finalrez

Just my two cents -- I think it depends on the reason. Some of the schools I declined because I got better funding elsewhere. To me, less funding seemed like a totally sensible reason and so I politely responded with those specifics. On the other hand, I would of course not say something like, "I got into the program XXXX which is much more highly regarded in the mathematics community so I will not be enrolling at your school."

I've basically narrowed down my decision to these two PhD programs, but like many others here I'm having trouble making up my mind. Santa Cruz makes me a little nervous because it's a relatively new (unranked) program, and it is more expensive than NC State. On the other hand, the department seems to be working on some very cool stuff (mostly non-parametric Bayes as far as I can tell). I'm interested in computational methods and the Bayesian focus is an attraction. Also, I thought it might be easier to get lost in NC State's program because it is so much larger. Here's some of the things that I'm wondering about: Stability of funding -- particulary, is NC State funding typically provided at the advisor or the department level? Attrition rates and relative difficulty of first year exams/oral defense Placement rates in academia or industry. I wasn't able to find much information about either program Any other advice? Thanks in advance --

I've worked for a number of years as a software engineer, and I've used a number of languages and frameworks along the way -- Python, Django, PHP, Java, JavaScript, SQL, JQuery, HTML, C, Zend, Clojure, R, Stata, SAS, Matlab... I may be forgetting some but you get the idea. For your purposes, I'd highly recommend learning Python if you want to learn a new language. I wouldn't bother with C -- it's too low level for most research. C is good when you need routines that are very efficient, or when you need to do things at the byte level. The downside is that C takes many more lines of code when compared to higher level languages. You'll spend lots of time writing algorithms and routines that come for free with other languages. For research, programmer efficiency is paramount, and C just won't serve you well on that front. Also, C is not object oriented (as opposed to C++). Java is a very nice language, but again, probably not the best choice for research. Java is higher level than C, and is sort of the de-facto standard to which most other object oriented languages are compared. Java is almost certainly one of the most popular langauges used in industry (i.e. not in academia) today. For a variety of reasons (strongly typed, good encapsulation, interface support, etc.) Java is good for enterprise type projects for which many programmers must collaborate. Java also has great library support, probably better than any other language, and it also has great integration in terms of testing and continuous integration for large projects. Some people argue that Java is more verbose than Python, and that argument definitely has some truth to it, but newer versions of Java have gone a long way towards closing that gap. Java is not, however, a scripting language. You must write in an object oriented manner when using Java. The fact that Java enforces an object oriented style might be an upside if you just want to learn about object oriented programming, but if you just want a quick and dirty prototype for your research it can get onerous. This object oriented style and built in access control schemes also make the barrier to entry relatively high for even basic programs. When you're just starting out this can be a burden because you have to read a lot before you can even completely understand a basic "hello world" program. Also, my completely anecdotal impression is that Java is not as popular as Python in academia. C# is basically Microsoft's answer to Java. I'm sure there are some more meaningful differences, but I don't really have much experience with C#. Honestly if you're interested in C# I'd just learn Java and then if you need to pick up C# later it will probably be relatively easy. VBScript is mainly good for writing macros in Microsoft products. There are some limited applications, but the syntax is poor, and I think you'll find it limiting relatively quickly. I'd skip it. PHP is used for building dynamic web applications. Unless your just need want to program websites, just skip this one. It also has a sort of klunky syntax, and its object oriented support was sort of tacked on in later versions rather than baked in, which can make it feel kind of klunky. JavaScript is a client side programming language for the web. It's a popular beginner language, but again basically useless for your purposes. SQL is a database query language -- it's a very good skill to have in terms of employment and research as well, but it's more of a complement to 'real' programming languages, and so I would skip it for now. I wouldn't bother learning a functional language like Haskell or one of the Lisp dialects until you are firmly grounded in object oriented design. I say that not because I think an object oriented style is better, but simply because object oriented langauges are so much more popular -- functional languages are used comparatively rarely in both academia and industry. For statistical programming your choices mainly SAS, Stata, Matlab, R, and Python (and maybe Mathematica? I don't have much experience with it). SAS is a very powerful tool, although not really a programming language in the same way that say Python or Java are programming languages. In that sense, it's something more similar to Matlab. SAS is especially popular with informatics, epidemiology, and bioinformatics types of research. Personally I didn't care for SAS's syntax very much, but I'm sure there are others on these forums who would disagree. Again, completely anecdotally, in my opinion SAS's popularity seems to be waning. Stata is sort of the same, again not really a programming language, and it is popular with economics folks because it has good regression support for sometimes esoteric models. I'm not sure if it would serve you well for engineering types of problems though. R is an open source language that is very popular with stats types and other disciplines as well. It has a strong community and seems to be growing in popularity because of a newfound interest in "big data." Overall, R is a good research tool and is popular in acedemia. R is probably not a bad choice, and is in some ways similar to Matlab, so you'll probably be able to pick up the basics quickly. I think that pretty much covers it for the languages that people often suggest as a beginner -- and now I'll put in my plug for Python. Python has good object oriented support, but is also powerful as a scripting language. Python is also versatile -- you can use it for research, building web applications, and desktop applications. Python's syntax is among the best of modern languages, and you'll be able to pick up the basics quickly. Python is also supplanting R in some datascience applications because of some great statistical libraries like Numpy and Pandas. If you want to get a job as a programmer Java might be a better choice -- there are more programming jobs for Java programmers. However, the programming jobs for Python tend to be more interesting IMO ;-) Also, consultants and analyst jobs (even outside of the tech industry -- i.e. banking etc.) tend to want people who know Python far more often than they want people who know Java.

