Jump to content

health_quant

Members
  • Posts

    22
  • Joined

  • Last visited

Everything posted by health_quant

  1. If you're interested in enrolling in a doctoral program after the MS, one big benefit of attending Brown's program is its small size. With fewer than 10 students entering each year (~3 being doctoral students), it may be easier for you to establish closer relationships with the faculty there. Letters of recommendation will be extremely important for further graduate study, and solid references may help to obtain a position in industry (though I don't know how common it is for employers to check recommendations for MS-level positions).
  2. I think you and I are in similar boats this year, and maybe even the exact same boat. (I am assuming that by second-tier stats program, you are still referring to a top 10 or top 15 stats program, as tiers on gradcafe seem to be used to distinguish between programs within that upper echelon.) I think that what earlier posters have noted about academic job placements (i.e., publications outweighing school names) really is the most important factor to consider if you want to go the academic route yourself down the road. The top biostats programs seem to do consistently well in preparing their students in this way (though biostat_prof has noted a general tightening in the job market), and if the stats program has had good placement recently, then I assume they've been supporting students well in this respect, too. Of course, any success has a lot to do with the students themselves and their relationships with advisers. Regarding private sector employment, I imagine that the relative importance of name and program-type will depend on the industry. If it's a highly statistical position, then I imagine that your hiring will be done by other statisticians, or at least people with an awareness of the general strength of stats/biostats programs. In this case, I feel like a candidate who would be competitive for an academic position would also be quite competitive for the private sector. I suppose if the private sector job really emphasizes applied, collaborative research, then a strong history of that may be quite helpful. April 15 is loooooming. I hope you're having an easier time deciding than I am.
  3. I'll echo Kimolas and recommend getting more research experience (and checking with your undergrad professors is a great idea). Your math background is far from weak, so I don't think that taking the math GRE will give admissions committees much more information on you as an applicant. For what it's worth, a friend of mine was admitted to Washington's PhD program in statistics without an undergrad major in math or stats. He did, however, have a lot of relevant research experience.
  4. I would agree with you regarding Stanford and CMU, but UW versus Cornell, Michigan, or Wharton is definitely debatable (especially the latter). Among the new and/or incoming ML-related hires at UW: Carlos Guestrin (formerly a prof at CMU in ML/CS); Emily Fox (formerly a prof at Wharton); Ben Taskar (CS/Stats at Penn Engineering/Wharton). There's definitely a big push from UW to build up ML in the stats department (while integrating with UW CSE), while biostats and stats have historically been linked quite closely.
  5. Uromastyx is right on. For the school, extending an offer to someone who isn't completely committed to their program could tie up one of their admissions slots for weeks and delay them from extending the offer to someone on their waitlist. Just like we'd like to hear back from schools as soon as possible, schools want us to decide quickly, so that their own waitlist doesn't evaporate while those waitlisted students take other offers. If this is your top choice, let them know. It seems likely that they want to admit you, but they just want some assurance that they're not wasting their time.
  6. I don't know for certain, but if they're doing more rounds of reviews, then it seems their offers wouldn't be limited to people who attended that day. Given the strength of their program, I imagine that Brown's yield is good, but it's still possible that they could exhaust their initial list of interviewees without filling every slot, as those initial people are probably receiving multiple competitive offers. I think that at this point, we should consider no news to be good news (which also applies to any other schools from whom we haven't heard).
  7. Brown's biostats program is very careful with its admissions process, as they only aim to admit 3 or so doctoral students per year. During the (first? only?) interview/recruitment weekend this year, there were only about 8 applicants present per arm of their public health program (biostats, epi, health services research). Despite being relatively small, the biostats department is doing some extremely interesting work. They're definitely among the top picks of the schools to which I applied. Fingers crossed that we all hear some good news soon.
  8. As others have pointed out, all of the aforementioned are top-notch programs, but it seems that even within this upper echelon of schools, distinctions are still commonly drawn. When we're looking at the total population of schools, I agree that these are nearly all top-tier programs. As far as this thread goes, the distinctions seem to boil down to drawing sub-tiers within that topmost tier.
  9. Same here. I'm scratching my head because I've gottten responses from Washington (biostats) and Wisconsin (stats), both with full funding, but not a peep from UNC. I thought the concordance of admissions from these places would be pretty high...
