Justin Wolfers on trade across the disciplines

What is interesting to think about are the terms of trade between economics and all these other disciplines. We are clearly a net exporter to political science and sociology. But at this point the trade with psychology is almost all one way. We are a near-complete importer. I wonder why we haven’t been bigger exporters to psychology. I think it has to do with the research method. Like political scientists and sociologists, economists are almost all about the analysis of observational data. And then there are second-order differences. Formal political scientists write down a model before they observe data; informal ones don’t. Ethnographers observe four people; survey researchers observe 4,000. But it’s all observational. But when I watch and speak with my friends in psychology, very little of their work is about analyzing observational data. It’s about experiments, real experiments, with very interesting interventions. So they have a different method of trying to isolate causation. I am certain that we have an enormous amount to learn from them. But I am curious why we have not been able to convince them of the importance of careful analysis of observational data.

Here is the much longer interview.  I would cite an additional factor.  We as economists can export models to political science and sociology and at least pretend that maybe those models work.  Even if they don't, someone can publish by knocking those models down.  When we try to export models to psychology, it's too obvious, too early on, that our models are limited in their ability to deal with context-dependent phenomena.


Doesn't economics justify its unrealistic foundations with the macroscopic "fruitfulness" criteria? Methodology of Positive Economics and so on? So naturally it exports to disciplines focusing on generating "fruitful" conclusions, while importing from disciplines with actually realistic microfoundations. It's all about the comparative advantage...

A few of us in the econ dept talk to people in the stats department, but there isn't as much dialog as I'd like. My friend in the stats department always jokes that when his students are having a hard time finding a topic for a paper he suggests they go over to the psychology department, pick a paper at random, then simply correct all the mistakes in their statistical methodology. I think his exact words were, "Even a first year stats grad student could find two or three major errors in every paper they write."

I don't know anyone in the psychology department, so I don't know how much truth there is in this, or if its all just teasing. I have no idea what they teach people over there. I just assumed that they're equally concerned with issues like selection bias, censored data, measurement error, endogeneity, etc. and have the same toolbox of models as we do regarding measurement of treatment effects in all their fully, semi, and non-parametric glory.

Is this correct, or is my stats friend right in that they commonly overlook things a statistician or econometrician would classify as an elementary mistake?

At the core of today's economics ismthe import of a bogus picture of "science" and knowledge taken from philosophers and physicists.

You can track it -- Mach and Mill and instrumentalsim and operationalsim and other forms of fake physical science taken up by Schumpeter and Samuelson and Friedman and the econometricians -- you can point to specific places where they explicitly acjpknowledge their debt to these bogus pictures of science and knowledge andmusemit to justify their reseach program.

Need I rehearse the fact that the foundations and scientific justification of most of tofpday's economics remains in tatters, and remains an ebarrassment to science -- as many leading sciences in the legitimate sciences have said?

You have failed to shiw how a pure math construct provides causal explanations of empirical phenomena, you have yet to expalin how one good models or representative agent models have anything to fdo with the world, you have yet to justify much of econometric practice, macro practice, etc., etc.

And all of these research programs are in debt to a bogus borrowing from failed philosophical pictures of science and knowledge for their justification.

"Worse, their models, assumptions and simplifications are sometimes caricatural, for example dividing individuals into 'rich' and 'not rich' as we have seen very recently, and assign specific behaviors to these classes of individuals."

It's not a caricature to conduct research that presumes people with different wealths act differently, because they do. If the behavior is arbitrarily assigned, that would be inappropriate, but that isn't what mainstream economics does. Do you have an example of research that does this?

"Economists may well perform "careful analysis of observational data", but the careful seems to be in the data collection, and not in the analysis. Fitting the models and deciding what interactions and complexity to leave out, seems to be largely informed by political fashions and personal bias."

Are you talking about macro or applied micro? In macro, the models are calibrated, and are often not statistical in nature. In micro, they argue plenty about what interactions to include, robustness, endogeneity, and whether IV or treatment effects or randomized trials is the best approach.

You are probably right that there is more empirical science going on than my caricature. Also, I was describing how macroeconomics appears to this relatively uninformed outsider - after all macroeconomists impact government policy and therefore my life, directly. I should have made this distinction.

But, please, then also address my central argument, that macroeconomic assumptions and simplifications verge on the caricatural and will never explain such complex systems as 'the economy'.

Well Justin, spot on.

Only lab experimentation on specific theories will bring fundamental progress to economics.

BTW: Already 60 years ago Ludwig Mises's (often misunderstood) Praxelogy recommended to economists not to get sucked into the observation trap.

Rgds from Austria.

Maybe it's because sociology and polysci have even bigger BS quotients than economics. On the plus side, econ did export Malthus's idea to biology.

