Humanomics

This is exhausting. I don't have a lot of interest in comparing people's CVs and syllabi here, in order to draw inferences about pipeline likelihoods in social science. If folks are interested in pursuing these questions formally, there is a lot of cool stuff coming out of the sociology of science. If anyone wants to learn more about the MACSS program at Chicago specifically, feel free to message me. Marshall333: You can also consider the fellowships like Data Science for Social Good, at the Computation Institute, for your gap year after EME. Google the Toyota Institute at Chicago for other opportunities if you want to get closer to computer scientists. If you're interested in this kind of work longer run, you might want to politely decline the offer in personal emails to the faculty you were interested in working with at MACSS. That can be an opportunity to explain that you have elected to ensure a strong theoretical base for your training before you go applied or try to innovate, and to express an interest in collaborating in the future. Many of these people have large pools of money across UChicago, connections to independent institutes like, eg, Sante Fe, and they have very strong reputations in their respective disciplines (Padgett, Raudenbush, and Evans specifically). It's worth keeping a channel open.

Some of the economics faculty might take your position in the MACSS program as a signal that you're interesting and creative. The trouble is with the smart signal. I'm afraid Chateauheart may be right. Faculty may be skeptical of how selective MACSS is, and/or how strong its treatment effect is. Although Professors Raudenbush and Hong are very well respected by our economists. Generally the MAPPS, CIR, and other MA programs in the Social Science Division are consolation prizes for kids who aren't good enough to get into the SSD's doctoral programs. MACSS is not taking indirect admittances in that same way, but MACSS may (unfortunately/unfairly) suffer the same stigma early on. Students who do well in MAPPS and CIR are genuine false negatives who bust their butts and compensate for the noise they suffered in initial admissions decisions. But a large portion of them do not do well and move into industry with a lot of debt and a relatively weak signal of technical competence. For the folks in the viewing audience, these are important considerations when you're thinking about paying for a stepping stone masters anywhere. In the first three years here, between 2016-2019, I would not recommend MACSS unless you: (1) know you want to use it for industry, for which you will have access to top research jobs, and very interesting and fun ones at that, or (2) if you have an incredibly good idea of the kind of computational work you want to do in an interdisciplinary social science doctoral program afterward I have a friend in George Mason's Computational Social Science program who is extremely dynamic, is giving talks at major institutions in government and the academy, publishes relatively often in major news magazines . . . and still does not expect to place well into a traditional social science department on the market.

EME has a strong record of placing economics students. Email James Evans about your plans beyond the program. He keeps relationships with top economists, statisticians, and folks from other disciplines. He'll have some idea of how the signal will read and what kind of opportunities you might have. You're likely at the top of their pile if you got into EME, so he'll want you to come. But he's very fair and will not mislead you about your chances of placing beyond the program. If you're interested in frontier quantitative methods and want to work with network scientists, statisticians, AI / machine learning / topic modelers / et al. long run, the Duncan Watts RAship and Gentzkow/Shapiro RAships are very good options after EME. Another option is to go to EME and self train in whatever ancillary methods/questions you're interested in, in order to have something to pitch to an RAship application.

Sup guys, long time no see. Hope everyone is handling the wait ok -- decisions will start rolling in soon! Stigler Center is building out and hiring an RA. Disclaimer: I can't tell how old the job posting is. https://jobopportunities.uchicago.edu/applicants/jsp/shared/position/JobDetails_css.jsp?postingId=647148

Does Economics4Life's IP address trace to Italy?

Pretty sure it means that committees are optimizing over some convex map of credentials, and that a lot of one variable can compensate for a little of another.

Is the person who gave this advice an older faculty at a lower ranked school? Ph.D. admissions and the job market used to work like this, but does not anymore given the size of the applicant markets. If you want to start a research discussion with faculty you're interested in, that's fine. It is not a requirement and only in the minority of cases will result in such a positive impression that the professor lobbies for your admission to the program.

120-130 should do it.

Bjorn Lomborg makes the kinds of arguments OP is looking for, I believe, and has caught major flack for suggesting that there are much more important human health and development goals to invest in than environmental protections, because it's say quite cheap and very important to deal with malaria which we know kills people with high probability than it is to do very expensive things to quash CO2 emissions when the probability estimates of various outcomes are all over the place (this is not a statement about the probability of whether anthropogenic climate change exists, which even most academic critics agree is 1). As far as the positive/normative distinction, I think it is a dichotomy that is a useful thing to think about, but denotes the ends of a continuous interval. The same is true of form/content, degree/kind, and a number of other famous "elemental [false] dichotomies" that scholars have always argued over. Generally I agree with chateauheart that it's not necessarily meaningful or productive to force people/groups/disciplines to one side of that interval or try to pin them down to a location within it. Economics has the advantage of having reduced the dimensions of its discussion to those chateauheart pointed out. There are costs and benefits to such institutional/organizational parsimony.