  10. The level of theory seems to vary quite a bit across Biostats departments. For Michigan, the Biostats doctoral students take the same requisite graduate, upper-division sequence in mathematical statistics as Stats doctoral students, but are not required to take the measure-theoretic course in probability. Perhaps it's also worth noting that Michigan's more measure-theoretic coursework is cross-listed in the Math department. Similarly, Penn's measure-theoretic probability sequence (with stochastic processes) is cross-listed under the Math dept. The doctoral-level stats course in mathematical statistics is taken concurrently with measure-theoretic probability, so it seems unlikely that their mathematical stats work(at least at that level) requires measure theory. Penn's Biostats department doesn't require measure theory (as it has its own sequence in probability and mathematical statistics), but evidently, many of the theoretically-inclined Biostats students take the aforementioned probability sequence through the Stats/Math department as electives. The University of Washington's probability and mathematical stats courses are the same for the Stats and Biostats departments. From what I understand, students from both programs take the same theoretical sequence for the first two years. (I believe UWashington is known for being one of the more rigorous biostats programs.) Wisconsin-Madison's Biostats program is just a concentration within Stats. In this case, the probability and math stats coursework are the same, i.e., lots of measure theory. I would imagine that biostatisticians with strong theoretical backgrounds (from the stronger programs) might still be viable in stats departments, but given the number of Biostats departments across the country (and industry positions), I don't think there's much reason for them to pursue those jobs. When you flip through the top stats journals (e.g., JASA, JRSSB), you'll see an abundance of biostatisticians along with statisticians. For us, the distinction seems to arise more from our relative emphasis on biomedical research applications. Even with Stats departments, research in cutting-edge methodology (which of course requires extremely strong theoretical backgrounds) is gaining in importance. We can see that with how the University of Chicago's been updating its Stats department, and with the number of hires from other backgrounds (especially EECS) for machine/statistical learning, and topics in non-parametric Bayes. Really, it seems like much of the disciplinary boundaries between Biostats, Stats, and CS are blurring now more than ever. More traditional topics in theory, (e.g., probability and stochastic processes) still seem as much the domain of mathematicians and applied mathematicians as of statisticians.
  11. haha. so true. spending all our time on thegradcafe doesn't help with maintaining a broader perspective...
  12. Their placements in biostats departments are quite good (high numbers of graduates in well-ranked programs), but I know less about their placements in stats departments.
  13. This is definitely going to be a tough decision, so don't feel pressured into rushing it. Whether you decide to attend Berkeley or Minnesota, you'll be in a very good position down the road. Some points in favor of Minnesota (as my last post was kinda pro-Berkeley): As others have already pointed out, Minnesota is a top-notch program, and at least in biostats, comparable to (if not outright stronger than) Berkeley. If you're planning on going into a non-academic (or even non-statistical) career, the Berkeley name may count for more, but within biostats, people will know what Minnesota means. As others have also pointed out, there's no guarantee that you will be able to stay at Berkeley. Your dad has a good point in that another round of applications during your second year will be taxing, and certainly distracting from any academic work/research in which you might otherwise be engaged. Should you decide to move into a different program, you will likely have 3 years in your next school (if they have an accelerated program for MS students like Michigan and Minnesota). In this case, your connections with faculty there may become quite strong, but they may not be as strong as connections you might forge over a full 5 years in one program. (Of course, they could be just as strong...who knows, right?) Also, some programs may not accept transfer credits, in which case you're in for another 5 years at least. The plus-side to this is that you'll really know your stuff once you leave, but the con may be that this could be overkill for any non-academic posts. Just another 2 cents from someone who still hasn't chosen a program himself.
  14. I think Berkeley is still worth serious consideration. Being a funded MS student at a relatively small department should afford you a number of good research opportunities with well-respected faculty in biostats (and possibly stats). Assuming that you make a good impression there, you would be extremely competitive for doctoral programs at the top schools (Hopkins, Harvard, UW Seattle), and you would presumably still be at least as competitive for Minnesota then as you are now. I may be biased though, as I'm currently a funded MS student (at a comparable biostats program to Berkeley's). For what it's worth, I was accepted at the University of Washington this cycle, which likely wouldn't have happened without the research experience and recommendations from the MS.
  15. I've also been thinking about these UW admissions figures. It may be hard to determine whether those acceptance rates are consistent across the MS and Ph.D. strata of applicants, as those numbers look like they're aggregated across all graduate students. (It would be strange if UW's doctoral admissions rate is that high, even accounting for the yield %.) UW is interesting in that they fund MS students (or at least most of them), but the cost of funding an international MS student would still substantially less than that of funding an international Ph.D. If we had more granular data, we could do some interesting analyses. (Agresti actually uses male/female admissions data stratified by different departments at Berkeley as an exercise/example in Categorical Data Analysis.)
  16. ah, beaten to the punch by seconds!
  17. If you apply to any schools using SOPHAS, your GPA will be calculated several different ways, with some measures aggregated across undergrad, grad, and presumably non-degree coursework. (There will still be a separate GPA for undergrad, which, as cyberwulf noted, will not include that non-degree course.) Also, some schools state on their websites that they're less concerned with exact GPA calculations than they are about performance in the relevant coursework (and whether that coursework provides the appropriate background in math/stats).
  18. From what I understand, Brown only admits a few students from the visitation/interview day and waitlists (most of) the rest. I don't know how their yield compares to other programs though, so an eventual acceptance may be possible even without attending the interviews now. Also, keep in mind that Brown's program is tiny, so for those of us who already have experience in stats, they may place more of a premium on fit than some of the larger schools. I'll post an update if Brown mentions anything interesting during this year's interview day.
