(Note: This is the first post in a series about graduate life in statistics. For links to all articles in the series, click here).
Here are some key points to consider when choosing graduate programs in statistics and biostatistics.
1. What’s the difference between biostatistics & statistics?
When I first applied for masters programs in statistics, I had little to no idea what biostatistics was. To the untrained eye – in my case, a liberal arts undergraduate student – the subject biostatistics gave off a connotation aligned with phylums and petri dishes, things I had been hoping to avoid since roughly 10th grade.
Ironically, however, biostatistics is not the intersection of statistics and biology; instead, biostatistics is mostly just statistics applied to fields within or related to public health. In four years of a biostatistics program, for example, I didn’t take a single biology course, and neither did the majority of my classmates.
However, there are definitely differences between the two fields, and this distinction manifests itself in a few ways. First, on average, graduate students in biostatistics might be expected to interpret, collaborate, and write more often than those in traditional statistics programs. At Brown, for example, all biostatistics students were required to attend a weekly ‘journal club’ meeting, in which we met with students from epidemiology and health services regarding recent advancements in public health.
A statistics program, meanwhile, might be a bit more theoretical than one in biostatistics. At UMass, for example, PhD students in statistics were required to pass a competency exam in advanced calculus, while Brown’s biostatistics qualifying exams included the more applied topics of survival analysis, longitudinal analysis, and multivariate analysis. While biostatistics students might face some restrictions on the types of data they can use for a dissertation, statistics students would have opportunities for projects beyond the scope of public health – for example, sports or agriculture – that biostatistics students might not receive.
Each field is growing, and growing rapidly, and many students could be successful by taking either track. In large part, your decision should come down to (i) what type of research you want to do and (ii) if working in public health sounds like a neat thing to do. For me, my answer to (i) was applied, and my answer to (ii) was a resounding yes, which is why I’m happy that I ended up in a biostatistics program.
2- Thinking PhD? Have a reason why
One of the most important things that I have picked up on over the last several years is that the rate of attrition of students going directly from an undergraduate program to a PhD track is noticeably higher than their peers who came in with either a masters or real world experience. A PhD program in statistics might be the natural extension of an undergraduate major in mathematics or statistics, but the differences between the lifestyle of a PhD student and an undergrad are severe.
Doctoral programs, on the whole, consist of a few tricky stages. For starters, qualifying exams are entirely different than any undergraduate final. They are longer, more in depth, and take, in most instances, months and months of studying. This, in retrospect, is slightly ironic, as the students taking PhD qualifying exams, on average, are actually the types of students who would have nailed most undergraduate courses without studying!
Once a PhD student passes his or her qualifying exams, by and large, there are no classes to take and no exams to study for. So on one hand, you have hours of free time to spend studying the subject that you love, but on the other hand, that level of independence makes it easier for students to get lost.
For students fresh out of undergraduate programs, these types of adjustments may be a bit more difficult. In my case, it took me six years to figure out the best field and program for me – had I started Brown’s program at age 22, there is very little chance that I would have made it through unscathed. If you are still an undergraduate, my advice is to hold off on graduate school until you know its the best decision for your future.
3- The pluses and minuses of masters programs in statistics
A popular route to start a career in statistics is to do a masters program. One benefit of the masters track is that there tends to be more diversity with respect to topics and school. Further, most masters degrees can be completed in two years or less, and some schools will allow for more flexibility in their masters programs than their doctorate ones, most of which involve a full time commitment. An additional benefit of masters programs is that they allow students who are unsure about their future in statistics a few years to get their feet wet in advanced coursework, without the long-term commitment.
However, while the majority of doctorate programs in statistics or biostatistics will provide several years of guaranteed funding, funding for masters students varies substantially. Some programs will offer full tuition coverage and a small stipend, while others will cost upwards of $50,000 per year. It is not uncommon for a masters student taking the class for free to be sitting next to another masters student who is paying several thousand dollars a course to be in the identical spot. As a result, within several programs that offer research assistantships (RAs) to masters students, competition for funding is fierce. Masters students, be prepared to fight for funding, and don’t be surprised to get frustrated from all of this, and to end up paying lots of money for your advanced degree.
One alternative option is to attend masters programs that guarantee funding and small stipends through teaching assistantships (TAs). This is more frequent at state schools – for example, in New England, UMass funds masters students that help teach their undergraduate courses in statistics – and comes with a reasonable weekly teaching requirement. For some students, teaching statistics might be a turn-off, but I found that it was a great opportunity to work with faculty and to improve my communication skills. Further, in my opinion, there’s no reason to pay $50,000 a year to get roughly the same education (with the same book, even!) as one where you can get paid a small amount to attend.
4- So you’ve picked a few programs that you like. What should you look for?
There are a few things to look for as you sort out the pluses and minuses of statistics and biostatistics programs.
a) Attrition vs. retention
In one well-regarded state school program, rumor has it that half of the approximately fifty students taking the PhD exams fail, and eventually drop out with a masters degree. If you are applying for PhD programs, don’t wait and find this type of information out after you enroll. This seems like common sense, but if you want a PhD in statistics, find a program that tends to keep the majority of students that enrolled while also wanting to receive a PhD in statistics.
A program’s retention rate can be an important question to ask of all schools, both big and small, and can tell you alot about (i) the faculty’s commitment to working with students, (ii) the difficulty of the school’s qualifying exams, and (iii) other interesting opportunities at the school that keep students integrated. Also, ask the rate to both faculty and students that you meet with.
PS – if a school tells you its retention rate is 100%, it’s lying.
b) Past student successes
There are several opportunities for graduate students to get recognized for their research or teaching. For example, JSM and ENAR, two of the major conferences in statistics and biostatistics, each host yearly student paper competitions. And for American students, the National Institute of Health is one of several agencies that funds pre-doctoral students.
Don’t these sound like nice things to have on a CV? Of course! As a result, it is worthwhile to use google or a few emails to find out if students at the program you are interested have earned these types of awards. Also, don’t punish the small schools if they don’t appear to often on these lists – they have a smaller sample size of students!
While these types of honors are not the only indicator of a strong program, they do speak to both the faculty’s interest in their students and the level to which a critical mass drives each program’s research agenda.
Next, in Part II, I’ll explore “How to thrive in a graduate program in statistics”