Questions about Masters Programs in Statistics

Graduate School
1

First, let me apologize because this is gonna be long. There are a lot of things I’m worried and wondering about. I come from a Chemical Engineering background from USC, graduated in 2015 and have been working.

What is the general consensus on rankings for Masters Programs in Statistics? The US News rankings (http://grad-schools.usnews.rankingsandreviews.com/best-graduate-schools/top-science-schools/statistics-rankings) have duplicates listed, I’m not too sure why, but it definitely doesn’t make me feel like that list is legitimate. Regarding rankings, how important is it to go to a top 10 or top 20 school versus a lower tier school? Do drop prospects fall significantly like they do if you go to a lower ranked law school, or is it more like medicine?

Most of the top ranked programs require three letters of recommendation. I’m wondering, how important are these letters? I think I will be asking three former professors, in classes I did exceptionally well in, but they are unlikely to be able to “provide detailed anecdotes and examples to support their assertions” (from stanford’s stat MS admission faq) other than that I aced tests. The best thing that these professors will be able to say about me is that I’m prepared for graduate study in a quantitative field because I aced their (heavily quantitative) classes and got the highest mark on exams one or two times. That’s it. I don’t have special relationships with them nor did I do research with any of them, and these types of classes are just where you have homework, do math, code in MATLAB a little, and take exams. I’m not sure what kind of anecdotes admissions are looking for, but my recommenders certainly wouldn’t have any of me.

With that said, what are my chances at getting into Masters Programs in Statistics, from the top 20 programs to any decent programs whatsoever? Here are my stats:

-Undergraduate degree in Chemical Engineering from USC, GPA 3.7, and the relevant courses I took were Calc III (multivariable), Diff Eq, one applied statistics for chemical engineers course, and one Matlab course. Received an A in all except for Diff Eq (A-).
-I have been working as a Materials Engineer for over half a year, but it is more of a glorified secretary role where I collect data. One good aspect of it is that I use Excel heavily and have written a bit of code in VBA in order to gather data automatically or quickly.
-I have not yet taken the GRE but I assume I would do pretty well, was in the 98th percentile on verbal on the SAT and should be able to get a perfect math score.
-My programming experience is sparse, I can write code in VBA and Matlab, and have done the intro courses on Python and SQL in Khan academy, but that’s it.
-I will take Linear Algebra in the Spring of 2017 at a community college, along with Probability theory course most likely.

Lastly, do MATLAB and VBA qualify as statistical software/programming languages? Berkeley recommends applicants to have coursework in “at least one statistical system (such as R) or a computer language (such as Python)”. I’m wondering, what language do admission committees prefer applicants to have experience with? Am I fine with the MATLAB I used in college and the VBA skills that I can build upon? Or do I need to learn R for example, starting now (since I’m hoping to apply by this years deadlines in December/Jan), and continuing to learn it/enroll in a course after my applications have been submitted?

Appreciate any help on any of my questions, thank you so much!!

2

I don’t know the answers to most of your questions, but I can say that MATLAB and VBA are not statistical languages. Those would be SAS, SPSS, STATA, R, Python, and even SQL to a lesser extent.

Consider getting a recommendation from your employer. Some professors like for students to write a first draft of a lor and give it to them to tweak and sign.

I imagine the opportunities from lower ranked schools would be good but not as good the opportunities from the higher ranked schools. Just a guess. . .

3

The U.S. News list has duplicates because some schools have more than one MS program in statistics or something related to statistics. For example, the two Harvard programs are the AM program in statistics and the MS program in biostatistics. If you click on the school’s name it tells you that - I clicked on Harvard and discovered that the #3 program is the biostatistics department and the #7 program is the statistics department. UC-Berkeley’s statistics department is #2 and their biostatistics department is #13.

For a master’s program in statistics, unless you want to do work that is adjunct to academia (i.e., as a staff statistician at a university medical center or a academically-oriented think tank) it’s probably not very important to go to a top-ranked school. Statistics in general is in such high demand these days that if you get a good education at a respected school and learn some skills to enhance it - like a computer language, or SQL, or both - then you should find your employment prospects good.

Letters of recommendation are important for graduate programs. Stronger recommenders will definitely have some anecdotes or personal relationships to discuss in their letters. But many a student has gotten into a master’s program, especially, without having particularly close relationships with professors. So just do your best.

Any statistics program is going to require you to have taken linear algebra, so it’s good you’re going for that and probability theory. (If you can, I recommend you take it at a four-year college instead of a community college.)

MATLAB actually can be a statistical language depending on how you use it, but I doubt you’ve used it that way. R is going to be the standard that most stats programs are going to use and prefer, and so yes, you should learn R. You can find yourself a good R book and start teaching it to yourself if you don’t have access to a class (it’s free and open-source).

For a full-on statistician SAS is probably the second-most useful, although it’s hard to access outside of a university or corporate setting as it is insanely expensive. SPSS and Stata is mostly used by social scientists and both are pretty easy to learn either if you already know R - but they are also easier to learn than R, and can be pretty cheap for students (both have GradPlans or student packs you can buy for under $200). R is more widely used these days, especially by hardcore statisticians, so I’d go with that one first (and it’s freeeeeeee!) If you go in only knowing SPSS, you might be in for a rude awakening - SPSS doesn’t use code (well, there is one, but few people use it well) and most robust statistical programs are best manipulated using code because it’s so much faster.

4

Thank you that was very informative!! I really need to decide if I should try to apply by the end of this years cycle (late Dec - early January 2017 deadlines) or wait a full year - in which I would have taken linear algebra/probability theory, possibly garnering better recommendation letters from those professors, and learned R. However, if its feasible for me to get into a decent program during THIS cycle, I’ll definitely hurry to get my apps in.

Therefore, what schools could I feasibly get into THIS cycle, considering the strength of my current application? Would those schools be considered “respected schools” that would result in good employment prospects? My current application would certainly lack great rec letters, and experience in a statistical programming language ®. I’m also missing those math courses (linear algebra/probability theory), although, according to the Chicago admissions specialist I spoke to its ok that I haven’t taken all the preparatory math courses as long as I’m planning to do so.

5

Personally I’d wait a year, but there’s nothing stopping you from applying to a few highly desirable programs, seeing if you get in anywhere and making a decision in April.

However, I think asking “what programs could I get into now?” and applying for those is the wrong way to go about it. I think you should do the opposite - only apply to the programs you’d really like to go this time around. The logic there is that if you apply to mediocre programs, you’ll always wonder if you just simply waited a year whether you could’ve gone to a better program. You don’t want the quality of your grad and professional experience determined by impatience.

So if it were me, I’d pick maybe 3-4 programs you were really concentrating on, apply to those, and see what happens - with the understanding that if you don’t get in anywhere then you can wait and reapply in 2017, and apply to more programs then.