It's interesting that at the same time as the price of elite college educations has exploded, the merit of the credential it imparts (as a measure of the noisiness of the signal) has diminished.

Lol. I was referring to fashions in the sociological sense -- you know -- where you get to stretch the definition of a word to claim it is a foundational mechanism driving social behavior across a bunch of contexts. So I meant orange skinny jeans and stainless steel coffee growlers. ;)

I think that's largely true, but only within a certain range of methods and propositions. That goes for any "knowledge community" I suppose, and a lot of this thread is me just crying because there is no heat and serve discipline out there that is already tuned in to the frequency I'm broadcasting on. So I don't want to trash economists or sociologists when I'm discussing interdisciplinarity here. It's just that it is in fact increasingly difficult in the number of disparate methods and postulates you want to bring together, because you will be asking people to pay considerable prices to digest and evaluate your research. Hell, it sometimes feels like I'm asking people to pay an unusually high price to sit in a theory course with me. ;)

Honestly I think any top 10 or top 20 school is really going to give you the freedom to do what you want to do. There is considerable room for creativity in economics to be sure. Are you going to go to Harvard, pass comps, and then do an ethnographic study of the introduction of a price system for traditional goods in rural Guatemala, then ship off to a tenure track job in economics at MIT? No. But once you've cleared comps and such, there is an assumption I think that you're purebred stock and there exists greater trust for you to venture off and do what you want if you're higher status. In my (relatively small) experience talking to people, risk taking and game changing is generally not encouraged and welcomed at lower ranked departments unless they have a speciality in such things like GMU. There is a result from organizational ethnography called "middle status conformity" that proposes and inverted-U function of conformity against status. Low status people often deviate because they are new to groups and do not understand norms and routines, because they have potentially lower investment in the group relative to others and draw fewer resources from the common pool, etc. High status people often deviate because they have a stock of clout that they can spend, and frankly, they make the rules to some degree. But middle status people conform to and police norms most vehemently. That's some of where I'm deriving these conjectures about strategies in economics programs from.

You should bring that note over to Phil Mirowski during his office hours and see what he thinks about it. ;)

Well that's cool, cuz I was just making all of that up on conjecture. Lol. I've never done a matching or random study in my life, and have read probably 15 experimental papers and 2-4 matching papers in my life. Yep, agree. Does the Handbook of Experimental Economics have a good exposition of these tests of the level of randomization achieved? Have you read Delia Baldassarri's paper here? They got an enormous field sample out in Africa and were really thorough about collecting network data while running various games. https://michael-falco-p2nl.squarespace.com/2011-centralized-sanctioning-and-legitimate-authority-promote-cooperation-in-humans

I'm into conjecture land, here, because I haven't read enough RCTs and matched studies to know, but I would imagine that the reason some are arguing for matching is because people rarely get the large in the law of large numbers one needs to make randomization meaningful. Moreover, both (1) and (2) are primary concerns for observational studies, which are bread and butter. If true randomization is the gold standard, and specification of regressions can get quite messy and pose serious problems even beyond correct specification, maybe matching would fall somewhere between those extremes by requiring you to answer (1) and (2) but retaining some of the magic of neatly controlled experimental situations? What do you think of that? I think it's also important to use empirical methods that force you to go back to your theory and think about intervening mechanisms (which I think per our conversations you for example have done a great job of re: the "computer game" learning effects of many economic experiments), and matching is something that forces you to do that. Of course, there is economy in using things that reduce the number of dimensions of your empirical space, but randomization could I think become another excuse to just breeze past discussion of confounds, in the same way that significance testing did a few decades ago. I think "gold standards" in any methods discussion encourage that kind of behavior, because we start using very simple heuristics as decision criterion rather than deliberation. Just something to consider. I'm actually working on (meaning I have a manilla folder I haven't touched in a month) a proposal for a modification on a series of recent public goods experiments that I need to get your input on in a couple months, and for that I will inevitably be randomizing.

I'm not totally sure. I think Stanford and Harvard have good ones. Other posters know a lot more than me -- I didn't research or apply to political economy programs because I'm not particularly interested in poltiical institutions (not enough feelings​ involved ;) ).

