statsguy

I saw how admissions worked and there are basically two relevant categories: auto-admits (e.g. finish a 4.0 combined BS/MS degree in 3.5 years at an Ivy, two publications in good stats journals, and LORs from superstar researchers) extremely-strong candidates (someone like you) Once you are in the extremely-strong category, there are diminishing returns on improving your application given the time and energy you need to spend. Taking 7 vs 6 advanced stats classes is unlikely to change much, whereas 4 vs 0 makes an enormous difference. I remember one year someone got a major bump for his astronomical score on the comically difficult GRE Math subject test (scored in something like in the 98%th percentile). Candidate rankings can also be really subjective in this category, and unfortunately a random component comes into the equation. If there are 20 candidates in the "extremely strong" category and there are 10 offers in the initial batch, the initial 10 can look very different depending on the committee because it's so hard to separate applicants in this group. A lot will also depend on who else is applying (there is definitely minor variability from year to year) and where others choose to go. There are some years where we had to send 15 offers to get 10 accepts, some years we had to send 11 offers to get 10 accepts. There are more uncontrollable things as well which I wont mention here... but you get the idea. Apply broadly and you should get into at least 1-2 top programs. Edit: I also agree to not delay graduation. It's one thing to switch majors junior year which forces you to delay graduation, but in your case it'll likely just raise the bar higher.

Taking 4-6 weeks off in the summer, 1-2 weeks off in the winter, 1 week in the spring, would have been totally acceptable in our department. However, if you're looking at getting a good academic position (need to publish lots of papers), or even just looking to finish comfortably within 5 years and go into industry, it becomes incredibly difficult to put in the shear number of hours if you take 3-4 months off in the summer. Plus there's the issue of atrophy as well when you totally tune out for so long. Prior to when I entered, there were stories of students who backpacked all summer or worked full-time at hipster coffee shops while bartending or playing gigs at night with their band, and it would take them 7-8 years to graduate only to get a mediocre industry position. The department put an end to this by only guaranteeing 5 years of funding, pushing the qualifying exam out to after year 2, instating mandatory summer research after first year, and most importantly, providing summer funding. If summers off is what you are looking for, I don't think you'll find it in a Stat PhD program.

It will vary by department, when you take the quals etc. In my case it was: After year 1: research project with a small stipend. Basically to get you used to doing research, reading journal articles etc. I read several articles of interest, reproduced the results while at the same time teaching myself to write code. Some students already entered the PhD program with an advisor in mind so they basically jumped right into research. TA-ships were available for a limited number of students but we all got a small stipend ($2500/summer) to help keep us afloat. Being a summer TA would've doubled summer funding to $5000. After year 2: prepare for quals which are taken in August. Do research or a consulting project. Again, small stipend ($2500) with TA-ships available (another $2,500) if desired for a limited number of students (this was highly discouraged for those studying for quals). After year 3: start getting serious about research and preparing for the prelim oral exam. Do research. TA-ships ($2,500/summer) and grad instructor positions ($5,000/summer) available again. Some students did an industry internship this summer as well. After year 4: By the end of year 4 you should've finish your prelim oral exam. Work on dissertation/research and be a summer TA or grad instructor. Some industry-bound students also did an internship this year.

In our department (a top 15ish ranked school) an unusually low GRE quantitative score would be a red-flag. I was kind of shocked to hear that from the grad program director himself, as I had always thought that test was a formality.

I think this is field-specific. I will finish my PhD in Statistics this Spring and there are currently TONS of industry positions available (with starting salaries ranging from 90-120K/yr). We've had graduates in the past few years go to Google, Microsoft, Pfizer, Citibank, etc... even all the international students with immigration related issues are almost always able to find employers willing to sponsor them.

For applied fields such as data mining, machine learning, or operations research, linear algebra will be EXTREMELY important. I cannot stress this enough. Know it as well as you know basic arithmetic. Numerical Analysis is good too. For theory, Real Analysis (Rudin level), Measure Theory, and some Functional Analysis are important. Having a decent understanding of programming is good too. Obviously any Applied Statistics, Machine Learning, or Operations Research courses will help.

Don't sweat the verbal score. Half the grad students in Math/Stat PhD programs are foreigners whose English is (sometimes) pretty bad. No one will care about that score for Math/Stats PhD programs. Maybe they will for Econ; I'm not that familiar with Econ PhD programs. Having a Master's will help because hopefully it means you have a solid foundation for PhD programs. Although be prepared to redo a lot of the coursework. The undergrad GPA will most likely be ignored if your MS GPA is that good. Your MS coursework is important too. I'm assuming you took Real Analysis, a year of Mathematical Statistics, maybe some applied courses... Good letters of recommendation are important too. Just as a warning, you may want to apply to more schools. Departments everywhere are cutting back on the number of students they are accepting due to funding issues, whereas applications in some fields (such as Statistics) have spiked in the fast 4-5 years.

