Possible Ph.D. fields besides Statistics

[richardw]

Hi,

 

I started this post as a personal advice seeker but then I realized this may become a general discussion. So I removed my personal information and kept the question the same: What other Ph.D. fields/programs are out there for statistics background students beside a statistics Ph.D.? 

 

PS: This is not a question for those who want to go to academia and have certain research interests in mind before going into Ph.D. I know this is strange because Ph.D. programs are generally meant to prepare people for the academic, but the situation is that there are way more Ph.D. than it's needed for academic, and the industry do require Ph.D. for certain positions sometimes. This is intended for people like me who came from a statistics background but feel lost and confused at all those schools/programs because there are so many possibilities with statistics.

 

1. Traditionally, statistics Ph.D. programs are quite theoretical. The top departments that offer statistics Ph.D. programs usually require students to have a strong mathematical background, to have experience with real analysis, and GRE math subject test is "strongly recommended but not required".
2. There are some existing subfields of statistics that have their own programs, like biostatistics and actuarial science (no Ph.D. programs, some master's maybe). I have friends in both disciplines, and what they have in common is that they all have a clear picture about their future. Biostatistics Ph.D. = Pharmaceutical Company. Actuarial Science = Insurance Companies or Big 4 accounting companies. (Again, besides the academic option)
3. The Recent heat of "data science" brought machine learning and statistics closer. On job markets, these fields are always listed together under "requirements for data scientist". More departments are offering data science programs (mostly master's). CMU had the joint ML and STAT Ph.D. for years.
4. I have friends come from econometrics, finance, marketing, and psychology Ph.D. programs. They usually need to get data, through survey or simulation, and build models to solve problems in their respective fields. The skill set is usually similar as an applied statistician. However, I heard that most business school Ph.D. went to academic.

 

In terms of schools/departments, from the listed fields above, there are the usual statistics under arts and science, business schools, engineering schools, public health, and social science. In statistics, some of them do traditional theoretical research and some collaborate with other departments all the time.

Here's my question:
If you have a BS or MS in statistics and you are doing or have done Ph.D. in a field beside statistics but use a lot of statistics knowledge, could you please discuss:
1) what are the pros and cons to "change major"? 
2) what other knowledge/skills are required for stat majors to be a better fit at your field? 
3) If possible, could you talk about what departments/industries people usually go after getting their Ph.D. in that field?

I hope people can benefit from this discussion. I'll come back and provide my own experience later.  Thanks.

 

 

[ace589]

I am not a statistics major.  But I do have relative information.

There are PhDs in actuarial science.  Usually called Risk and insurance manangement.  Although acturial science graduate programs aren't common.

Phd in Economics is common path taken by statistics or other math majors even more so than ugrad econ majors because grad level econ is almost all math where ugrad is not.

Phd Industrial Engineering or Operations Research or Management Science type programs are also common paths taken from math and statistics majors who don't want to go into pure math research.  They are often considered branches of applied mathematics but more applied and may be more marketable in industry despite being similar.  Courses in non-math areas are included in human factors, manufacturing, science, engineering, and business.

[Edotdl]

Depending on your background/interests, Stats to ML/CS is probably not that uncommon.