It also dawns on me, that since economics is a high-prestige discipline, that has had a substantial influence on other disciplines, and does have quite a bit of research on its fringes that resembles sociology and political science, even some things that look like computer science and so forth, that it may appear whence coming from an economics undergraduate that it is not that far of a leap over to another discipline. That impression is incorrect. Switching disciplines, especially at a stage like that between undergraduate and graduate school where one has no experience working with researchers outside the home discipline, and only nominal engagement with primary research and ancillary reading outside of one's core courses, can be (and has been for me) an enormous shock. The people are different, the attitudes are different, the fashions are different, the standards and modes of debate are different, the topics of interest are different -- and that's all auxiliary to the fact that the core of other disciplines is usually miles from even the economics/other-discipline crossover type research one has likely been exposed to. For those who want to change economics from the inside, attending a traditional program and eventually branching out, or attending higher ranked business or political economy programs and still publishing in top econ field journals and generalist journals, is probably the only way to do it. Going to another discipline puts you in another discipline, full stop. You can keep your buddies in econ. You can go to the workshops. But they call one set of people economists, and another set of people political scientists, for a variety of reasons that aren't immediately visible until you've been on both sides of the fence, and it is easy to neglect these.

Nice work, though, coffeehouse and PhDPlease. This is why we in (analytical) sociology are pushing hard to get data supplements and methods descriptions published as riders at all times with studies. Nobody should have to sit around and argue about how they think the sampling might have been done. ;)

Generally we believe matched groups (the canonical example would be the twin studies that have been done in psychology) and measures of mean difference between them to be much better indicators than "random" samples. I imagine this is an artifact of how research design has actually played out in field and lab experiments, along with observational data adjustments like propensity score matching and factor analysis, because there is nothing inherently superior (as far as I can see) about matching groups along traits/dimensions of potential bias, to randomizing it away, in theory. But I bet there have appeared a lot more "randomized" assignments out there with poor control on variation than there have matched assignments that have expressed and planned control on salient and predictable confounds.

I found repetition and extension of basic principles to new applications to be extremely useful for developing intuition. To that end, I sincerely recommend heterodox-y treatments of topics that argue with the contours of standard theory. The main advantage of reading criticisms of economic theory, for me, was to learn neoclassical economic intuition really well. Many criticisms, after all, are unconvincing and can be resolved with standard economic thinking.

I really don't have the patience for the mathematics; I'm just too interested in and comfortable with thinking about other things to stay focused on it very long. I consider dropping out and getting an internship at Reason or applying to GMU on a weekly basis. But I will probably stay where I am, muscle through the first couple years, and try to seek refuge on the business school market eventually. I have two advisers at Chicago who are either pioneers in their field or considered leaders far and away, both of them sympathetic to economics and to hypotheses that aren't drawn from the set ({race}v{class}v{gender})^{oppression}. I feel pretty silly talking about my research options in terms of "where are the libertarians!?!" and "OMG the lefties are after me!" but I didn't create this situation where political dispositions correlate so strongly with research methods and theories. I'm salvaging what I can out of a difficult situation (that I clearly created myself - face palm); thanks for the support.

I love how IRBs are all good with people lying to professors to prove that there's discrimination, but won't give a similar pass on anything relating to "vulnerable" populations.

Also noteworthy: everyone discriminated in the study, making the blame of rich white men for institutional racism look pretty silly.

These studies select on the dependent variable: "if I put agents in an extremely low information environment and ask them to make predictions on one another's behavior conditioned on very few signals, will they make conditional inferences about one another based on cheaply obtained signals like large demographic categories." The answer to that question continues to be, unsurprisingly, yes. Evidence like this, about short run inferences made when people are not reasoning deliberatively and clearly are economizing on computation, get extrapolated to make claims about the long run persistence of subconscious racism in high-information interpersonal environments. (Similar claims get made about the malleability of preferences in behavioral experiments -- these are always effects measured in the short run and say nothing about the long run durability of preferences or not.) I don't see anyone running audit studies or lab experiments to see if people abandon these kinds of inferences as the cost of other signals comes down and people get to know one another better. And I'm not sure we can just shame people into disassembling their Bayesian expectations of one another's behavior. What we can do, and I think have done a fantastic job of, is convince people that the signal they're conditioning on in these types of situations is extremely poor. Race is an extremely poor signal, as is gender, of individual level characteristics, precisely because those categories include so many individuals. What we want to encourage, is for people to reduce the weight they place on these initial inferences and remain willing to pay (with time and attention) for more information about one another. Hammering on the idea that institutional subconscious racism and sexism are intractable social problems and absolutely dominate human decision making across contexts accomplishes precisely the opposite of its intent.