  19. PittPanther: Are you a domestic or international applicant? Given your background, I'm really surprised that you weren't accepted or at least waitlisted at UW. Sisyphus1: I think you mentioned in an earlier post that you're an international student. Given all your time in the US, have you tried to become a permanent resident?
  20. Of course, the difference between the undergraduate and graduate levels will be depth, and the relative depth of the coverage will also be contingent on the field itself. I'm not surprised than an undergraduate, intermediate econometrics sequence (in a fairly mathematical program) might cover nearly the same material at a comparable level to a graduate-level sociology sequence. (Some of the minor differences would likely include a greater emphasis on probit versus logit for analyzing dichotomous outcomes, but whatever.) However, many undergrad econometrics and grad-level quant soc classes are hamstrung by the students' lack of exposure to probability and advanced statistics. Your background and current approach (hats off for using an engineering stats book) seem like they should prepare you well for being more than a "regression monkey" but be sure to check any prerequisites for graduate econ classes. Again, the assumed level of mathematical preparation is typically much higher for doctoral students in economics. Also, although many programs may not go so deep as measure theory, it's still worth keeping in mind that (forgive the bromide) you could end up losing the forest for the trees. As you've noted before, the structure of the first year in econ programs often differs from that of sociology, where much of the focus for the former is simply on preparing students for a qualifying exam. For most of these exams, the content will skew heavily toward the mathematical details. Here, I would think you would be better off knowing the distributional theory for regression/residual diagnostics (more applied) than knowing Black-Scholes or Karush-Kuhn-Tucker inside and out. It's not an either-or scenario, but I would guess (I don't know for certain) that many programs reserve more of the practice with the relatively applied concepts for third-semester courses in econometrics (or leave that for students to work through independently). My friends' experiences are similar to jacib's, in that they've taken plenty of courses in stats departments, but rarely go into econ. One advantage with the stats coursework is that many of the non-phd-level stats courses are designed as service courses for graduate students in other departments. If I recall correctly, some programs like Michigan, Penn, and Stanford allow doctoral students in other programs to pursue concurrent MS degrees in statistics. A program like that should also be able to get you where you want to be. I'm really not trying to discourage you from taking an econometrics sequence, but would just remind you that you'll have finite hours each day that you'll also want to devote to soc theory and developing ideas for research. If you're looking into doing work with corpus linguistics, I would also recommend looking at the course listings in computer science, and maybe starting to familiarize yourself with Python and/or Perl now. R (with which it seems you'll gain some experience) will be useful for your statistical work, but most people who do serious work in NLP end up doing quite a bit with scripting languages.
  21. If you're serious about learning what's going on under the hood and you've had mathematics through multivariable calc, pick up Statistical Inference by Casella and Berger. This will take you through probability theory and mathematical statistics at the upper undergrad or lower grad level. However, if you're not planning on doing a lot of heavily quantitative research and/or teaching yourself a lot of advanced methods in the future, C&B (and what follows) would likely be overkill. Alternatives to C&B: Mathematical Statistics (Wackerly, Mendenhall, Scheaffer) - similar topics to Casella & Berger, but at a lower mathematical level, in my opinion Mathematical Statistics (Rice) - a lower mathematical level than WMS, but some with weaker math backgrounds may find it as a good intro to the topics Mathematical Statistics (Bickel, Doksum) - slightly higher level than C&B; this one does a better job of emphasizing estimation of multiple parameters, while C&B sticks more to single-parameter estimation You will need a good background in regression to make use of all the above statistical theory. For that, you might try the following: Introductory Econometrics (Wooldridge) - this is used for the first-year quant methods courses in many soc programs (e.g., Penn's and UNC's); his upper-level book, Econometrics, is commonly used as an alternative graduate text on econometrics to Greene's Introduction to Linear Regression Analysis (Mongomery, Peck, Vining) - a good alternative to the introductory text by Wooldridge, and written more from the statisticians' perspective; note that you should have a background in multivariable calculus and linear algebra to get the most out of this book Linear Models in Statistics (Rencher, Schaalje) - good for self-study, as all solutions are provided in the back; the first few chapters review the essential linear algebra, but again, you should have already be familiar with eigenvalues, eigenvectors, spectral decomposition, etc. A first-year sequence in quant methods for soc will likely cover the basics of ANOVA, standard linear regression, logistic regression, and Poisson regression (all three of which are encompassed by generalized linear models) and possibly touch on multilevel and longitudinal structures, survival analysis, and causal inference. With stats, you can go much, much deeper than what I've listed above. That being said, I don't think it's necessary to know all of the above to be a good quantitative researcher in the social sciences, provided that you do have a solid understanding of the assumptions and limitations of whatever techniques you use. As for taking the first-year graduate econometrics sequence: I wouldn't recommend doing this unless you have an extremely strong background in mathematics. Many of the top programs assume knowledge of real analysis and some familiarity with mathematical statistics before entering the program. Some top-10 programs' econometrics sequences even begin with an overview of measure-theoretic probability, and dive right into asymptotic properties of estimators. These are not trivial topics. FWIW, my own background is in math (undergrad) and biostats (grad).
×
×
  • Create New...

Important Information

This website uses cookies to ensure you get the best experience on our website. See our Privacy Policy and Terms of Use