are three required sequences of classes (i.e. 200, 201, 202) with three classes contained in each (e.g. 200A, 200B, 200C). The 200 sequence contains topics in theoretical statistics and probability and is focused mostly on proofs. The 201 sequence is a little more applied, focusing on sampling, experimental design, regression, and the like. The 202 sequence contains statistical computing, from basic R and Python to Monte Carlo methods. (Side note: UCLA uses a course numbering scheme I hadn’t encountered before, as courses in the 10s are lower-division undergraduate classes, like STAT 10, STAT 13, etc., for non-statistics majors, the 100s are upper-division undergraduate courses for statistics majors, and the 200+ courses are graduate level.)

In addition to the technical course sequences discussed above, we also have a weekly departmental seminar course, which all graduate students are required to attend (although it’s open to the public as well). Every week in the seminar is a different invited speaker, presenting their work. The speakers tend to reflect the interests of the professor who is organizing the seminar — when Mark Handcock is in charge, there are many talks on networks, for example. My first year in the program, Mark Hansen organized the seminar and we had a variety of not-traditionally-statistics talks, including both the technology head and a design technologist from Stamen design, a law professor, and a statistician who specializes in the ethics of data. The goal of the departmental seminar is to expose students to research interests that may be outside the expertise of the department, and allow them to see the variety of possible sub-fields available to them. It’s been likened to a required Meetup group for us grad students.

To complete the coursework requirements, PhD students must have 54 units, or about 13 classes. After completing the 200-series, most students have approximately 4 additional electives they must complete in order to have their required units. In my case, I chose Spatial Statistics, Data and the Media Arts, Programming Media, Computer Programming Languages and Systems, and two quarters of “Data, Data Practices, and Data Curation.” Most of my electives were related to data, computing, and data visualization/art. For those of you keeping track at home, that’s more than four courses, but I enjoyed coursework and had quite a variety of interests.

# Masters program

Thus far, I’ve only discussed the PhD program. For Master’s students, the requirements for coursework are slightly reduced (although they take the 200-level courses, too), and students typically spend only two years at UCLA (one year to complete coursework and one year to write their thesis). Once Master’s students have completed their coursework, they can move on to writing their thesis. I’m not as familiar with the Master’s program, but if you want to know more, you can check out the department’s information for applicants.

# Proving your worth

After finishing coursework, the next step in a PhD program is typically to prove that you know the material from the required courses. This is often one of the things that separates Master’s and PhD programs– Master’s students are required to take classes and write a thesis; PhD students have to take classes, prove they know the material, and then write a dissertation.

In many graduate programs in mathematics, proving you know the material requires students to complete a grueling series all-day exams covering all the content from the required courses. These are called qualifying exams (or “quals”). However, a few years before I joined the Statistics department the faculty voted to discontinue quals, even for PhD students. Instead, grading for courses was set to make the classes even more rigorous, meaning that once you’ve passed all your coursework with the required GPA you are considered qualified. This lack of quals makes the UCLA statistics program little outside the norm for math-related PhDs. However, when I was looking at graduate programs I was honestly drawn to the department by the lack of quals. It felt much less stressful for me to pass classes individually than to remember everything on one stressful, make-or-break day.

# What's next?

In the next episode of this post, I’ll talk about the process for “advancing to candidacy,” writing and defending the dissertation, and show you some of the more fun aspects of grad school. For a taste, I’ll just tell you that we tend to have fun at conferences together:

Nice post! BTW I stole you’re CV template from your website

Thanks! And that’s what the code is there for. Trey Hunner deserves most of the credit, but maybe I’ll do a blog post– “TeXing your CV– is it worth it?”

Amelia, I really enjoyed your post about your experiences in grad school!

Thank you for your information.I really enjo4 your posts.I am holding Ms in applied mathematics(numerical analysis) and my Bs was In physics(from Iran).I am working as math faculty in college right now.I am very interested to data science.I was wondering if you recommend the ucla ms or phd stat program.Do you think , it’s better for me to get the master or directly apply for phd? I alsp wants tp know about the stipend amount.Thank you

If you have a strong math background, I think that going directly for the PhD could be fine. But it depends on where your data science interests lie– if you want to continue in academia with data science, the PhD makes sense. But if you’re more interested in industry, I think a Masters is fine. For more about how I paid for grad school, see my other post– http://datascience.la/paying-for-statistics-grad-school-aka-the/ Between GSRing and TAing I’ve been able to live on my support.