Someone who would prefer to remain anonymous writes:
A lot of the papers I’ve been reading that sound really interesting don’t seem to involve economics per se (e.g., https://web.archive.org/web/20070104045027/https://home.uchicago.edu/~eoster/hivbehavior.pdf), but they usually seem to come out of econ (as opposed to statistics) departments. Why is that? Is it a matter of culture? Or just because there are more economists? Or something else?
And here’s the longer version of my question.
I’ve been reading your blog for a couple of years and this post of yours, “Is an Oxford degree worth the parchment it’s printed on?”, from a month ago got me thinking about studying statistics. My background is mainly in engineering (BS CompE/Math, MS EE). Is it possible to get accepted to a good stats program with my background? I know people who have gone into econ with an engineering, but not statistics. I’ve also been reading some epidemiology papers that are really cool, so statistics seems ideal, since it’s heavily used in both econ and epidemiology, but I wonder if there’s some domain specific knowledge I’d be missing.
I’ve noticed that a lot of programs “strongly recommend” taking the GRE math subject test; is that pretty much required for someone with an unorthodox background? I’d probably have to read a topology and number theory text, and maybe a couple others to get an acceptable GRE math score, but those don’t seem too relevant to statistics (?). I’ve done that sort of thing before – I read and did all the exercises in a couple of engineering texts when I switched fields within engineering, and I could do it again, but, if given the choice, there are a other things I’d rather spend my time on.
Also, I recently ran into my old combinatorics professor, and he mentioned that he knew some people in various math departments who used combinatorics in statistics for things like experimental design. Is that sort of work purely the realm of the math departments, or does that happen in stats departments too? I loved doing combinatorics, and it would be great if I could do something in that area too.
My reply:
1. Here are a few reasons why academic economists do so much work that does not directly involve economics:
a. Economics is a large and growing field in academia, especially if you include business schools. So there are just a lot of economists out there doing work and publishing papers. They will branch out into non-economics topics sometimes.
b. Economics is also pretty open to research on non-academic topics. You don’t always see that in other fields. For example, I’ve been told that in political science, students and young faculty are often advised not to work in policy analysis.
c. Economists learn methodological tools, in particular, time series analysis and observational studies, which are useful in other empirical settings.
d. Economists are plugged in to the news media, so you might be more likely to hear about their work.
2. Here’s the syllabus for the GRE math subject test. I don’t remember any topology or number theory on the exam, but it’s possible they changed the syllabus some time during the past 40 years, also it’s not like my memory is perfect. Topology is cool—everybody should know a little bit of of topology, and even though it only very rarely arises directly in statistics, I think the abstractions of topology can help you understand all sorts of things. Number theory, yeah, I think that’s completely useless, although I could see how they’d have it on the test, because being able to answer a GRE math number theory question is probably highly correlated with understanding math more generally.
3. I am not up on the literature for combinatorics for experimental design. I doubt that there’s a lot being done in math departments in this area that has much relevance for applied statistics, but I guess there must be some complicated problems where this comes up. I too think combinatorics is fun. There probably are some interesting connections between combinatorics and statistics which I just haven’t thought about. My quick guess would be that there are connections to probability theory but not much to applied statistics.
P.S. This blog is on a lag, also sometimes we respond to questions from old emails.
Anyone with a master’s in EE likely knows more math than all but the mathiest Economists or Applied Stats people. Seriously, econ professors have likely never taken a course in ODEs, likely never in PDEs likely nothing in feedback control, likely nothing in Fourier analysis, likely nothing detailed on optimization (like numerical methods in optimization etc) of course there will be exceptions but if you look at the course requirements for econ at undergrad, masters, and PhD these things typically just aren’t there in the catalog, and the econ professors I’ve talked to admitted as much.
Applied Stats is much more likely to be full of linear algebra and a bestiary of probability distributions than functional analysis, approximation theory, orthogonal polynomials, ODEs, PDEs, integral equations, Fourier analysis, digital filtering etc.
So an EE is likely to be WAY ahead of the game math wise.
Andrew, could you elaborate on your answer to #2, specifically: “…and what does this have to do with stat Ph.D. programs?” In your opinion, is “strongly recommending” the Math subject test a good idea, i.e., is it likely to be a reliable and valid measure of aptitude for stats grad school, conditional on other qualifications? Seems to me like overkill. The quantitative part of the general GRE, plus the usual factors considered by a program (coursework, experience, interests, recommendations) ought to explain practically all relevant variance in Math GRE scores.
