jjrousseau

Looks like this could be added: 2019 Profiles and Results: https://www.www.urch.com/forums/phd-economics/160952-profiles-results-2019-a.html

As mentioned above, probably the quality of your letters will matter a great deal if you apply now. Even with your grades, it's a gamble for the adcom at a T20 or maybe T30 to admit you from an unranked public university, and a strong and clear testimony to your econ intuition and research capabilities will go a long way making it feel less like a gamble. Connections from your profs may also help. It's hard for me to say with much confidence, but if your letters are great then perhaps 20-30 range? It's conceivable you'd do better or worse. In addition to many others on this forum, I've posted a ton about this. I've done both of these RAships. Feel free to search around the forum for some of my posts as I've definitely outlined the process before as well as competitiveness. But, your sort of profile - appears promising, but high variance because adcoms don't know what to make of your university - is the type that likely gains the most from these sort of positions. You'd be in it for the rec letter(s) (in addition to the money and gaining useful research skills/intuition/ideas). These letters will help dispel uncertainty about your potential.

No-name liberal arts does unfortunately hurt, both because it acts as a signal of the rigor of courses you've taken and because your profs are unlikely to be well recognized or have many connections among the admissions committees at PhD programs you might apply to. I do think you have work to do before applying, unless you're okay aiming pretty far down the rankings list (80-100, maybe? I'm not sure here). Unless you simply didn't mention it, one obvious area of concern is lack of research experience. Have you worked as a research assistant with one of your profs? Did you write a thesis? Did you write an applied research paper for any of your classes? This is just as important as adding obvious math classes like Calc III and Linear Algebra, if not more so. Econ MA is certainly an option, if you can afford it. I believe you can take your missing math courses there as well. But you'll want to scope out course materials ahead of time to the extent possible, as you may have a very hard time with the econ material without having the proper math background to start. In that sense, maybe a math/stats MS that allows you to take the obvious classes (Calc III, Linear, Real Analysis, something in Probability and Statistics at the level of advanced undergrad or early grad) would be worthwhile. In any of these, consider trying to work with a prof on research over summer/winter, and then including the prof you work for in your set of PhD rec letters. You could alternatively consider applying to full time research assistant jobs. I doubt your profile will be competitive at the more prestigious spots (NBER & profs hiring in top 10ish departments). It might be tough also to get recognized in the pool of applications at the Fed branches, but you can certainly try your luck there. You might try various think tanks (Mathematica, Urban, Brookings, etc etc). The more well known of these may still be somewhat tough to break into, but possible. There are many more out there besides the top names which could also be worthwhile. This sort of route has many benefits: (1) you make money, rather than paying it; (2) you strongly remedy the lack of research experience, and gain very useful rec letters from your boss(es) who directly observe your research work and testify to it; (3) often, these places will have tuition reimbursement programs, so you can take individual courses at universities offering that sort of setup and get reimbursed; (4) there's also more flexibility here if you decide a PhD isn't for you, in that these jobs will also be useful in getting you to a next step in industry, whereas it's less clear an econ MA would do as much for you there. You should discuss these sort of options with folks who know your situation a bit better as well. Try reaching out to your profs to talk. Ask about previous students from your school who have pursued PhD's. Also, if you do try to take one-off courses, I highly recommend Harvard Extension School's Math 23A, Linear Algebra and Real Analysis. This course is *hard*, so be sure you're prepared to do well in it (many course materials are online including section lecture videos on youtube). But it seems to be well received by PhD admissions committees, it's not terribly expensive, and it hits basically all of your missing math in one class (it covers linear algebra, multivariate differentiation, and real analysis).

I see. Yeah I guess I'm mostly emphasizing the disclaimer. I don't think one has to be ill-suited to the job for connections to be insufficient; even a pretty good candidate shouldn't rely on connections to pull them through if there are obviously stronger candidates in the pool.

