Should I go for a PhD in statistics

[krislis]

Hi guys,

I am currently at junior at Cornell University, and last fall I changed majors from Astronomy to Operations Research and Information Engineering after 3 research jobs in astronomy/physics and realizing that the only part of physics I liked was the data science. I have taken Probability I & II and I am currently taking Statistical Data Mining and Stochastic Processes. I am now doing research for a professor in oceanography on measuring the change in chlorophyll at every point in the oceans. I have not been able to get really good advice on what I should do for graduate school. I want to work in applied statistics at research centers such as Pew, the UN, NASA etc. I think I'm leaning towards doing more social statistics research rather than now.. In Astronomy you had the option to work in theoretical (like string theory), more applied (like researching solar winds etc), or technological (like designing telescopes). Now I've been told that Statistics PhD programs are completely theoretical/methodological, whereas I was hoping to do a thesis on a more applied topic. So I guess my question is should I try to go for a PhD to then be able to get a higher position at a research center that requires PhD credentials so that I can finally do what I like doing? Or are there any programs that would let me do more applied statistics (that would be ideal)

I should specify that I love my statistical data mining class, and I have been enjoying all my probability and statistics classes, but I enjoy applying what I have learned to data rather than coming up with new methods.

Also I was told by a professor that getting a masters is worse than dropping out of a PhD program????

Edited March 2, 2017 by krislis

[Dora the Destroyer]

Master's is definitely still very useful. There are some programs that are focused more towards applying statistics (e.g. NC State) that you can check out. Or you can pick another field that you are interested in, and try to do more quantitative research with that field.

[abstract_art]

What's your mathematical background like?

[poopyhead]

1. Should you get a PhD?

If your interest is in research, then yes. If not, no. Don't be tempted by the fact that many job posts list PhD as "recommended"--you'll get more out of 5 years' industry experience than a PhD if your main goal isn't research. A MS might still be worthwhile (but also extremely expensive).

2. Should you get a PhD in statistics?

You're going to have to look at programs on a case-by-case basis. Two statistics departments could end up being very different in terms of what they research. Find the ones that does the kind of research that you're looking to do (deciding which programs to apply to is already not a simple task). In your case, expand the search to computer science, information science/informatics, IE/OR, and applied mathematics programs that specialize in data science, machine learning, and "big data." There are even data science doctorate degrees now (see: NYU, WPI). 

Edited March 3, 2017 by poopyhead

[krislis]

On 3/2/2017 at 6:31 PM, marmle said:

What's your mathematical background like?

I have taken Calculus II and III (B+'s) then Differential Equations (B+) and Linear Algebra (A-). I am planning on taking an analysis and/or algebra class next semester. My grades in my statistics and OR classes are better.

[StatsG0d]

12 hours ago, krislis said:

Hi guys,

I am currently at junior at Cornell University, and last fall I changed majors from Astronomy to Operations Research and Information Engineering after 3 research jobs in astronomy/physics and realizing that the only part of physics I liked was the data science. I have taken Probability I & II and I am currently taking Statistical Data Mining and Stochastic Processes. I am now doing research for a professor in oceanography on measuring the change in chlorophyll at every point in the oceans. I have not been able to get really good advice on what I should do for graduate school. I want to work in applied statistics at research centers such as Pew, the UN, NASA etc. I think I'm leaning towards doing more social statistics research rather than now.. In Astronomy you had the option to work in theoretical (like string theory), more applied (like researching solar winds etc), or technological (like designing telescopes). Now I've been told that Statistics PhD programs are completely theoretical/methodological, whereas I was hoping to do a thesis on a more applied topic. So I guess my question is should I try to go for a PhD to then be able to get a higher position at a research center that requires PhD credentials so that I can finally do what I like doing? Or are there any programs that would let me do more applied statistics (that would be ideal)

I should specify that I love my statistical data mining class, and I have been enjoying all my probability and statistics classes, but I enjoy applying what I have learned to data rather than coming up with new methods.

Also I was told by a professor that getting a masters is worse than dropping out of a PhD program????

Based on your research interests, i don't think a stats program would be a good fit. Maybe at some more "applied" programs it would work. If you're interested in social science applied statistics, it seems to me like a degree in economics (or with your current research background agricultural / resource economics) would be a better fit. You'll still need the math (analysis is always good), but you'll take classes in econometrics as well as economic theory (and a lot of social issues are explained by economic theory).