I like the Monkey Cage

The sister blog is a good place to reach a wider audience, also our co-bloggers and guests have interesting posts on important topics, but what I really like about our blog at the Washington Post is its seriousness and its political science perspective.

For better or worse, political science does not have a high profile in the news media or, I think, in the public policy world. We sometimes get quoted on some technical matters regarding campaigns, polls, and elections, but, other than that, political scientists’ inputs into the public discourse are sporadic. Sure, there was “bowling alone,” and “broken windows,” and “the clash of civilizations,” but I don’t get a sense that, when political issues come up, political scientists are the go-to sources for explanation and commentary.

That might be just fine. Maybe the contributions of political science are pretty much irrelevant to politics. It’s not my place to say that news coverage of these issues should change.

But one thing this all does mean, is that when political scientists do enter the discussion, it can be in a weird, contextless way, not representative of the field of political science and not necessarily making much sense at all.

One problem, I suspect, is that because poli sci is not on journalists’ radar screen, they have no way of evaluating the input they get from political scientists. It’s as if they were reviewing literature in translation from some distant northern land.

I was thinking about this the other day after coming across a horrible op-ed in the New York Times the other day written by someone named Gerard Alexander who is labeled as a political scientist.

I have nothing against Gerard Alexander either as a person or a political scientist. I did not happen to be familiar with his work but I see on Google that it’s been cited many times, and it might be excellent. He perhaps was having a bad day when writing that op-ed.

But that the editors at the NYT had no way of evaluating this piece. It was by a political scientist, and . . . who knows, really?

I took the opportunity on the Monkey Cage to extract something useful (I hope) out of Alexander’s op-ed, as to me it was an interesting, if unwitting, example of political polarization.

And this brings me to the point of today’s post. In the Monkey Cage, political science is presented in context. I’m not saying every post on the Cage is perfect, far from it. Some recent Monkey Cage claims that I’ve publicly disagreed with include “Liberals smell better to other liberals than to conservatives,” “Here’s how a cartoon smiley face punched a big hole in democratic theory,” and “Could non-citizens decide the November election?” But that’s fine: research is messy, and we also have room to give second opinions. But there is a context, there’s a steady stream of political science, rather than what we see in the NYT op-ed page, which is random things that happen to pop in.

I’m hoping the Monkey Cage will have a positive impact, not just at the Washington Post, but also at other media outlets, as editors start to sense that political science is a field of research and scholarship and not just an excuse for people to polemicize.

37 thoughts on “I like the Monkey Cage

  1. This is an interesting comment about political science and it’s broader impacts. I come at this as a political scientist, myself. I have noticed a similar dynamic in the area that I study, education policy and politics. In that domain, it is interesting that economists seem to really dominate the narrative when it comes to debates over policy change—things like debates over methods for teacher and principal evaluation, school accountability, etc. That can be a problem because although economists have a lot of experience and tools that the can bring to the task of trying to measure these things, I think they are really fish out of water in making policy recommendations because they have very little grasp on how policies are implemented in practice. Knowing that we can measure teacher value added in some sort of econometric analysis—Economist: Hey, with our model we can identify the “best” teachers! Political scientist: How are school districts and principals likely to interpret and use those results to improve practice? Economist: …crickets…—does not mean that those evaluation methods, when adopted into law or regulation and implemented on the ground in the nation’s 14,000 school districts will actually work as intended. Political scientists have a ton of great ideas about implementation, but (and I blame ourselves for this) we have done a bad job of contributing those ideas to larger policy debates.

  2. Imagine if some political scientist came along and made such profound progress that centuries from now they’d be looked on as a Newton of political science. That is to say, this individual permanently and dramatically changed political science for the better and was of great benefit to mankind.

    What would they have done? What’s the big achievement political science is working towards?

    P.s. Political Science should have kept the name “civics”. Nobody respects two word subjects, especially if one of those words is “science”.

    P.p.s. The only difference I can see between Political Science and Economics is which wrong theories they ask you about on their respective qualifying exams.

    • Perhaps you should read a bit more broadly, Anonymous. These folks clearly have made important contributions: Herbert Simon, Elinor Ostrom, Robert Axelrod, Robert Putnam, Gary King, etc . . . That’s not just my opinion. Check the citation counts.

      • “Citation counts” reflect academic interest. But as a general measure of broad usefullness what are the big contributions of Poli. Sci.

        Genuinely curious, not trying to be sarcastic.

        Like, what are some big ideas from Poli. Sci. that have had impact beyond academia. Something that has helped the average voter or changed our processes.

