Preparing for PHD

#1
Hello All,

I just wanted to gather a few opinions on pursuing a PHD in Statistics given my background. I would like to begin a PHD program in maybe 3 or 4 years so I wanted to see what type of preparation is necessary.

I have an undergrad degree in Business and an MS in Technological Entrepreneurship.
The only math related courses I have taken are Quantitative Analysis for Business 1 and 2 and Business Statistics. I also conducted a research paper for my honor’s thesis in which I collected data and used multivariate hypothesis testing to find correlations between 10+ variables.

To be eligible for a PHD program I was thinking I would take a few grad courses in stats as a non matriculating student and do well on my GRE.

Just wondering if this would be enough or do I have no shot since I do not have a purely mathematical background.

Any and all help is greatly appreciated.
 

noetsi

Fortran must die
#2
There won't be one answer to that question, because different universities have their own requirements and they differ a lot potentially. It might be useful to consider some of the universities you are thinking of attending and ask them what their requirements are.
 

Mean Joe

TS Contributor
#3
I just wanted to gather a few opinions on pursuing a PHD in Statistics given my background. I would like to begin a PHD program in maybe 3 or 4 years so I wanted to see what type of preparation is necessary.

Just wondering if this would be enough or do I have no shot since I do not have a purely mathematical background.
You seem to think that a special kind of person is required to get a PhD in statistics. Well, maybe. But I'd say you can do it.

What would you think about a person with "the perfect background for a statistics PhD" going into a field like electrical engineering? Wouldn't that be possible, or would that be a waste? People can do what they want to do.

3 years off to take some grad courses to prepare you for PhD?

I'll start by saying that the graduate level math courses I took covered a lot of the same stuff as the undergrad math courses. But in the grad level they expect you to have a certain background (eg be familiar with proof by contradiction, some basic number theory such as modular arithmetic, etc). So I think you maybe should start with undergrad math courses to prepare you. They'll be easier in the sense that they require less of you, but you will still be learning essential stuff.
(My apologies, I thought you were going to focus on math-related courses. But looking back now, you just mention taking grad level stats. Anyway, the general idea still holds I think; you can get a good background by taking the undergrad courses.)
 

noetsi

Fortran must die
#4
You seem to think that a special kind of person is required to get a PhD in statistics. Well, maybe. But I'd say you can do it.
I think you are being too humble there Mean Joe. As some who struggled through a (much simpler) master's in measurement and stats I think its fair to say you have to be pretty decent in math to get a PhD in stats :p

My own PhD was in a very different area, but I want to point out something that I believe is generically true of that degree. Which is, it takes a long time, wears you down, and by no means is a guarantee that you will receive a position in that field (nor does it generate signficant income advantages assuming that is something one cares about). Maybe stats is different in both regards of course than other fields.

So if someone is interested in a PHD don't underestimate just how much work (and pain for most) is involved in it. And don't have pie in the sky assumptions about what it results in.
 
#5
Thanks for the responses. The two programs I’ve been looking at so far are UC Irvine and UC San Diego because I live in the area. My end-goal is to become a data scientist or statistical consultant.

I know most people pursue a PHD for academia or research but I would like to go into the private sector since there is quite a bit of demand for people with a business background that possess strong analytic capabilities. (For this reason I'm trying to find a program that emphasizes applied statistics and is less heavy on theory)

So far I have been trying to learn the tools of the trade. I’m currently taking the following online courses for SAS:

• SAS(R) Enterprise Guide(R) 1: Querying and Reporting (EG 4.3)
• SAS(R) Enterprise Guide(R) 1: Querying and Reporting (EG 5.1)
• SAS(R) Enterprise Guide(R): ANOVA, Regression, and Logistic Regression
• SAS(R) Programming 1: Essentials
• SAS(R) Programming 2: Data Manipulation Techniques

My goal is to at least be SAS Base Programmer certified by the end of the year. When that is complete I will focus on learning R & SPSS (and continue with SAS since it is widely used in the private sector).

When I comfortable enough with my skills I was thinking of approaching some PHD candidates in my area and volunteering to help them with their projects. They will obviously know a great deal more than me; however, I figure there will be some low-level non-administrative stuff I can help with. Not sure how feasible this is but just something I was kicking around.
 

noetsi

Fortran must die
#6
Thanks for the responses. The two programs I’ve been looking at so far are UC Irvine and UC San Diego because I live in the area. My end-goal is to become a data scientist or statistical consultant.
I am a data analyst and it's important to stress that academic research (which is what a PHD is) commonly is very different from what is stressed outside academics. There are skill sets, such as practical data manipulation - aka SQL, preparing reports for non-technical audiences, and networking that are actually critical in that environment that receive little if any attention in academia (at least from my experience but I spent a long time at universities).

I think the courses you are taking on line are great to be an analyst -I am not sure how much you will use them in academic statistical programs (that appear to lean more to R these days). But they are useful to know regardless.

If you go into a phd with the idea of working outside academics its very useful to make contacts before you graduate and look at the type of work actually done in real world organizations related to statistics. I think you will find they differ a lot from academic research.

It's fair to say many here won't agree with this sentiment.:)
 
#7
I am a data analyst and it's important to stress that academic research (which is what a PHD is) commonly is very different from what is stressed outside academics. There are skill sets, such as practical data manipulation - aka SQL, preparing reports for non-technical audiences, and networking that are actually critical in that environment that receive little if any attention in academia (at least from my experience but I spent a long time at universities).

I think the courses you are taking on line are great to be an analyst -I am not sure how much you will use them in academic statistical programs (that appear to lean more to R these days). But they are useful to know regardless.

If you go into a phd with the idea of working outside academics its very useful to make contacts before you graduate and look at the type of work actually done in real world organizations related to statistics. I think you will find they differ a lot from academic research.

It's fair to say many here won't agree with this sentiment.:)
Thanks again for your input. As far as academic research I want to focus more on data mining large data sets. As a result, the PHD program at UC San Diego is especially enticing as the San Diego Super Computer Center is located there and has some interesting stuff going on involving data mining/analysis.

I’m sure I will have a clearer picture of exactly what I want as I get closer to my anticipated start date. I’m still in the very preliminary stages of planning for a PHD and a lot of things can change in 3 or 4 years.

I will certainly make every effort to network and garner a better Idea of what the private industry looks for as far as data analysis goes.

May I ask where you obtained your PHD and what you do in your current role as a data analyst(day to day)? Also, what tools do you use most often?
 

Mean Joe

TS Contributor
#8
I think you are being too humble there Mean Joe. As some who struggled through a (much simpler) master's in measurement and stats I think its fair to say you have to be pretty decent in math to get a PhD in stats :p
Haha the thought crossed my mind that I should edit my post; I was not being clear originally. What I should have said is that you may not have a good background right now for a PhD in statistics, but you can get it. It will take work, but it's learnable.

PS isn't measurement the stuff you learn in first week of Physics? Just kidding. But my friend used to make fun of Chapter 1 of pretty much any College Physics book.


"A whole chapter on measurement, just like for special relativity. I'll specialize in that."
 

noetsi

Fortran must die
#9
lol

In my program, which was in the College of Education, "measurement" meant equating, IRT and the like. How you actually measured latent abilities you could not see primarily. Which in honesty I found far more difficult than statistics.

As someone who keeps dropping out halfway through a matrix algebra text (about to start again) I only wish math was learnable for everyone if you tried. My point is that many people simply don't have the capacity to learn advanced math and for them theoretical stats is probably not the way to go.