I've been an RA at a T5 institution and was involved in our hiring process of new RA's. Connections will help get your application noticed. They may be particularly useful for someone whose application under-represents their ability (for example, the student who didn't realize until late on their interest in research and has a few weak grades on their transcript, who might normally get cut in a hasty scan of resumes). *However, connections are not sufficient for getting the job.* Most positions like this will have you do a coding task in later rounds, and will have interviews with the PI's. There will be a pretty direct signal of your ability here. If you do not have the requisite quality for doing the work, do not expect connections to keep you in the game. I wonder if Rohanps is extrapolating from PhD admissions, which do seem to be more influenced by connections (mainly, who's writing your letter). Note that there is not an easy way to demonstrate to a PhD adcom that you will be a good researcher, so there is perhaps more room for this sort of thing mattering. However, most of RA work depends on coding skills. This is easily and directly gauged via coding tasks. Don't expect PI's to fool themselves into hiring an obviously less talented RA because of a referral.

Your undergrad professor letter writer being well known is an important boost (see the recent thread on Top 10 admissions). Yeah, if you're at one of the stronger (in terms of undergrad) top 30 programs and have letters you anticipate will be strong - including one from a well-known prof - this all bodes pretty well. So I'd adjust slightly to the stronger end of what I said above. As in, you would probably expect a 10-15 admit, decent chance of getting a top 10 if you apply to many of them. I hesitate to get anyone's hopes up, though. If your financial constraints allow it, I think it's always useful to approach things safely. Personally, I'd probably apply widely across the top 15, a few in 15-25, maybe even one or two in 25-30 as strong safeties. But my risk aversion may be relatively high. Anyway, you should absolutely supplement info you get here with asking your writers directly what they think. And know that at the end of the day, there will always be noise in the process and you can never be too sure. Good luck!!

OP, I think for readers who are future applicants it would be helpful to more clearly distinguish whether the problem of attending particular undergrad institutions is (1) having relatively useless letters of recommendation / connections or (2) the actual undergrad institution per se. Letters can be remedied, as you touch on regarding MS/MA or RA positions. It is not at all infeasible for a 4.0GPA from an "unknown state school" to get an RA job in a top department (of course undergrad pedigree is still useful here, but far from deterministic) and then apply to PhD programs having letters from well-known top-5 or top-10 department economists. My prior is that this back door to the back door is effective, but readers would probably benefit from your distinction here. Directly: If someone comes from an "unranked state school" with great grades etc, and RA'd full time with a top 5 professor(s) who will write a solid letter of recommendation, what is your view on the extent to which a weak undergrad institution penalty still applies? (And, agreed with above that it's great you chose to share your insights.)

Two important remaining unknowns are (1) the strength of undergrad reputation at your school, as this varies a bit around the econ top 30 range, and of course (2) the strength of your LORs. On average, your profile probably gets into a 10-15 or 15-20 program. Fed is pretty consistent placing in this range. Adjust this slightly up or down if your UG institution and/or letters are above/below average. I doubt your real analysis grade will have much impact, given the rest of your grades. Top 10 is certainly possible, again I'd probably say that having above average (1) and (2) will really help here. I would roughly recommend applying a lot in the 10-20 range, a couple/few in the 20-30 range that match your research interests mostly as safeties, and a couple/few in the top 10 also prioritizing match quality. I would apply to lots of top 10 (along with the same 10-20 and 20-30 apps) if your inside info about (1) and (2) is strong, or even if it's roughly average but you're somewhat risk seeking and willing to take the financial gamble.

I've worked at both Fed and NBER, and was involved in our recruiting of new RAs. I'll second tutonic that 3.5 tends to be a rough cutoff, but we certainly looked at applications from folks with slightly lower (3.4, maybe 3.3) GPAs, especially if there seemed to be solid research experience and the transcript wasn't awful in especially relevant courses. Fed is less competitive than NBER generally, which I think matches with NBER being a stronger value if you want to continue on to grad school. Our group at NBER was hiring 2 people and received something like 120+ applications. We screened to about 30 to receive a coding task evaluation. From there, maybe 5 got interviews. (Bear in mind these numbers may seem a bit less strict if you consider that a few groups at NBER hired at the same time, and there was likely a lot of application overlap.) I have less of a sense of the numbers at Fed, but am quite confident it's an easier gig to land. Also as tutonic says, once you pass the first screen, you want to showcase your coding especially, as well as your general aptitude for research and critical thinking. If you get a coding task, do not take it lightly; it will likely be the deciding factor from that point on. If you get an interview, be sure interviewers know of any research assistance work you've done or your own research projects for classes etc. Show them you can work with data and you can think about research intelligently and explain yourself coherently. I will say, while there were tons of applicants to NBER from highly-regarded undergrad schools with GPA's of 3.6+ or 3.7+, the coding tasks were generally somewhat underwhelming. I point this out to emphasize again that you want to try to stand out at this point. If you are worried about your GPA (I'm inferring from the original question), rest assured that it will be largely forgotten if you can write immaculate code and present thoughtful results in later stages of the process, and those with impressive GPAs are not always submitting the quality you might expect.