        • Rahul:

          Setting aside big ideas for the moment, I think the posts on the Monkey Cage have a higher level of reasonableness (in favoring common sense over grabby contrarianism, and in a generally data-based and analytical attitude) than, say, op-eds in the NYT or articles on politics in Slate or whatever. That’s just my guess—I haven’t done a thorough study—but I think that political science promotes a useful seriousness about political issues. The field of poli sci does have its biases, but overall I think it provides useful frameworks about thinking about politics, and I’m often frustrated with journalistic treatments which can seem to start from scratch each time. To the extent that even if a political scientist appears on the NYT op-ed page, it can be for something that lacks insight. What I like about the Monkey Cage is that it has a certain grounding that comes from being an academic field. Again, not always—I’m still annoyed about that smiley-face post, that was something that would be more in place at Psychological Science—but no institution is perfect.

        • Andrew:

          Of course, reasonableness is great. And grabby contrarinism bad. And I’m sure there’s tons of good work in Poli. Sci. that fits that bill.

          The part I worry about is usefulness. i.e. It is perfectly possible to build an elegant multi-level, heirarchichal model with no crappy NHST & disseminate reasonable, replicable, non-hyped conclusions that meet the best standards of scientific rigor *but* everything about an issue that hardly matters.

          It is that sort of thing that I worry about in Poli. Sci. It may have too many, honest, scrupulous, talented academics producing work of marginal utility.

          I agree that utility / impact etc. are hugely subjective. But IMO Poli. Sci. consists of too many people working to produce too little of real impact. Where are the big ideas? Where’s the impact on political processes or voters?

        • Rahul:

          My papers are here and here. Some political science papers that I think are about issues that matter include:

          The Great Society, Reagan’s revolution, and generations of presidential voting

          The mythical swing voter

          Moderation in the pursuit of moderation is no vice: The clear but limited advantages to being a moderate for Congressional elections

          Understanding persuasion and activation in presidential campaigns: The random walk and mean-reversion models

          Economic divisions and political polarization in red and blue America

          Segregation in social networks based on acquaintanceship and trust

          Public opinion on health care reform

          Bayesian combination of state polls and election forecasts

          Income inequality and partisan voting in the United States

          Partisans without constraint: political polarization and trends in American public opinion

          Rich state, poor state, red state, blue state: What’s the matter with Connecticut?

          Voting as a rational choice: why and how people vote to improve the well-being of others

          An analysis of the NYPD’s stop-and-frisk policy in the context of claims of racial bias

          A broken system: the persistent pattern of reversals of death sentences in the United States

          Standard voting power indexes don’t work: an empirical analysis

          Empirically evaluating the electoral college

          Enhancing democracy through legislative redistricting

          Why are American Presidential election campaign polls so variable when votes are so predictable?

          Systemic consequences of incumbency advantage in U.S. House elections

          Is this work only of marginal utility? Maybe I’m not the best to judge. I’m not curing cancer here. On the other hand, policy on cancer and other issues is made in large part by our elected government, and I think it’s valuable to understand our political system in all its forms.

        • Personally, I think your fantastic work on Radon sampling or Arsenic in B’desh well water & all your Methods work has far greater *impact* than any of these Poli. Sci. papers. I guess I’m biased.

        • Rahul, one might reasonably argue that no piece of knowledge on governments and voters might have the impact, especially if limited in its dissemination, would have the kind of impact as papers on radon or arsenic. But, that doesn’t mean something worthwhile wouldn’t be lost if we abandoned any effort to gain such knowledge.

          A lot of people are very concerned about polarization and partisanship. It is arguably better to have some knowledge than be in the dark about such things. For instance, commentators assert again and again that things like open primaries and gerrymandering reforms should can alleviate these issues. But, research actually shows these factors play at most a minor role.

          Many commentators went totally apeshit when Jeremy Corbyn was elected for Labour, but research on the median voter theorem and retrospective voting helps make sense of what is actually going on.

          Understanding voting and campaigns is meaningful and worthwhile work, I think, and political scientists have made progress on this stuff. Their impact definitely is lesser that of natural scientists, activists, and policymakers, but they are shedding light on important questions.

          I suppose that political scientists could amass compelling evidence for theories that could explain the links between the electorate, policymakers, and events better and make a bigger impact on discourse and debates. But, the latter problem may reflect the insularity of academia and its consequences more so than a reflection of meaningless work.

          Also, with respect to another commentator in another thread, the Analyst institute hires a lot of political scientists and Daron Shaw worked for Bush’s campaigns in 2000 and 2004. Bush managed to win despite strong growth and Presidential approval for President Clinton.

          James Robinson co-wrote “Why Nations Fail” on the impact of inclusive social institutions on sustainable economic growth. Chris Blatt has done a lot on cash transfers which are improving the quality of life for the some of the world’s poorest people. Martin Gilens is doing work on inequality in political representation which is at the heart of having a democracy. Branham, Soroka, and Wlezien have an interesting response to that.

