Should I apply to statistics or biostatistics PhD programs?

[zestar07]

Hi,

I am applying to stats/biostats PhD programs, and I want to go into industry after. Sorry if I'm beating a dead horse, but should I apply to stats or biostats programs? I'm thinking about the following:

1. Job prospects. Is biostats more limited than stats? How do employers view a PhD in stats vs. biostats?

2. Will research topics be more limited to public health and medicine if I go into biostats? I'm primarily interested in biostats right now, but I don't want to restrict myself in case my interests change.

 

 

 

 

[Stat Assistant Professor]

No, I think you are fine with either stat or biostat. There are plenty of Biostatistics PhD graduates who work at tech companies like Amazon, Facebook, Google, Microsoft, etc. (conversely, there are also a lot of Statistics PhD grads who work in big pharma and government agencies like FDA, etc.). If you want to get a job like that, you really just need to perform well in their technical interviews and challenges (which often includes brainteasers and probability/statistics questions at the advanced undergrad level).

A PhD is a long commitment though, so you might want to consider what kind of research you would enjoy doing more, and if prospective programs have enough faculty working in your areas of interest.

[StatsG0d]

1. I would say no. I'm in biostats, and I have gotten offers / interviews on Wall Street, government, postdocs, etc.
    
2. I think many applicants (and indeed many statistics PhD students) assume that biostatistics departments only do applied work. This is true for some biostatistics departments, but certainly not all. Biostatistics departments in the top-10 are very engaged in methodological work. IMO, the fundamental difference between statistics and biostatistics programs is that in biostatistics the problems are motivated by medicine / public health and in statistics they are motivated by that as well as other things.

Here are some "hot" research areas in statistics where a biostatistical application exists:

• Spatial-temporal: medical imaging (e.g., finding tumors)
• Bayesian: Bayesian clinical trials (e.g., interim analyses)
• Machine learning: precision medicine (e.g., finding the treatment at the right time for the right individual)
• Computational statistics: genetics
• Dimension reduction: genetics
• Causal inference: any observational public health study

The only real constraint with biostatistics is that your methods must be relevant to some kind of public health or medical setting, broadly speaking (e.g., basically anything dimension reduction related will be relevant to genetics).