Physics Envy?

Political scientists Kevin Clarke and David Primo placed an op-ed in yesterday’s New York Times arguing that the social sciences need to overcome their ‘physics envy.’ I take the main point to be that we ought to value theoretical contributions even if we cannot directly test their empirical implications. Vice versa, we can learn a lot worth knowing about the world from empirical studies on important questions, even if these studies do not have important theoretical implications.

To the extent that this is the argument, I agree (and I made a similar point on these pages just a few days ago). However, I don’t really see what this has to do with ‘physics envy.’ As Clarke and Primo point out, physics has a division of labor between empiricists and theorists. Physicists value theories that cannot be tested (like string theory) and observations for which there is no good theoretical explanation. Clarke and Primo believe that social scientists are stuck in a ‘high school textbook version’ of science. But even a high school student who occasionally watches the Big Bang Theory understands this. So are social scientists dumb?

I guess I am a little confused about this supposed inferiority complex. Clarke and Primo motivate it as follows:

They often feel that their disciplines should be on a par with the “real” sciences and self-consciously model their work on them, using language (“theory,” “experiment,” “law”) evocative of physics and chemistry.

So are they saying that social science theories shouldn’t be called theories or that the experiments social scientists use aren’t really experiments?  I didn’t see an argument for that in the op-ed. And do social scientists really model their experimental work on physicists or on psychologists?

Again, I am sympathetic to creating more room for theories that cannot directly be tested as long as they teach us something important about the world. I also favor increasing our valuation of well-executed empirical work that tackles an important substantive question but that does not make a notable theoretical contribution. I just don’t think we need to beat an inferiority complex to get there.  I have to admit that I found the last sentence of the piece a bit annoying and borderline insulting:

Rather than attempt to imitate the hard sciences, social scientists would be better off doing what they do best: thinking deeply about what prompts human beings to behave the way they do.

So social scientists have been so busy imitating scientists in order to boost their self-esteem that they no longer think ‘deeply’ (whatever that means) about human behavior? Not sure I recognize this.

8 Responses to Physics Envy?

  1. Ben Stanley April 2, 2012 at 9:01 am #

    Well put. I’ve seen this article floating around social media in the last couple of days, and the issue it addresses seems fairly simple to resolve. If you make use of scientific methods to answer research questions, you are a political scientist. If you don’t (and you’re not a theorist), then you are something else – a ‘political analyst’ perhaps. I try to wear both hats, for the simple reason that there are some interesting questions for which I don’t feel scientific methods are appropriate. But either way, it’s possible to think deeply about your topic, and most politics scholars of my acquaintance – whether scientists or analysts – do precisely that. Perhaps the authors have a point about the tendency for some scholars to prize the selection of methods over the selection of interesting research questions. But it’s a point that needs proving.

  2. Matt Grossmann April 2, 2012 at 9:30 am #

    I am sympathetic to the idea that we should value theoretical contributions independent of their empirical testing. Rational choice theorists, it seems to me, are already given quite a bit of space for data-free theorizing. But I find myself with Justice Kennedy, in search of a limiting principle. How do we know that theories can “teach us something important about the world” without empirical data? Should sociological theorists also regularly publish updates to new ways to explain politics through mimicry, institutionalization, or debt slavery? Should we empower separate communities of agent-based modelers to construct different models of the world starting with distinct assumptions about political behavior? Should biological theorists create dozens of ‘just so’ theories about political behavior based on evolutionary psychology? If not empirical predictions, we may be forced to evaluate the plausibility of starting assumptions.

  3. Joe Bruns April 2, 2012 at 9:50 am #

    For example, only among social scientists would such a debate take place. One thing I do wish, however, is that social scientists be more careful in letting the data speak for itself. I’ve seen too many ‘studies,’ particularly in education research, that the data are shoehorned to fit a desired conclusion.

  4. Larry Bartels April 2, 2012 at 12:24 pm #

    There is a 200-page Oxford Press book, _A Model Discipline: Political Science and the Logic of Representations_. It is perilous to attempt to convey the argument of a 200-page book on the philosophy of science in a 750-word op-ed piece (and do New York Times readers really care?); but how many of us would resist that temptation?

    The book is mostly *not* about the value of “pure” theory, but about the limitations of “hypothetico-deductivism ” as a way of doing business. Why do we spend our time “testing” theoretical models and “rejecting” silly null hypotheses? Couldn’t we find something better to do with our summers than EITM? That sort of thing.

  5. Erik voeten April 2, 2012 at 1:47 pm #

    Larry: I understand that but as written the oped succeeds mostly in confirming tired old stereotypes about social scientists. I suspect that is what the NYT editors wanted and I agree that few readers will care about the philosophy of (social) science debate that we need to be having.

  6. Luke April 2, 2012 at 5:45 pm #

    From the perspective of economics, the book “More Heat than Light” by Philip Mirowski, argues that General Equilibrium Theory is largely derived from 19th century thermodynamics, essentially replacing heat with utility.

    Since DSGE models continue to play a large role in contemporary macroeconomics, to the extent that Mirowski’s theory holds up, the authors could have a point…

  7. Mike April 2, 2012 at 11:31 pm #

    I get Bartels’ comment above and I’ve read some of their other work, which I think is cogent, but…

    “The ideal of hypothetico-deductivism is flawed for many reasons. For one thing, it’s not even a good description of how the “hard” sciences work. It’s a high school textbook version of science, with everything messy and chaotic about scientific inquiry safely ignored.”

    This is no doubt true. But why shouldn’t we hold formal models to these very same standards?

    If I’m reading their argument correctly we are supposed to abandon the HD model – a simplified map of how science works that has definitely brought us somewhere – because it’s false and fails to adequately account for how messy the real world is. But at the same time it’s totally OK for formal modellers to veer off into more or less mystical nonsense because formal models are just simplified maps and they’re bringing us somewhere?

    I get the impression that if they evaluated rational choice with the same standards as they evaluate philosophy there wouldn’t be much left to evaluate. But perhaps it’s warranted to place higher empirical standards on philosophers of science than on scientists?

  8. Andrew Gelman April 2, 2012 at 11:39 pm #

    I think physicists have a bit of physics envy too.