        • “one might reasonably argue that no piece of knowledge on governments and voters might have the impact, especially if limited in its dissemination, would have the kind of impact as papers on radon or arsenic.”

          I don’t know about that. Ending Crony Capitalism for example would seem to have about as a big an impact as curing cancer. I could think of other things too.

          The sad fact is though, no matter how dismal a field is, the paid hacks who make a living of it will look around, rank everyone’s work, and announce to the world the highest ranks were “impactful and genius”. Just watch to Oscars or look at Economics to see what I mean.

        • @Nadia Hassan

          Just to clarify, I’m not at all arguing for any abandonment of effort.

          I’m merely wondering if there’s too many people working here producing too little impact.

          Indeed, the median voter theorem offers fantastic insight but were Hotelling, Black, Downs from Poli Sci. Departments?

        • Because political science has an incredibly broad scope, it’s difficult to list off all major contributions in the field. So I’ll limit my discussion to what I know best, U.S. politics.

          Keith Poole and Howard Rosenthal developed a roll call scaling method — NOMINATE — to estimate the ideological positions of Congress members. This allowed them to develop an ideological history of the House and Senate and to determine the ideological distance between the two parties over time. They have found that congressional Democrats and Republicans are further apart today than they have been since the Reconstruction. We are living in a period of unprecedented polarization, in other words. This finding helps explain why we see such fierce battles in DC today, and why Obama could not usher in a post-partisan era. Moreover, we can expect that subsequent presidents will be no more successful in fostering bipartisan coalitions.

          Poole and Rosenthal’s NOMINATE scores have been used in numerous studies. The obvious, big question is, What’s the cause of partisan polarization? Working with Nolan McCarty, Poole and Rosenthal have found an extremely strong relationship between increasing income inequality and partisan polarization in Congress. The two parties, driven by the widening gap between rich and poor in the United States, are motivated to serve radically different socio-economic groups. By doing so, they only polarize further.

          This research has only raised more questions. Does partisan gerrymandering encourage greater polarization? The answer is no. Is the electorate as polarized as the legislators who represent them? The answer, again, is no, indicating that a serious breakdown in democratic representation has occurred. But research, based on decades of polling data, also finds that certain subgroups within the electorate — the most active, most knowledgeable voters — have adopted increasingly extreme policy preferences. So politicians are apparently responsive to voters who are most likely to donate money and participate in elections. Finally, congressional research has found that the majority parties now pass legislation that moves policy away from the median voter and toward an ideological pole. So we have increasingly extreme legislators, passing increasingly extreme bills that serve increasingly extreme and relatively narrow constituencies.

          I have just touched the surface of political science contributions. The field is massive; most researchers look beyond the United States and conduct research with an international or cross-national focus.

        • I could sit down in 5 minutes and come up with a consistency measure for roll call voting, as could any competent statistician and probably most political scientists. It’s an indication of the poverty of the field that this is given as an example of a contribution. How about not-obvious insights into the political process?

        • That’s a fair point about citation counts, Rahul. As far as broader practical usefulness, the speeches and other documents supporting the Nobel Prizes won by Herbert Simon and Elinor Ostrom suggest broad applicability of their ideas. Axelrod’s path breaking work on the evolution of cooperation has huge substantive relevance for policy responses to conflict or violence. Although not the original author of the concept, Putnam’s explorations of “social capital” have informed much work at national, state, and local levels by agencies of government and nonprofits and other groups outside it. Finally, the suite of statistical methods that King has developed, often growing out of political science applications, are used by dozens (hundreds? thousands?) of researchers and policy analysts around the world each day.

  3. as editors start to sense that political science is a field of research and scholarship and not just an excuse for people to polemicize.

    One has to wonder how much political science has to offer when someone untrained (Nate Silver) can come in with a cursory knowledge and put together the one website that outperforms what any other political scientists does regarding a campaign? Or how the Obama campaign hired technical people rather than political science people for their field operations. Or “predictors” for presidential elections for which factors continually have to be added/modified after each election and which aren’t terribly accurate in any case and completely ignore demographic change. Related to the last, I’ve already mentioned that quantitative political scientists, with their issue spaces and modelling, have completely missed the racial aspects of American presidential campaigns for the last half century. Citation counts aren’t particularly relevant when they are basically incestuous with no real application outside the academic field.

    • +1 for citation counts. It’s the same set of people who write journal articles, review them, read them, cite articles etc.

      Those counts are hardly a good metric of broader usefulness. Neither necessary nor sufficient.

      PS. How many citations did Daryl Bem get?