I thought it was much more similar to the SAT than to the ACT. I'm a bit older though, so when I took the SAT it was the version on the 1600 scale.

I don't think there's a full enough profile here to speculate on your chances of admissions to specific schools, but to your question "Will being a white female give me an edge in statistics and data science programs?" the answer is probably yes (even if it isn't politically correct to say so). You will get more of an edge for being a domestic student than for being female. You will also probably get more of an edge in stats programs than in biostats programs, but stats programs are typically a bit more competitive so that may even things out. As an aside, my opinion is that it is in really poor taste to apply and take a PhD slot if you have no intention of finishing the PhD, but that's another topic...

According to ETS GRE scores are valid for 5 years (http://www.ets.org/gre/revised_general/faq/). Every school that I've looked at that mentions a cutoff also goes by the 5 year rule. I assumed that the 5 year cutoff was from the time the application was submitted, but I actually saw one school that did it by matriculation date. I guess for you it wouldn't matter -- you would be fine either way. I haven't seen any schools that have a 4 year cutoff, but obviously your scores are older than 4 years. If I were applying to a department like that I would just go ahead and ask if they'll take your scores anyway (since the ETS site says that scores from July 1st 2008 are still available). It can't hurt to ask!

The first thing I would say is don't be deterred by the profiles that you see on gradcafe. Many schools publish data about GRE scores, GPAs, etc., and if you look at the average admitted candidates, you'll quicky realized that there's definitely some self-selection bias going on in the profiles you see on the forums. I do think that statistics has gotten more competitive as a discipline in recent years, but certainly not so much so that you won't have any chances of admission. It sounds like you have a reasonable academic background. Have you taken multivariate calc? It sounds like you probably did, but if not that's one that's required by most programs. Your CS experience might be a little light for programs that are heavy in machine learning, but I think you probably have considerably more programming experience than most recent grads. Your GPA/GRE might be a little low for top 10 programs, but your grades and scores are absolutely good enough to keep you out of the auto-reject pile for many good stats schools. Not having any published work isn't really that big of a deal. Having published work is actually more of the exception than the norm. OK, so now that I've convinced you (hopefully) that you are a viable candidate, I'll get on my soapbox. It sounds like you may be a bit unsure of exactly why you want to go to graduate school. Rather than look for diverse departments, my advice is figure out exactly why you want to go to graduate school in the first place. Once you do that, your school selection will become much easier. Wanting to learn more is a great reason to go to graduate school, but what do you want to learn about? What questions do you want to tackle? Problems in biology? Astronomy? Time series financial data? It looks like you have an interest in psych -- you might like doing classification/cluster analysis kinds of problems (since you mentioned machine learning) in biostatistics or computational bio programs, or maybe working with clinical trial or survey data. I know others will disagree, but I honestly don't really think that wanting to get a job (in the general sense of "getting a job") is a great reason to go to graduate school. Definitely don't go to graduate school if you're just trying to put off choosing an occupation. IMO it's smarter to choose an occupation that you think you might like, and if you hate it, you can go back to graduate school as a career change kind of move. It's a lot harder to do it the other way around. Also, if you're smart about what job you choose, your graduate application will be stronger. Plus, you will likely be paying quite a bit of money to explore your interests in graduate school, whereas on the job some sucker will pay you. If none of that convinced you, then my final advice is try to find large programs. This serves three purposes. First, it will enhance your chances of admission. Second, it will be easier to expose yourself to more ideas as you explore your interests, and you will be more likely to find an advisor with interests close to yours. Third, it will make it easier for you to network and get a job once you graduate. Sometimes programs will tell you directly the number of current masters and PhD students, and other times you will have to infer by the number of faculty members.

