bc2012

I'm interested in quantitative social science research, and was wondering whether a PhD in Statistics would be appropriate. Ideally, I would do a PhD in Applied Statistics, but I can't seem to find such programs. Many PhD in Statistics programs heavily emphasize proof-based theory, but I'm much more interested in methodology and application. I was initially thinking an econ phd, but their methodology doesn't incorporate some stat topics I'm intersted in (little bayesian outside finance, not much multilevel modeling, no machine learning, etc.).

INSEAD is great, but it lacks the resources of a university. For example, if you want to build programming or statistics skills, there are much fewer options than USC/Stern/Columbia.

I suggest RA work over a Masters degree for recommendation letters. You get to know a prof better from working rather than just sitting in on an economics lecture.

All these sound normal. "Crazy" would be a school accidentally sending acceptance emails to rejected applicants.

Do some RA work for econ profs and get 2 LORs from that, and I think you have a good chance at Top 20. Yes, Top 20. I find profile evaluations on this forum pretty ridiculous. If you "own" your application, and really show you can offer something other applicants cannot, then you'll have a good chance.

Do you actually say what schools you have offers program? Is it rude to decline to name specific schools?

I have an interview day on March 15 on the west coast, then another on the east coast March 16. These are both interviews (NOT visits for accepted students). Is attending on the fixed dates important? I could attempt a red-eye flight, but probably won't be very alert for the second day.

Think back to college acceptances. It's not like people decline their safety the same day as getting into a better school. There are so many factors involved. People worked hard to get into PhD programs and they deserve the time to consider their choices carefully.

Saw Columbia on GC. Wonder if it's micro or macro. I believe they interview one, but not the other.

Mac are pretty to look at, but terribly inconvenient for doing work. Some problems I encounter: 1. You can't open up multiple sessions of Stata. Sometimes my prof calls and asks to look up a sumstat really quick. It would be nice to open a separate dataset without having to close your current session. 2. Foreign characters screw up, making merges based on name very difficult. 3. Powerpoint incompatibilities. Sure, you can blame it on Microsoft, but if you want to do something about it, then get Windows.

OMG, this thread is going all over the place. Let's recap: ORIGINAL QUESTION: What to wear to visit day at program where you were accepted? CONSENSUS: anything, really - be it t-shirt&jeans or business casual. VARIANT QUESTION 1: What to wear to visit day when interviewing (i.e. not accepted yet)? CONSENSUS: business casual or full suit & tie. VARIANT QUESTION 2: What to wear everyday once in grad school? CONSENSUS: doesn't matter, though slack/button-down would be nice when teaching

I'm wearing suit & tie for onsite interviews. But for visits where you are already accepted, does it matter what you wear?

Is there a dress code for visit days? I was planning polo and jeans, but should I do business casual?

For complicated data manipulation, those guides probably won't help. Try asking your neighborhood grad student or RA :)

Just hope a non-native speaker will read it and not notice.

Is there usually a formal dinner involved or social events with other interviewees?

I agree adcoms are looking for future researchers, but I still stand by my statement that "There's nothing wrong with doing a PhD to do a teaching career." I mean, there are plenty of folks that go into industry careers, which is farther removed from academia than teaching. I can think of two profs at my department with PhDs from Princeton and Columbia who only teach (both undergrad and grad courses) and advise. If teaching at a university is your passion, then go for it.

There's nothing wrong with doing a PhD to do a teaching career. I'm at a top 10 and there are several professors who are lecturers and do student advising too. Here's a Harvard PhD candidate who's goal is mainly to teach: About the Author

I'm debating taking a Python class too. This is the syllabus for my class. Is anything here useful for research (data manipulation, web scraping, etc)? I have no compsci background so I'm not sure what the terms mean. Introduction. Python Overview. Basic data types. Control flow. Sets. Lists. Dictionaries. Files. Functions. Iterators/Generators. Lambda expressions. I/O and Command Line. Regular expressions. Object oriented Python: Classes, Objects, Methods and Attributes. Exceptions. Modules. Standard Library Standard Library II: GUIs, Pickling. 3rd-party libraries: Scipy/Numpy, Pygame. Unit testing. Performance, optimzation, profiling. Parallelization. Language Integration. Python best practices.

I'm not saying getting good grades is a definite sign of cheating, but one does get suspicious after witnessing so many copying other people's homeworks, having solutions from previous years, having pdfs of solution manuals, passing notes during exams, looking at online lecture slides during an exam while "in the bathroom" (i finished an exam early and went to the library and found such a person), etc. I think a few (well, a lot actually) instances just gave me a very bad impression. Again, I think it's highly cultural as academic cheating is more normal and acceptable (Asians cheat like mad on US college apps | GlobalPost)

Someone PMed me about the PhD program. Figured I reply publicly: I think the PhD program is much better. The professors teaching PhD courses this year are great; I think because they are full professors instead of adjunct. All the profs are friendly; just the MA advisor is mean so you'd never have to deal with him. There are fewer students so it's more tight-knit and hence more collaboration rather than cheating. Also, it's probably quite difficult to cheat in a rigorous PhD program when all the profs know you and class sizes are tiny. I'm an MA student so my judgment of the PhD program is purely based on observation.

I personally am not a fan. Columbia is a great university, but not this particular department. I'm not sure if stat departments elsewhere are similar. GETTING IN: I'm pretty sure they'll accept anyone willing to pay. There was a student who was taking calculus and linear algebra, which were supposed to be prereqs to admission. I was an econ major at a Top 20 UG with a 3.6 GPA. The only relevant courses I took were calc 3, linear algebra, intro stat, and econometrics. PROGRAM: Only 8 classes; no thesis. There are 4 required classes plus 4 electives. The stat department courses are all theoretical; they just started offering an applied track this year which is just 2 additional applied courses. If you want proof-based courses, the stat department is just for you. If you want stats with applications, I recommend takings classes in engineering, computer science, biostat, business, economics, or poli sci. FACULTY: Most grad classes are taught by adjunct faculty, including the program advisor so don't expect any guidance/support from him. I find him very mean. Most of the professors are nice, but not that great at teaching. It seems the good teachers only teach undergrad classes. STUDENTS: Another reason to take classes outside the stat department. Prepare for a culture shock. 95% of the students are Chinese. You probably won't hear English in the classrooms aside from the professor. They are also notorious for cheating (on exams, copying homework answers, having solution manuals, etc). By no means am I trying to be racist, but cheating is much more prevalent and acceptable in Chinese cultures (Asians cheat like mad on US college apps | GlobalPost). Most of them will score 100s on the homework. Exams are fair-game if the professor takes measures to prevent cheating; otherwise, don't expect too much of a curve. OVERALL: Despite the above, I think the fact that you can choose any 4 electives provides a lot of flexibility to customize your program. Just be sure to check out classes in different departments and do a lot of course shopping. Take the time to educate yourself; don't expect the department to guide you. If you want a friendlier student atmosphere, I recommend QMSS which has easier required classes but you can choose tough electives. HINT: After you do two semesters of full-time, there is a very deep discount if you do another semester (only $8000 to take up to six courses). Of course, you shelled out $40,000 to get to that discount.

Why does everyone think a 3.7 is low? I'm in the program now. My gpa was 3.6, major 3.3(yes, lower) and my only math classes were Calculus III and Linear Algebra.

Is "Experimental Design" useful? I saw a syllabus and didn't recognize most of the topics (factorial design, mixed design, latin squares, etc)

Panel data falls under GLM? I thought GLM was poisson, ordered logits, multinominal probits, etc. and these some a bit difficult to explain to non-technical people too.