prowag

I had a great deal of fun this weekend during my first visit to the quadrennial State of the Parties Conference at the Bliss Institute on the campus of the University of Akron.  I learned a great deal, had fun meeting new people, talking with old friends, and working a bit with co-authors.

I presented some work on my joint project with Ted Carmines and Mike Ensley examining the uncertainty facing American political parties given that elites divide their preferences along one ideological dimension and the people divide their attitudes along two. Our panel will be on C-SPAN in the coming weeks, probably after Thanksgiving. I will post a link when my 15 minutes of fame come.

Also on our panel were two interesting papers giving evidence for the interesting idea that dislike for the other party is a driver of partisan polarization in the electorate.  The first, from David Kimball, Bryce Summary, and Eric Vorst provided evidence that partisans are systematically likely to find that the other side makes them feel angry and afraid. Alan Abramowitz showed evidence of increasing negative affect for the other party as well; his analysis of the American National Election Study data showed that people’s placement of the distance between their own ideological position and the position of their party has not changed much but that the distance between their position and the other party has grown dramatically.

Though I missed the presentation as I was across the hall watching another panel, one of my favorite papers was one from Thad Kousser, Scott Lucas, Seth Masket, and Eric McGee that sought to deal with the particular difficult issue of estimating how much endorsements actually matter to voters. Using a unique, multi-pronged, research design that included a survey experiment and a regression discontinuity analysis of a primary election, the authors found support for the idea that endorsements do matter, to the tune of 10-15 percentage points in a primary under some conditions. Particularly interesting to me was that the survey experiment showed that the background of the candidate mattered with respect to the size and statistical significance of the post-endorsement bump; that is a novel finding about the contextual factors influencing the impact of endorsements.

Caitlin Jewett’s paper dealing with primary elections presented evidence that the Republican Party’s recent reforms to their primary elections have not achieved their goals of giving more voters voice in the nomination process.

Moving away from elections and to governing (and a bit more inside-baseball political science wonkiness), Mike Ensley, Michael Tofias, and Scott de Marchi’s paper shows that since the well-used DW-NOMINATE scores measuring the ideological position of lawmakers’ roll call votes generally operate in a way that assumes lawmakers die with their ideological boots on, some of the non-linear movement that lawmakers do is being masked. Simplifying a bit, the way that the DW-NOMINATE scores work, lawmakers don’t really change how liberal or conservative they are while serving in Congress – and if they do – the change is linear.  Ensley and friends use a procedure that shows that some lawmakers do shift their ideological position across their terms (and not always in linear ways).

I have to change flights now, but I will write more in the coming days about other interesting work presented at the conference.  Thanks to John Green for organizing yet another useful meeting assessing the state of the parties.