I agree with awells. It absolutely is possible to write a strong essay that doesn't fall strongly on either side of the issue. That said, it is much easier to end up with a muddled essay if you don't pick a side (especially if you find yourself at the end of the essay without time to revise seemingly conflicting viewpoints), which is typically why the test prep places recommend just choosing a side arbitrarily and going with it.

Yes, keyboard inputs do work on the actual exam. I had the same question before my exam, and I couldn't find anyone who could tell me the answer... Keyboard shortcuts also work in the Kaplan practice tests, but not in the Barron's practice tests. At least in terms of test flow, the Kaplan tests are much closer to the actual exam. The closest of course is the ETS powerprep software. I don't remember if shift+home, ctrl+x etc. work on the exam or not. My gut feeling is that they do, as I don't remember running into any issues like that, but again I'm not sure. If you're concerned, I would try it on the powerprep software -- if the shortcuts work on powerprep they will almost certainly work on the actual exam. The real GRE is virtually identical to powerprep.

I'll preface this by saying that I'm not a professor, or a graduate student - I'm just a guy on the internet who is also applying to statistics programs. So you probably shouldn't believe anything that I say... 1. My guess is that it depends largely on what you mean by "decent." The best schools (Stanford, Berkeley, UW. etc.) are frankly probably out of reach given your limited math background and GPA. The big omission in your math background is linear algebra. Real anaslysis is also nice, but not necessarily a requirement. Even assuming that you are a US native, it's honestly not worth applying to anything listed in the top 10 on US News (or probably even a little lower) unless there's something else stellar in your profile that isn't listed above. Some other thoughts -- Applied statistics will be an easier sell than the more theortical programs given your limited math background. You may have more luck applying to biostatistics programs instead of pure stats. I think that biostatistics would still be perfectly suitable if you want to be a data scientist, particularly if you try to tailor your learning for data science types of problems. Biostatistics programs tend to be a little less competitive, and often weigh the verbal component of the GRE a little more, both of which work toward your favor (assuming your GRE projections are correct). You could also make a good sell for many biostatistics programs with your neuro background. Finally, you might want to take a look at computational biology programs (especially if you're into computers/programming), which also tend to be feeders into data science type gigs, and for which you might be more qualified. 20+ is a lot of programs. My guess is that you actually will get better results applying to a smaller number of programs in the range that you are qualified for, and more carefully refining those apps. I wouldn't worry about the D in your final quarter too much - especially if it's not in a math class. No one cares if you got a D in Canadian Literature. Even if it is from a math class, a single bad grade isn't going to ruin your app. I would be a lot more concerned about your "upper division" GPA of 2.6. I'm not totally sure what you mean by "upper division," but it sounds like your GPA declined over time. Adcoms like to see improvement over the course of your academic career, and getting low marks your junior and senior year could hurt you. Hopefully this is obvious, but be ready to field questions about why your GPA declined, why you got a D, etc. with a plausible reason that is not 'partying too much.' I hope that isn't too discouraging, and it's also worth pointing out that you are in an enviable position of being able to improve your profile. Cheerfully for you, from talking to professors, I've found that among GRE scores, letters, and grades; grades, perhaps surprisingly, are the least important. Also, If you took linear algebra, real analysis, and a couple stats/probability courses and got good grades (and maybe some awesome recs) you could greatly enhance your app. Keep in mind though, that for many programs GRE scores are often used more as a cutoff, so the difference between say a 163 and a 167 won't necessarily matter that much except when candidates otherwise have very similar qualifications. 2. Probably not, although if you did well it could help. The subject test is notoriously competitive, and not many applicants (especially domestic applicants, and especially masters students) bother to take it. That said, I think that there are other areas in your profile that you could more easily improve. My gut feeling is that taking a few more math classes and doing well will improve your chances much more, and probably also be less work. 3. This is the million dollar question. If I were in your situation, I would try to emphasize why I wanted to study stats, especially because you are coming from a background that is a tad atypical. Why does statistics excite you? You might briefly explain why your grades aren't stellar if you have an appropriate reason, but I wouldn't dwell on that aspect too much. Focus on what you've done well, and why you're passionate about learning statistics! 4. From what I've gathered, reaching out to professors and directors isn't nearly as important in math/statistics as it is in other disciplines. Statistics grad students rarely have a viable research idea entering the program, and probably for that reason it isn't as important to try and identify and contact a specific group/mentor beforehand.