If you're already well-versed in Stata and MHE, it's not that learning Python/R is useless. I just personally would prioritize it below these other things, including doing well in your MSc program. If you're solid on all the rest and twiddling your thumbs, by all means pick up a new language.

Maybe others will know better about the F specifically, but when you do apply, I'd probably recommend taking a relatively large list of schools. Given your strong research background but somewhat rocky grades, you may see more variability in acceptances. You may get top 10 admissions from schools that love your research background, and top 30 rejections from schools concerned about the F or withdrawals.

And, if you do these things, highlight them in cover letters / interviews!! Talking about any of these in an interview for an empirically-minded RA job would really be impressive, and would certainly give you a greater edge than Python/R for all but the handful of positions that *absolutely require* previous Python/R experience. Python/R are certainly useful, I don't mean to discount them entirely. But it seems to me that, while some small portion of RA jobs may really care about knowledge in a specific language, the majority just want RA's who are strong in code generally and in empirical economic research know-how. (Even among those that want experience in a specific language, Stata is still the language of choice for a good chunk.) So while not knowing Python may disqualify you entirely from a small proportion of jobs, improving your general coding skills and empirics will make you that much more competitive for the many remaining openings. It also likely has more value-added for you down the road. This seems an easy tradeoff to make, from my view. An afterthought: I would also guess that any position so invested in Python/R that they won't give you time to pick it up based on experience in another language is also going to only be after applicants who already have extensive experience with Python/R. As in, your self-study of Python/R would have to be deep in order to be very useful, and still you may well be competing against someone else who has already worked full-time using the language. All around I just don't think it's the most efficient way to spend your time boosting your RA application.

Check here as well for postings: Econ RA Listings (@econ_ra) | Twitter Generally, you will be limited as not all positions sponsor visas. Postings should mention if they can sponsor; if not you should check before applying. Others may have different opinions on this (I think @startz and/or @Kaysa often advocates for Python knowledge), but I would probably not pour too much time into worrying about Python/R. The more important thing in applications is to show your general coding skills: can you write clean, efficient code that someone else could understand without too much headache? I think being an expert in one language (like Stata) is the better strategy here compared to being adequate in many. Most recruiters will understand that coding skills are transferrable, and that you can pick up the nomenclature of a new language pretty easily. I would only spend time learning Python/R if it has no marginal impact on your grades in your master's program, and even then I might recommend spending that time reading papers or Mostly Harmless Econometrics instead. Or, learning to use Git (check out the awesome GUI "SourceTree"). Or, reading Shapiro/Gentzkow's "Code and Data for the Social Sciences".

That's right - if it's the same professor I had, he likes to make proofs very central to the course. And exams (if I remember correctly) are largely based on memorizing a number of proofs. Like I said, it is not an easy course: it covers a lot of material at a pretty high level. You can judge best for yourself whether you would be able to perform well int hat environment. You can find a bunch of the course materials and lecture videos online. For example, the main assistant for the course (initials KP) has many review section videos posted on YouTube that are easy to find. I recommend checking these resources over for a better sense of what to expect. Regarding the online format, I would say resources are not a big concern. Harvard Extension does a solid job of recruiting a handful of TA's for the course who hold office hours and review sessions regularly via online video chats. They also make it easy to connect online with other students in the class. If you have internet problems, sure that would be a concern. But I would say otherwise the online factor likely does not detract much if at all from the course, and it is pretty easy to be in touch with TA's or other students.

Also.. Consider taking a field course instead of the full first year sequence (especially in place of macro, if that's not your field of interest)? Taking even one of first year micro or metrics and doing well (an A) is a pretty good signal already. Beyond that, it might be worth spending your time developing your research interests. You'll cover all of first year again, anyway. You know best for you, but food for thought.