      • Brin advised the Obama people to hire programmer types rather than polysci types. A simple algorithm for support (probability of registration times probability of turnout times probably of support) was the unit of analysis, along with constant experimentation.
        I would suggest that Hartman (the subject of the puff piece you link to above) was successful because of her Caltech education, rather than any political science background. Note from the article:

        Hartman, with no prior polling experience, was hired as the chief survey methodologist for the analytics department of Barack Obama’s 2012 re-election campaign. Within nine months, she’d built one of the largest polling operations of that election cycle—one that accurately predicted how the election would turn out.

        “It was my background in statistical design,” she says, “and lack of strong prior beliefs on how to properly conduct surveys that allowed me to completely revamp the polling process from the ground up.”

        The obvious question is why someone run a successful polling operation without prior polling experience or study in polling in political science, since political scientists have been polling for 60 years. Wouldn’t this experience be helpful or even necessary? Evidently not.

        As further evidence for my assertions, consider

        http://www.technologyreview.com/featuredstory/509026/how-obamas-team-used-big-data-to-rally-voters/

        The head of analytics for Obama in 2012 was a someone (Dan Wagner) who had done polling for Obama 2008 and DNC since then, fairly successfully (he’s not a political scientist). Hartman was under him and while I’m sure her contributions were positive they would have been done without her. But the article notes:

        After the voters returned Obama to office for a second term, his campaign became celebrated for its use of technology—much of it developed by an unusual team of coders and engineers—that redefined how individuals could use the Web, social media, and smartphones to participate in the political process.

        The key is coders and engineers. I will also state that what Obama did was not conceptually difficult (even political scientists could understand it) but the difficult was in the database management and integration. Having a billion dollars helped a lot here, aside from getting non-political scientists.

        • If you go the no true Scotsman route, you will always be correct.

          A key here was Hartman’s “lack of strong prior beliefs.” Wagner is not a political scientist, but he’s not a coder or engineer either. Both Wagner and Hartman brought substantive expertise to the mix, and Hartman acquired some of that expertise in her PhD program at Berkeley.

          After the 2012, Wagner and other started Civis Analytics. Hartman and others started Blue Labs. Both firms employ statisticians, coders, engineers, and political scientists.

    • Numeric, I think you’re misunderstanding what political science is for and how it influenced the things you mention. It’s not necessarily in a political scientist’s wheelhouse to be running day-to-day polling forecasts or to be running presidential campaigns, at least no more than an academic physicist gets involved in designing the latest type of toothpaste. However, Silver has drawn extensively on political science in figuring out how to aggregate polls (including some of Gelman’s work) and what other factors to use earlier in the election period. Similarly, political campaigns have used a lot of techniques derived from political science, especially on voter contact strategies and the use of RCTs. The last example is completely off — the role of race in political campaigns has been extensively studied for 50 years.

  4. Numeric, I think you’re misunderstanding what political science is for and how it influenced the things you mention. It’s not necessarily in a political scientist’s wheelhouse to be running day-to-day polling forecasts or to be running presidential campaigns, at least no more than an academic physicist gets involved in designing the latest type of toothpaste. However, Silver has drawn extensively on political science in figuring out how to aggregate polls (including some of Gelman’s work) and what other factors to use earlier in the election period. Similarly, political campaigns have used a lot of techniques derived from political science, especially on voter contact strategies and the use of RCTs. The last example is completely off — the role of race in political campaigns has been extensively studied for 50 years.

    • The purpose of political science faculties is to teach pre-law students, and others who want a semi-respectable major (interesting fact–Colin Powell at CUNY had a geology major because it was “respectable”, and not too hard. Political science, like most liberal arts, was not “respectable”).
      Gelman is a statistician with a taste for slumming (probably because he’s verbal). Factors have been known since V.O. Key (the standing decision) and then minor perturbations which relate to war and the economy. Much effort has gone into describing these perturbations but aside from fairly obvious signs (war sucks particularly after the first “rally around the flag” phenomenon and a bad economy hurts, but there’s different models of how it hurts and they change from election to election and it’s all statistically indistinguishable, though of course different models aren’t tested against one another because it would show, well, they are indistinguishable). I remember when people used to predict Congressional seat changes in off-years before an election but they don’t do that anymore because none of the models really worked for the last three off-years–now presidential predictions are all the rage but there only accurate +-5 points and we have to go back to 1984 to find one greater than that. And what RCT did the Obama campaign (with the analytics run by a non-political scientist–see the link above and Dan Wagner) use? And regarding race, what RCT model (presidential vote on the LHS) actually uses race on the RHS?

    • Designing toothpaste is firmly grounded in concepts & equations that we owe to physicists, chemists etc. for their origins & development. No pharmacist is going to deny the strong influence of physics & chemistry on his working craft.

      Can we say the same about the applied practitioners of polls, forecasts & election campaigns? Is their practice strongly dependent on Poli. Sci. contributions?

Leave a Reply to Andrew Cancel reply

Your email address will not be published. Required fields are marked *