Sunday, June 24, 2012

Are we prognosticators?

I don't wish to provide a detailed rebuttal to Jacqueline Stevens' article in today's New York Times, as Henry Farrell has already done this extremely capably. But I do wish to focus on one of Stevens' central points: that political scientists are primarily prognosticators. As she writes,
[I]n terms of accurate political predictions (the field’s benchmark for what counts as science), my colleagues have failed spectacularly and wasted colossal amounts of time and money.
In other words, we are at our most scientific when we act as soothsayers. As Henry noted, prediction is simply not what we do. Thumb through any of our scholarly journals, and you will find almost no forecasts of future political events. We are in the explaining business, not the predicting business. The one notable exception is the round of scholarship that generally comes right before a U.S. presidential election, where a small number of us use a few different models to anticipate the outcome.

And the point of these forecasting articles is not to offer predictions so much as to test theories of how elections work. The claim being made is not "I can predict elections!" but "I can predict elections with just two variables!" There is an enormous difference between those two statements. The first claims the magical powers of a seer. The second makes an argument that elections turn on just a few key factors, and that other things that political observers tend to dwell on prior to elections (charisma, narrative, spending, etc.) may be just noise. Further, by providing those variables or explaining how they can be obtained, it invites other researchers to check their work and offer their own conclusions. The first statement is a boast, the second is scholarship. To the extent we do engage in forecasting, we are almost entirely engaged in the second form.

Obviously, I disagree with what Stevens wrote, and I find the piece disheartening on several levels. Of course, political scientists are not required to march in lockstep with their colleagues. We're free to disagree with each other -- and often do -- about the best ways to conduct our profession. My concern here, though, is that this piece will be used as a justification to cut support for political science, even though it is based on what I perceive to be a serious misinterpretation of just what our discipline does.


  1. Seth, my reaction was that this is the final death knell of Perestroika. Having succeeded in creating an alternative outlet (Perspectives), in allowing for opt-out of the APSR (coming soon), and in promoting qualitative research (multiple sections, panels at APSA), they are unwilling to declare victory. Apparently the next step is to defund the profession.

    Read Pious's comment on the Stephens piece, where he claims all the good research in Political Science has been funded by private sources.

  2. Hans et al might be too modest to point this out, but the party-elites-choose-presidential-nominees framework laid out in "The Party Decides" worked pretty well in 2012. It "worked" both in the sense that it PREDICTED the outcome ex ante (Romney had an early lead in endorsements and maintained it) and it EXPLAINED why this outcome occurred: Romney was the most conservative candidate who could rally (most of) the GOP and win. So, using Stevens' one-case-proves-a-general-claim approach, I guess I win, and the NSF can continue funding political science.

  3. Hey,Sir

    Your publishing on is a power full site for us.Now a days it site has become vary popular & demanded site for all of us.So I like it. I think not new because there are a lot of individuals have already publish it or study it, but its okay. Please observe this , in your publish there are few terms that you not modify with properly after you used Search engines convert for convert it from your mom terminology into British.As I want to share Palscience is an awesome science website that covers the latest breakthroughs in health and medicine, Nuclear medicine, technology, space, gadgets, and science

    Thanks for sharing.