Harvard biostatistics

[rosebud123]

I have heard a few people claim casually that Harvard’s biostatistics PhD program is very “applied.” Could anyone elaborate what this might mean or whether this is in fact a general consensus about the program?

[abstract_art]

When people talk about programs being applied/theoretical they're usually talking about one of two things: coursework/quals or the research being done by professors in the program. I'm guessing when most people say Harvard is more applied they're talking about the coursework/quals. Harvard is a lot lighter, in this regard, than other top biostat programs where measure theory/more theoretical coursework is required (e.g. UW, Hopkins, UNC). However, that doesn't mean there aren't professors there doing theoretical research (off the top of my head there are a few bayesians who do pretty theoretical work there), just like the programs with more theoretical coursework/quals have professors who do applied work.

[rosebud123]

Is it fair to say that when people label certain work in (bio)statistics "theoretical," they mean it roughly as a euphemism for "involving advanced mathematics"? For instance, I feel like I see highly abstract work in, say, models of causal inference referred to as "applied" simply because the symbols on the page aren't terribly imposing

[cyberwulf]

10 hours ago, rosebud123 said:

Is it fair to say that when people label certain work in (bio)statistics "theoretical," they mean it roughly as a euphemism for "involving advanced mathematics"? For instance, I feel like I see highly abstract work in, say, models of causal inference referred to as "applied" simply because the symbols on the page aren't terribly imposing

Yes. Of course, within each sub-area of biostatistics there is a range of mathematical sophistication. For example, within causal inference there is plenty of heavy mathematical lifting in the theory of semiparametric efficiency for doubly robust estimators.