In Defense of a Theory

I’m a big fan of political scientist Seth Masket, but what he wrote in defense of the theory of presidential nominations expounded in The Party Decides intrigues me. First, he notes that his initial defense caused some blowback:

Last week at Mischiefs of Faction, I wrote about the book The Party Decides and the extent to which the 2016 is causing trouble for it. I argued that the Republican race isn't providing a very clean test of the theory, while the Democratic race is. This invited substantial pushback from Jonathan Chait, Justin Grimmer, and others, suggesting that I was either protecting the theory from actual testing, or that the theory itself was untestable. So I'd like to clarify a bit and suggest just what evidence we've gathered this year.

Then, he goes on to explain more fully:

Now, in my recent post, I focused on what I see as the central claim of the book: The party generally gets what it wants. That is, when party elites have selected a favorite candidate, they're very good at making sure that candidate wins the nomination, even if public opinion doesn't favor it initially. The Democratic case struck me as a particularly good test since that party has very clearly indicated a preference for Hillary Clinton but Bernie Sanders is doing well in polling for early contests. The Republican example was, I thought, muddier, since party elites have indicated a dislike for Trump but haven't, for any number of reasons, picked a favorite.

To re-state, a Trump nomination—not a win in one or several primaries or caucuses but the actual party nomination—would be a failed test of the theory. The theory suggests that a party should not nominate someone who's so clearly a threat to what party insiders want just because he's rich and famous. But the reasons for that failure might be unclear. If the party tried to stop him and failed, that would be a much clearer refutation of the theory than the party not even trying.

This is the heart of his argument:

Finally, it's good to reflect on just what a failed test of either theory would mean. I share Christina Wolbrecht's elation that political journalism is arguing over the falsifiability of political science studies. This is very healthy! But let's think about what a Trump nomination would mean. In general, if you have a pretty good theory that confronts an inconsistent data point, you have three options:

1. Treat the data point as an outlier. It's a stochastic world, and even the best theories will occasionally have cases they can't account for. Maybe the Republicans in 2016 were a mess, but in general the theory is still right.

2. Update the theory. Maybe it's mostly right, but some recent shifts in the political system (the rise of social media, party polarization, etc.) need to be accounted for.

3. Throw out the theory. At some point, it made sense to acknowledge that the Earth wasn't at the center of the universe rather than to keep updating and complicating a theory that just wasn't explaining much anymore.

Should Trump somehow win the nomination, it's not obvious which of these three paths should be followed. Any worthwhile theory should be able to survive a bad observation. Sometimes a three-pack-a-day smoker will live to be 90; that shouldn't shake our belief that tobacco use shortens lives. Of course, the theories laid out in The Party Decides rest on relatively few data points—presidential nominations from 1980 forward—so a failed test should certainly prompt some reflection. And if we end up grouping other recent contests (Mitt Romney's nomination amid tepid early party support in 2012, Barack Obama defeating early party insider marginal favorite Clinton in 2008, etc.) as similar failures, we might come to the conclusion that this theory doesn't actually describe current party nominations all that well.

But again, we're not there quite yet. This is, after all, a theory about nominations, so it might be helpful to see who actually wins this one.

This reminds me so much of what happens when polls show an increase in “independents” in polls measuring party identification. Usually, the polls are dismissed with arguments that most “independents” are really partisans in disguise; that is, they “lean” towards one of the two parties in terms of voting behavior and political attitudes. So, let’s ignore the results.

I don’t accept that argument and can’t wait for the day when “denial” becomes impossible. That is, what are they going to do when the disgust that Americans feel for both parties results in a new party system? Was the rise of the rejection of a partisan attachment as a precursor to a realignment still going to be denied? I hope more people would understand that "the times, they are a-changing.” Among young Americans, ethnic minorities, as well as others, there is growing evidence that the parties are not representing Americans to the extent that they did in earlier eras of stable partisanship.