Importance of program ranking for industry

[bayessays]

7 hours ago, trynagetby said:

FWIW, off a gut feeling I feel like the impact of a Biostats vs Stats would matter more as prestige of program goes down. E.g Harvard Biostats vs Harvard Stats probably doesn't make a different in industry. Duke Stats vs Duke Biostats probably matters quite a bit more. Rutgers Stats vs Rutgers Biostats is probably huge difference.

Rutgers is probably a bad example because they are the same department, just a concentration, but I think this is definitely true in general. (EDIT: Rutgers seems to have changed this recently or I was just completely mistaken.  I thought Rutgers just had a biostats concentration within their stats PhD just like UW-Madison, but now it seems it's in the school of public health...) There are stat programs in the 60s and 70s on US News that are very good, but I don't think the same is true of the lower-ranked biostatistics programs. @statsguy , I understand what you are saying, and all I'm saying is I don't think it applies to today's *top* biostat programs.  If you have a biostats PhD from Hopkins/Harvard/UW/Michigan/UNC, and probably a few schools below that like Penn/Brown/Minnesota, there's going to be a range of machine learning options to take, you'll learn a lot of theory, and everything I see suggests you will be able to get top data science jobs.  But as prestige goes down, the divergence between stat and biostat programs increases greatly.

[untzkatz]

The ranking thing makes sense, I went to a UC and its not particularly that high up. What sucks is that most of my classes were also in the stat department and we were not even taught by our own biostat faculty. But of course the industry hiring managers don’t know this. I took extra classes in supervised+unsupervised ML+time series too albeit all at undergrad level. And I found I really liked those way more.

But I guess it does seem especially at a lower ranked school you are probably better off even doing something like EE, CS, DS to do real statistics than biostat, if you don’t have the math prereqs to get into a stat program.

No doubt though Biostat is a stable career, and for the people that just value that aspect or don’t want to really keep up anymore (like say they have a family and kids to raise etc) it can be a good option. But especially in my 20s I can’t stand doing outdated things, and especially dislike regulatory work. 

Even within classical stats, people don’t seem to be fond of more rigorous methods because a lot just want you to be a robot following the FDA guidelines. I want to get out of the Biostat field asap but I am competing against all these programmer bros who may not know stats as well but can do “ML production systems”. That is one reason I wanna do a PhD so I can focus on statistical ML/DL 

 

[statsguy]

8 hours ago, bayessays said:

Rutgers is probably a bad example because they are the same department, just a concentration, but I think this is definitely true in general. (EDIT: Rutgers seems to have changed this recently or I was just completely mistaken.  I thought Rutgers just had a biostats concentration within their stats PhD just like UW-Madison, but now it seems it's in the school of public health...) There are stat programs in the 60s and 70s on US News that are very good, but I don't think the same is true of the lower-ranked biostatistics programs. @statsguy , I understand what you are saying, and all I'm saying is I don't think it applies to today's *top* biostat programs.  If you have a biostats PhD from Hopkins/Harvard/UW/Michigan/UNC, and probably a few schools below that like Penn/Brown/Minnesota, there's going to be a range of machine learning options to take, you'll learn a lot of theory, and everything I see suggests you will be able to get top data science jobs.  But as prestige goes down, the divergence between stat and biostat programs increases greatly.

I don't disagree that a Harvard Biostats PhD will be able to find a good job. My point is that if someone is dead-set on going into industry into data science, why bother with Biostats, even if it's a good department? I'd rather go to a lower ranked Stats department, do a dissertation in some sort of ML or AI or other hot topic that is computation heavy, take relevant electives (e.g. NOT things like survival analysis), do an internship (these help immensely), and go from there. 

In fact, I know someone who did a Biostat PostDoc at Hopkins, flopped on the academic market, and took a job at Google. So it's possible. But I also know of data scientists at Netflix, Amazon, who graduated from >35 Stats departments and had no issues finding great jobs.

Going the Biostat route just seems unnecessarily harder. And even if you do get a sweet DS positions, you'll probably have some catching up to do.

Edited April 27, 2021 by statsguy