If you eventually want to get a PhD in Statistics, take Real Analysis. But if you just want an MS so you can do applied work, I wouldn't take it. I know MS students in my program who came in with degrees in Biology or Physics, even Psychology, having only taken Calc I and II in undergrad, and they are doing fine. Again, just to warn you, you will be taking a lot of classes that are useless in the field of Statistics by completing an entire BS in Mathematics (ie Number Theory, Abstract Algebra, Diff EQ etc...) Now if this interests you, great, go ahead and take them just for fun. Since you already have an MS, I don't know how much more school you can stomach. I don't know what the pre-reqs are for MS programs in Biostats, but I am aware that the pre-reqs are lower for Biostats PhDs, largely because (as you said) there is much less theory and the programs are generally not as mathematically rigorous.

Go straight into Linear Algebra + Stats classes. These courses use only basic integration and differentiation techniques; you can easily brush up on what you need to know, especially if you have already taken Calc I + II (even if it was a while ago.) You will be wasting your time by taking the entire Calc III course since you only need to know how to take partial derivatives and do multiple integrals. Also, don't bother with DiffEQ. This way, you can start an MS program in 1-2 years rather than in 3-4.

For convenience, here they are: http://graduate-school.phds.org/rankings/statistics/rank/_M______________________________________________________________U Keep in mind that the data for these rankings was collected around 2006 (IIRC), so these rankings may already be slightly out of data. I actually think the 2010 Newsweek rankings were much more accurate. I'm really surprised at the NRC rankings.

In the US, almost every good pharama biostatistics job requires and PhD in Statistics or Biostatistics. It is absolutely required for advancement to a management or supervisory position. There may be some jobs available with a Master's Degree, but there is a very low ceiling for those people in terms of salary and advancement in pharma-- you will likely be stuck as a SAS programmer and work under a PhD for most of your career.

The PhD does offer slightly more job openings higher salaries, but not nearly enough to justify the extra 4 years he spent in graduate school after he got his MS. Again, this is according to him, other opinions may differ. The PhD is ideal for academia.

I'm getting my PhD in Statistics but have a friend who finished his PhD in Industrial Engineering at a very good school. He got a job a doing process/quality control at a large company, starting salary was around $85-90k (and this is in Houston, where the cost of living is very low.) However, he kind of regrets not just leaving with the Masters and going straight into the workforce, maybe eventually picking up an MBA. In general, if you want to work as an engineer, the MS seems to be the best degree when time spent in school, lifetime earnings, job openings etc... are all considered. I think a PhD in Statistics with a dissertation having something to do with IE/OR (for example, statistical quality control, machine learning etc...) is a much better plan than a PhD in IE or OR, if you absolutely want a PhD and want to work in industry

I took a look at the UH Math page and would recommend the MS in Applied Mathematics (MSAM). I recommend the MATH 6382;6383 and MATH 6360;6361 sequences. You may need to take some pre reqs your first year for that. For all of your electives, take courses in Probability and Statistics. You should be able to finish in 2 years if you are motivated.

Your two year plan seems really solid. With an undergrad degree from Penn and two years worth of Math/Stats classes at U of H (not at all a bad school) you can definitely get into top 30 programs (assuming your grades are pretty good at U of H.) Lots of people do their entire UG at schools much worse that U of H and get into good grad schools. I would personally skip the CS courses; maybe take a one semester introductory course in programming just to learn some basics (ie while/for loops, if/then statements, etc...) which will help, but any more than that is overkill. Skip the other CS classes and take as many upper level / graduate level Statistics classes as possible (Regression, Time Series, Stochastic Processes, Design of Experiments, etc...) This will definitely make your application look good. Do not take any Algebra/Topology/Number Theory/Complex Analysis. You will absolutely be wasting your time (I took all those in UG) unless you really enjoy the subject. You may be able to get away with skipping Diff Eq also. Also, you should be able to pick up an MS in Stats in 2 years, you may want to look into that. Again, another plus on the application. Also, if possible, try to get involved in a small research project while you are there. It will give you something to write about in your personal statement. Letters of recommendation are extremely important; make sure 3-4 professors know your name and like you. Do well in their classes.

In general, I would ignore all of those Statistics unless you know how the site obtained those numbers. A US citizen applying for an industry position will likely have several offers to choose from. Academic job placement is much lower for ALL applicants (maybe around 30-40% depending on the year), so those that are absolutely set on academia are probably pulling the overall average down since it may take them several application cycles to get a faculty position. However, most students go into industry if they are unsuccessful in getting an academic job. Some international students are unable to get industry jobs due to citizenship/visa issues (even though a company may want to hire them), so that again is probably pulling down the average. Starting salaries for industry are closer to 90k, and around 100k at big pharma companies. Academic positions start around 50-60k. So, to summarize, the situation is MUCH more complicated than those numbers that site reports. You have to look at each program carefully, which means talking to its students and faculty.