How can I prepare for a masters/PhD in statistics?

[Isusriourt]

Undergrad Institution: Ivy
Major(s): Cognitive Science
GPA: 3.77

Type of Student: Domestic Male

GRE General Test:
Q: 170 (97%)
V: 170 (99%)
W: 5.0 (93%)
 

Programs Applying: Masters in Statistics or Math - I want to get a PhD in Statistics, but I doubt I have the math background for that, and ideally I'd like to get math/stats research experience before I apply for a PhD if at all possible.

 

Research Experience: An undergraduate research thesis in cognitive science, not published anywhere
Awards/Honors/Recognitions: Distinction in the major, Psi Chi honor society in Psychology
Pertinent Activities or Jobs: None

Letters of Recommendation: My cognitive science thesis advisor, a computer science professor, a psychology professor

Math/Statistics Grades:  Calculus Functions of Several Variables: A, Game Theory: A, Design and Analysis of Algorithms: A-, Mathematical Tools for Computer Science (light linear algebra): A
Any Miscellaneous Points that Might Help: I've taken a number of 400-level computer science classes. I've worked through the Coursera course on Machine Learning from Stanford/Andrew Ng, which involves some linear algebra.

I've been a software engineer since completing my undergraduate degree, barring a medical leave of absence (coming up on two years back at work within several months). I became interested in statistics through working closely with statistics researchers at one of my jobs, and I'd like to go back to school to learn how to do research in statistics. However, I don't believe I have enough of a background in math, so I'm trying to study math in my free time and apply to masters programs in fall 2019 for a fall 2020 start date.

I'm currently working my way through a linear algebra textbook, and once I'm done I plan to move on to a more advanced linear algebra textbook and/or real analysis. Do I have enough of a calculus background, or do I also need to study differential equations? Is this all overkill, or would it not be enough to prepare for applying to masters programs in math/statistics? Should I consider a postbacc instead?

[omicrontrabb]

You mentioned you want to get statistics research experience before applying? Is that so you have a better idea of what a PhD in statistics would be like? Or to improve your application? It would be helpful to see what a research statistician does, but isn’t necessary for the application.

You are right, your background is probably too light on math courses to get into a top statistics PhD program. But otherwise you have a stellar application, good grades at an Ivy, extensive comp sci experience, GRE scores. There are not many statistics postbacc programs. Few people have done meaningful statistics research before applying to PhD programs. I would take linear algebra, real analysis and a probability class at a reputable university. I assume you’ll do well, which will satisfy the prerequisites and show you can handle advanced math. Alternatively, if you’re highly motivated, you could self-study math and take the math subject GRE. Those two options would save you a lot of time/money compared to getting a masters.

Lastly, depending on what exactly you want to research, you could also consider CS PhD programs. I’m not personally familiar with their admissions processes, but you do seem to have a strong background in CS. Some topics, like machine learning, etc, are researched by both CS and stats PhD students.

[Isusriourt]

Thanks, omicrontrabb! I mainly want to get statistics research experience before applying so I have a better idea of what a PhD would be like, rather than to strengthen my application.

If I want to keep working until I go to grad school and can't find any local options for linear algebra/real analysis courses on nights or weekends, are there any online programs that would be sufficiently reputable? I'm not yet sure about my ability to be prepared for the math subject GRE by the last 2019 testing date, given that I've never studied differential equations and haven't seriously tested my calc knowledge in years.

[bayessays]

With the ivy background, high GPA, and perfect GRE, you're obviously in great shape.  Obviously you're missing linear algebra - you should take this. Look for a  cheap state college that offers a course in it (ASU and other schools have an online one if there are no local options, I think).  I'm not sure what your financial situation is, but if you take linear algebra, you'd get into some decent PhD programs right now (you might even get into some if you just convince them you know some linear algebra).  Then you can get a definitely funded master's in route to PhD sooner, and cut your total time down if you do decide to do the PhD.   You could get into some top 5 biostatistics PhDs if you just finish the math pre-requisites - the first two years of all these programs is basically just getting the master's degree, and a lot of the students end up leaving with master's degree anyways.

[penguinbombs]

Just wanted to chime in that I had a similar academic profile and trajectory (non math major from an Ivy, took a bunch of non-degree math courses at a state school) and I had good results this year. I agree that with your computing skills, GRE scores, and more unique motivation for entering statistics, you'll probably be a very attractive candidate once you get Linear Algebra and Real Analysis under your belt. 

For online courses, check out the Harvard extension school. I don't say this because of the Harvard name (I don't think that carries much weight as an online extension program), but just because I remember them having tons of classes, so maybe you'll find what you need there. Also it seems like University of Illinois offers a lot of online math classes. https://netmath.illinois.edu/academics/netmath-courses-college-students. I'm sure ASU as mentioned earlier is a good option as well.

I ended up changing my work schedule and finding different jobs that allowed me to go to school in the day, so I didn't actually use any of these options, but give them a look, maybe they'll work for you.

(Oh and for many masters, you're probably a strong candidate already)

(Also about the math subject test, not to be a downer, but it's probably unrealistic to expect to get a good score given your preparation. It's extremely difficult covering a wide range of mathematics. I didn't submit my scores, but I was admitted to several schools that "highly recommended" the test, so probably isn't worth worrying about.)

Good luck!

[Bayesian1701]

5 hours ago, bayessays said:

With the ivy background, high GPA, and perfect GRE, you're obviously in great shape.  Obviously you're missing linear algebra - you should take this. Look for a  cheap state college that offers a course in it (ASU and other schools have an online one if there are no local options, I think).  I'm not sure what your financial situation is, but if you take linear algebra, you'd get into some decent PhD programs right now (you might even get into some if you just convince them you know some linear algebra).  Then you can get a definitely funded master's in route to PhD sooner, and cut your total time down if you do decide to do the PhD.   You could get into some top 5 biostatistics PhDs if you just finish the math pre-requisites - the first two years of all these programs is basically just getting the master's degree, and a lot of the students end up leaving with master's degree anyways.

Besides linear algebra wouldn't some more math like real analysis and mathematical statistics (undergraduate level) also be helpful?  Maybe not necessarily a class, but I think it would be challenging to have just the bare minimum prerequisites and not the optional "prerequisites" like real analysis and probability that would make first-year Ph.D. coursework so much easier.

[bayessays]

52 minutes ago, Bayesian1701 said:

Besides linear algebra wouldn't some more math like real analysis and mathematical statistics (undergraduate level) also be helpful?  Maybe not necessarily a class, but I think it would be challenging to have just the bare minimum prerequisites and not the optional "prerequisites" like real analysis and probability that would make first-year Ph.D. coursework so much easier.

Sure, a master's would be helpful too. More is always better - and if he wants to go to a top 10 program, he should take some more classes. But this guy is going to be able to handle the coursework, so I'm just suggesting what will help him get in the door at a decent school.