I note the relationship between this post and one back a few on the value of a liberal arts education.

Instead of reading Ayn Rand, economists need to read far more widely to gain a more realistic understanding of human nature. For Greenspan to believe fraud didn't exist in a free market so no regulation was needed shows an amazing ignorance of human nature.

Actual experience with science, introducing the scientific method, setting up experiments, and understanding few experiments take place in a lab. Perhaps taking some archeology or astronomy to understand the methods of going back in time. Some anthropology to understand the interrelationship between culture, climate, technology, resources, resource depletion, competition.

In fact, anthropology is a great example of a discipline that must integrate all the sciences.

Hayek's paper on Scientism would appear to me to be of high importance here. HE said in one of his UCLA interviews later in his life that he viewed Friedman's Essays on Positive Economics as perhaps more dangerous than any of the socialist works he was engaging with in the calculation debates.

I am sad that we no longer teach history of thought, and that methodological questions only come up when someone else understands enough to criticize us.

Hopefully somebody has some answers, things can get so depressing. Remember when the took the codeine out of NyQuil?

rajiani -- I'm typing on an Ipad.

Hayek's brain theory / psychology picture is at the heart of Joaquin Fuster's global brain theory and has been admired by the top psychologists, cognitive scientists, and neuroscientists for 3 generations. See:


“The main reasons for dwelling .. on Hayek’s model is simply that it has certain properties, absent from most others, that conform exceptionally well to recent neurobiological evidence on memory and that make it particularly suited to the current discourse.†  (Joaquin Fuster, Memory in the Cerebral Cortex:  An Empirical Approach to Neural Networks in the Human and Nonhuman Primate.  Cambridge:  MIT Press, 1995, p. 89)

Right, because they are false.

Nothing I've said is false.

This is a difference that matters.

I'd actually give econ a bit more credit than Wolfers. There are some really interesting resource models of cognitive control and attention that while they aren't quite ready to replace the dominate models from cognitive psychology, they have quite a bit of rigor and should be able to provide some testable hypothesis about cognitive processes to a level that isn't possible with current models.

However, about psychology's need for 'careful observations', I'd say observation is never as good as experimentation. Controlling variables is always important, and experiments just simply allow you to control more than observation. It's the careful part where psychologists can learn from economists. Economists, because of their hap-hazard data sets, have been forced to be very explicit about their assumptions, and trying to actively search for extremely good controls to dissociate the effects they see from closely related phenomenon. This is something psychologists could learn a lot from.

I deny that I am Jonah Thomas. So, yes, I deny I deny that I'm a cretin.

I'm surprised no has mentioned Paul Glimcher's Neuroeconomics here. He executes an impressive program of predicting experimental data on response times using economic optimization and game theory principles, simultaneously showing the weakness of mechanical reflex explanations.

Kuhn and Mayr and Hayek Hull and many others point out that this idea of "the scientific method" is purely bogus.


Example, it doesn't apply to Darwinian biology, an explanatory achievement which has done more to unify and change our understanding than anything you can point to which uses the (most bogos) "scientific method".

I don't understand what your objection is here.

Scientific method is a robust way to test hypotheses. It is not a way to create hypotheses.

Biologists often use scientific method to test evolutionary hypotheses. So what's the problem?

There are some issues that have not particularly been resolved. Darwin did not understand genetics and made some reasonable assumptions about it that were wrong. Nobody pays attention to that any more except for creationists. Now we have a much more thorough understanding, with medelism, and meiotic drive, and transposible elements, and genes that get pieced together from multiple DNA sequences with variable-length sections cut out, so that the same DNA can make multiple proteins depending on circumstance. It's a big complicated mess and the consequences are just beginning to get sorted out. But the hypotheses are mostly testable, in theory and increasingly in practice.

Darwin had ideas about how new species formed. Evolutionary geneticists came up with lots of other ideass based on the evidence of similar species, but we haven't seen enough species form to be real comfortable with any of them. Speciation is still kind of up in the air, unless there's been a breakthrough I haven't heard about. At some point we're likely to get testable hypotheses about speciation, and not just about some of the possible mechanisms for speciation.

Evolutionary thinking is a way to create hypotheses. Scientific method is a way to test hypotheses. The two are entirely compatible.

Greg, I have no idea what you are talking about. Mechanical procedure?

I repeat, scientific method is about ways to test hypotheses, and does not constrain ways to create hypotheses. Evolutionary thinking is about creating hypotheses. There is nothing incompatible there.

It happens sometimes that evolutionists come up with a hypothesis that is so compelling they believe it must be true without testing. This happens particularly with sociobiologists. And of course some economists do this a lot, particularly austrians. That can be a lot of fun. No need to confuse it with science, though.

Comments for this post are closed