The death of theory?

The field is moving away from developing or carefully employing theories and instead emphasizing the testing of empirical hypotheses through some combination of quantitative or qualitative analysis. Such work is not purely inductive or atheoretical, but theory plays a relatively minor role and most of the effort goes into collecting data and trying to draw reliable causal inferences from it.

Hence the paradox: theory is the most esteemed activity in the field, yet hardly anybody wants to do it anymore. John Mearsheimer and I explore this paradox in a new paper, and argue that this shift away from theory is a mistake.

That is Stephen Walt, on the field of International Relations. He points to a similar trend in economics.

Just back from vacation, I haven’t been able to read the paper closely. I see very good points, but these questions come to mind:

  • Is theory in absolute or just relative decline? More is being published now that 20 or 30 years ago.
  • Is this just the consequence of falling technological barriers? Until 20 years ago, the statistical programs and computing power weren’t available to do large-scale empirical work. So we would expect some rebalancing, even overcompensation.
  • By this logic, the marginal gains to new empirical work should be greater than the marginal gains to new theoretical work, no?
  • Can we separate novel theory from empirics? Yes, empirical papers have a much shorter half life, but how do you reject old and generate new theory without new empirical facts?
  • I’d argue that most of the interesting new theories in economics and politics–on growth, war, economic and technical change, democratization, etc–come out of new and unexpected stylized facts. But that would be a longer post.

28 Responses

  1. Pingback: Does the micro evidence tell us anything important about development? | Chris Blattman

  2. I see the trend away from theory as a reaction to the discrediting of dominant theories such as neo-liberalism. Every science is an interplay between facts and theories but sometimes we need less theory. (I think high-energy physics is also in that state at present.)

  3. @transliminal, I think that’s a really good question. Considering research needs funding and there are more often than not other hands guiding, I’m interested in hearing about how research is chosen, and whether these older theories will come back into fashion and make a (eventual?) reappearance in one form or another.

  4. It’s an interesting trend to have picked up on. As a graduate student, not yet professionally committed to a particular approach or set of tools, it does make me wonder to what degree ideological pre-suppositions are being whitewashed, overlooked, or unwittingly perpetuated by research programs that concern themselves first and foremost with ‘hypothesis testing’ and only secondarily with theory. I agree with Blattman that new number crunching tools and big data sets now provide a whole suite of more empirically fruitful avenues of inquiry–and this is certainly a good thing. I do wonder, however, to what degree, say, World Systems Theory, Dependence Theory, etc., are being discarded as potentially useful frames for hypothesis *generation*, as an (inadvertent?) result of this shift away from ‘theory.’ I’d be interested to hear what any career development academics have to say about this. (via @IDRNUBC; http://www.idrn-ubc.org)

  5. Fukuyama makes a similar point about development economics in this article about Albert O Hirschman.

    “Development economists spend their time these days performing randomized controlled experiments, in which a particular intervention like co-payments for mosquito bed nets are introduced into one group of villages and not into another matched set. This approach establishes causality with a level of certainty approaching that of the randomized trials used in pharmaceutical testing. But while such experiments are useful for evaluating the effectiveness of certain types of public policies, they all operate at a very micro level and don’t aggregate upwards into an understanding of the broader phenomenon of development. It is hard to imagine that all the work being done under this approach will leave anything behind of a conceptual nature that people will remember fifty years from now.”

    http://blogs.the-american-interest.com/fukuyama/2013/01/06/albert-o-hirschman-1915-2012/

    I think because of the incredible complexity of social phenomena, there’s always a tradeoff in social sciences between the broadness or conceptual usefulness of a question you can ask and the rigor with which you can answer it.

    I think after for too long being at former end of the spectrum – the Washington Consensus supported by many development economists based on little empirical evidences – the professional vogue has swung to the other extreme: answering small questions with extreme rigor (usually randomization) but limited or unclear applicability outside the specific context of the experiment. I think there’s a vast amount of virgin and fertile territory between those two extremes.