To the extent it does add extra predictive power, I’d guess that has more to do with screening for applicants who are most motivated to study for and take a test that’s a stretch. Maybe that’s the point? My concern here is for students with disabilities or special challenges. In my case, I love pure mathematics–literally read/do it for fun. But I also have ADHD and an anxiety disorder, so taking on the GRE in a subject I’ve not formally studied would’ve been a huge barrier. I don’t know what accomodations the GRE offers–at the time, I was undiagnosed on both counts–but just the idea would’ve been enough to drop those programs from my shortlist. Again, though, maybe that’s the point?
e. Compared to other fields, academic economists are particularly likely to be men, specifically white men who were born-middle-class. And, bless them, but they just have some real disadvantages when it comes to understanding a world they primarily experience filtered through the lenses of people exactly like themselves. And because doing economics is hard without selling your soul or making rich people angry, and because born-middle-class white men are simply better at everything than everyone else, as every fule kno, it’s obvious (to them) that they should be teaching every other field how to do their numbers rite.
I’m subtweeting* at least one real example here, with a little bit of spin. If I remember what it was/they were, I’ll come back and post the link(s).
There are many statisticians with an engineering degree. Some work in areas related to their engineering degree while others might venture into finance or clinical work. There are also engineers with no formal degree in statistics but doing statistical work-with mixed results.
I have worked with engineers daily for the past six years and at times before that. They get concepts and want to learn. Math is generally not the issue as with other types of scientists. Engineers can read formulas and graphs. The issue is attitude. They see variability as the enemy, not the parameter. They get obsessed with particular numbers, whether a cut-off or an observation.
DOE is a part of this work but most of the research in DOE seems to be about leveraging the data, not advanced combinatorics.
.
Economists study much more than money. They study institutions, policy design, human behavior, organizational design, social networks. There are several sub-fields within economics. If there is a question that involves people, there is probably someone in an economics department somewhere studying it.
Except in a very tiny number of applications, do you need ODEs, PDEs or Fourier analysis in economic; thats why it isnt part of “standard” maths for econ courses; you dont need complicated optimization techniques as most functions nused in the theory have global; maxima. In econometrics you can encounter badly behaved functions but standard techniques are more than adequate -you dont need to be a specialist: ” Seriously, econ professors have likely never taken a course in ODEs, likely never in PDEs likely nothing in feedback control, likely nothing in Fourier analysis, likely nothing detailed on optimization (like numerical methods in optimization etc) “
Well yes, especially if you just accept that theories built by people who have no idea about any of that math are good and adequate.
If, on the other hand, as an outsider looking in (as I am) you think “this is a complete dumpster fire” then the lack of any knowledge of dynamics or optimization techniques or signal processing or etc seems very very bad.
My own take is that economics is dynamics, it’s constant change, there are processes at multiple timescales, and of course some of them equilibriate quickly, like the temperature in a room, but some of them don’t ever equilibriate at all, like the temperature distribution across the globe. I think there are probably many processes which once set into motion have ultimately effects that last at least on the order of a lifetime. Things done in the great depression affect what is going on today for example. On shorter timescales, the dot-com bubble and 9/11 caused govt policy that led to a housing boom that led to a crash in 2008 which caused a massive pullback in housing construction, which caused the housing crisis and homelessness today… That’s all system-dynamics. If you aren’t modeling system dynamics using dynamic equations through time, you’re not doing (macro) economics.
+1 although I think that the disciplinary arrogance coupled with epic prediction fails (LTCM, Great Moderation, etc.etc) is what takes the cake for me. The final straw for me was getting into the climate economics literature, in part through this blog’s coverage of Richard Tol. For me that whole subfield crosses the Rubicon from annoying but ultimately ignorable to outright dangerous malpractice. It would be like if the field of civil engineering made a hard pivot to Lysenkoist mathematics while collecting accolades while socializing the risk of bridge collapses and dam failures onto everyone else. It’s malpractice. I won’t take any other branch nearly as seriously until they come for their boys in climate econ.
I guess I just want to +1 Daniel Lakeland’s comments about modeling dynamics. Folks with engineering backgrounds who bring tools like he mentioned to bear on problems are awesome people to have on the team.