Impeachment, party-switching, and political science

I have a bunch of long-form science fiction-related commentary in the works, but I want to take the time to get my thoughts in order on it.  So, here's this.

Anyway, the big Congress news is that Rep. Jeff Van Drew from New Jersey is about to switch parties.  He was one of only two Democrats to vote against the formal impeachment inquiry procedural thingamajig that Nancy Pelosi held a while back, and he'll vote against the articles of impeachment themselves.  Right now, that's a non-starter in the Democratic Party.  As Rep. Justin Amash (I-MI) was driven out of the Republican Party for supporting impeachment, Van Drew is being run out of the Democratic Party for opposing it.

Pure party line, no deviations will be accepted.

So, here's something interesting.  Timothy Nokken has a long line of research on people in Congress who switch parties.  You know what happens?  Their general voting behavior changes.  Like, a lot.  If you're looking for a citation, start with his 2000 Legislative Studies Quarterly article, "Dynamics of Congressional Loyalty."

If you want an example of what has happened since Nokken's original article... ya' know, replication 'n stuff (I'd kind of like to avoid a "replication crisis," like certain... unnamed other disciplines), let's check in on some House switcheroo-ers.  How about Parker Griffith?  Remember that little weasel?  No?  He won his first term in 2008, as a Democrat, and then decided that 2010 would be a hard year for Democrats.  So... he switched.  Er... oo... Ered.

I'm giving you data and replication.  Let me have my fun.

Anyway, the little weasel lost that GOP primary in 2010, as pathetic, little weasels should, which is not to say that the world is necessarily just, but just sayin'.  Anyway, we've got this cool thing, called "NOMINATE," which is our estimating procedure for congressional ideology.  We look at all of the votes that everyone casts, and on the basis of that, we put people on the liberal-conservative spectrum, with -1 being most liberal, and +1 being most conservative.  Basically, it's a souped-up version of factor analysis, devised by Keith Poole and Howard Rosenthal, improved incrementally over the years.

And when someone switches parties, we can conduct the estimation procedure before the party switch, and after.  Timmy-boy did that!  For every party-switcher!  Oo!  (er.)  In the case of Parker... hmmm, let me jot down his score with a pen...

Shut up.  That was funny.  (And I really like my medium nib Sonnet!)

Anyway, as a Democrat, Parker's score was -.045, by the current estimation.  Then, he did some gymnastics, and became a Republican.  Drumroll...

...

NOMINATE score as a Republican?  By the most recent estimation I downloaded*...

 .385.

That's a lot of movement.  That dragged him from dead-center to not exactly the center of the GOP, but not too far off.  I scrolled down in my STATA file to give you a reference name.  Devin Nunes!  His score in that same Congress?  .448.  So, yeah.  The dude named after my pen-- not my favorite pen, which would be the Lamy 2000, and I have yet to have a student named, "Lamy 2000," but give it time-- went from an ideology score dead center of the ideological spectrum to being spittin' distance from Devin Nunes.  I wonder what his mother and his cow thought of that?

So.  Did Parker wake up one day, and decide, you know what?  I changed my mind about everything!  Or, did party pressure do it?  Or, is it all posturing?  And how much do we infer, in the bigger picture, from the behavior of people who are not just weasels, but CAPITAL-W WEASELS?

Those are some hard questions.  The easy answer is that in some way, party matters, so get a clue, Keith Krehbiel.  (Obscure zing!)  Jeff Van Drew is about to change how he votes on a lot of things.  This is a scientific prediction, replicating existing models.

Science!

Why will this happen?  That's harder to say.  In Van Drew's case, he is under pressure from primary voters.  He has basically been shown that he can't win reelection as a Democrat if he doesn't indict a ham sandwich, sending it to trial with a rigged jury consisting of that sandwich's fixin's.  Mostly, mayonnaise, and lots of it.

So.  Much.  Mayonnaise.

What happens then?  Well, hasn't he shown what a loyal Trumpist he is, and isn't that the essence of conservatism and Republicanism today?  (I'm asking seriously, as a political scientist, now that the GOP has embraced tax increases from the tax "reform" bill, and tariffs.  Grover!  That's not even you, is it?  Pod person!  Pod person!)  Is that enough to win a GOP primary?  Maybe, if he starts breathing enough partisan fire, regardless of his votes.  See Abramowitz & Webster, negative partisanship.  Is Kevin McCarthy going to tell him he has to vote against every other bill Nancy Pelosi brings to the floor, or no committee assignments for him?  Is that it?

In my last book, Incremental Polarization, I wrote that party influence is most commonly "preference-preserving," by which I meant that parties cancel out the electoral pressure to vote insincerely.  If a Member of Congress has electoral incentives to vote against his or her sincere policy preferences, parties threaten punishment for the electorally-motivated vote, because if parties don't do that, they can't hold together and win their policy objectives.  But, that doesn't work when it comes to party switchers.  We don't even really know what they believe.

Are they anything more than opportunistic weasels?  And how much do we really learn by studying the weaselly-est of the weaselly?  I don't know.  But, I do know that Jeff Van Drew is about to start voting in a much more conservative way, on stuff completely unrelated to the reason for his party switch.

This?  This is science.  This is replication.  This is why political science works.  This is something certain... other disciplines in the social sciences don't do.

Take note.  That's why I'm bothering with this post.

Oh, and credit to Tim Nokken, obviously.



*Poole & Rosenthal update the estimations every Congress... cuz.  Complicated.

Comments