As I was writing this piece about the difference between conservative Republicans and liberal Democrats in Congress and why the latter don't act in the same ways as the former, I began thinking about those members who don't represent the ideology of their districts very well. How many of them are there, and how far away are they from their voters? In particular, I thought about the case of Scott Garrett, the congressman who represents the swing district in the northern New Jersey suburbs where I grew up.
Romney beat Obama in that district by 3 points in 2012, so you'd think it would be represented by a moderate Republican. And for many years it was (with somewhat different borders prior to the post-2010 redistricting), by Marge Roukema, one of the last of the moderate, pro-choice Republicans. But Garrett votes more like he comes from Alabama than New Jersey. In 2013, he was one of only 15 House Republicans to get the American Conservative Union's "Defenders of Liberty" award for having a perfect 100 percent score on the group's ratings, along with such noted crazy people as Michele Bachmann and Steve Stockman; Garrett got a 100 in 2012 as well.
Garrett is obviously way too conservative for his district, but for some reason the local Democrats haven't been able to find a strong candidate to beat him. So I decided to see what kind of exception he is. I took the Cook Partisan Voter Index (which measures whether a district is more conservative or more liberal than the country as a whole, based on presidential election results) for every district, and compared them to the lifetime ACU ratings for every member. The ACU ratings aren't perfect or all-encompassing, but they have the advantage of being easy to understand—the most conservative member would get a 100, and the most liberal member would get a zero.
So who's most out of sync with their district? First, let's look at the Republicans. I've displayed their ACU scores, and their district's PVI:
What we have here are eight members who represent somewhat conservative districts, but are almost perfectly conservative in their voting, and two members who represent districts that lean Democratic, but also vote almost perfectly conservative. Now let's move on to the Democrats:
Again, we have some Democrats who vote consistently with their party even though they represent districts that lean slightly to the right, and others who do the same though their districts lean only slightly to the left. Even though both parties have managed through redistricting to narrow the number of truly contested seats in recent years, there are still quite a few Democrats—19 in all—who represent districts that lean Republican in their PVI. Only five Republicans represent districts that lean Democratic.
Most of those representatives understand their vulnerability and vote against their party fairly often; the exceptions are the ones on these two lists (though not Nick Rahall, who makes the list because his district is extremely conservative, and as a moderate Democrat he's still quite far from them). But what we have here are mostly representatives who are pretty much down-the-line partisans representing districts where someone from the other party would at least have a theoretical chance of winning, and in some cases a pretty good chance.
Keep that in mind for a moment. There's something else important to note: when House members vary from their districts, it's almost always in the direction of the member's own party, and not away from it. In the case of these data, that's partly because we're looking at two different kinds of measures: if you always vote conservative, you can get a perfect 100 ACU score, but there are no districts that vote 100 percent Republican (although there are a few Democratic districts that come close to being 100 percent Democrat). But I'm pretty sure it would be true if you looked at other measures as well. You might find a few Democrats who got elected in conservative districts, but what you're much less likely to find are moderate Democrats who got elected in very liberal districts, or moderate Republicans who got elected in very conservative districts.
That's because of what the primary process produces, but it's also because people who run for office tend to have done a lot of thinking about politics, have clearly developed ideologies with hard borders, and have also risen through party structures that encourage ideological consensus. If you're a pro-choice Republican when you first get the idea to run for city council, by the time you've risen through the ranks and ten years later are ready to run for Congress, chances are you've changed your position to get with the team.
So the person who works her way up to a run for Congress from a swing district is going to be a lot like the person who works her way up to a run for Congress from a district that leans hard to one party. Both of them are going to stick with their parties almost all the time once they get elected, because that's where they've come from and (probably) what they truly believe.