Ideology


We’ve just released the July 2014 update of our state legislative ideology data set. Both legislator-level and state-level data are included, so you can use these scores for a number of different applications.

Big changes include:

  1. Expansion of the quantity of the data by 25% at the state level, and 11% at the individual level.
  2. Expansion of the breadth of the data to include 1993-2013.
  3. Filling in holes of the coverage even before 2012.
  4. Extensive cleaning and merging operations to minimize the effect of random noise.

Head over to the data site to check it out!

 

I have a new post over at the Monkey Cage blog at the Washington Post. It summarizes some of the posts I’ve made here, for an audience that probably hasn’t seen this blog before.

The one update is that the posts here used the most recent publicly available version of our aggregate state legislative data. That data includes the years up through 2011. The Monkey Cage post uses a preliminary update to our data set, that brings the data to 2013.

We are working on polishing this data for public release shortly.

In the meantime, here are the new plots using that data set. The graph below shows legislative polarization in each state, averaging across all the years of our data (approximately 1996-2013) and across both legislative chambers. Polarization is defined as the average ideological distance between the median Democrat and Republican in the state legislature.  Larger numbers indicate more division. The dashed line is the level of congressional polarization, included as a comparison (“US”).

state_polarization

About half of the states are even more polarized than Congress—which is saying a lot. At the same time, some states–like Louisiana, Delaware, and Rhode Island–have  relatively less polarized state legislatures. In Louisiana, both parties are fairly conservative, and in Delaware and Rhode Island, they are both fairly liberal.

One state that stands out is California. It is incredibly polarized. (And its most recent primary and redistricting reforms look unlikely to reduce polarization.) Unlike Congress, however, Democrats both dominate the state so thoroughly and no longer need to attain supermajorities to pass budgets, so this polarization is not as much of an obstacle to actual lawmaking in the California state legislature.

Another state that stands out is Wisconsin–the site of massive protests in 2011, a recall campaign against sitting governor Scott Walker, and even a physical fight between Republican and Democratic justices on its state Supreme Court.  It is perhaps no surprise that Wisconsin too is highly polarized.

Not only are states polarized, that polarization has increased over time. The graph below breaks down the trends in the ideology of Democrats and Republicans (measured by party medians) over time and across all 50 states. By convention, more positive scores represent more conservative preferences, and more negative scores represent liberal preferences. (One side-note: a data error exists in Washington State around the year 2000 and is being fixed.)

party_years

Most states have polarized over the past 20 years or so, but some more than others. Arizona, California, and Colorado are polarizing very fast. Nebraska—a state without formal political parties in its legislature—is polarizing very quickly too, though from a relatively low base.

Moreover, we are seeing asymmetric polarization, just as in Congress. Republicans have been getting more extreme faster than Democrats in more state legislative chambers, but this is by no means universally true across all states.

All in all, the picture we see in state legislatures is similar in many respects to Congress, but different in key points. The parties are pretty far apart on average, but that difference varies across the states. The parties are increasingly polarizing over time, but that too varies across state. Finally, we see cases of symmetric and asymmetric polarization. These new data on polarization at the state level—and the uneven pace of polarization across states—should help pundits and scholars figure out what’s driving polarization in our statehouses.

Over at American Legislatures, I show that, on the whole, polarization is asymmetric in the state legislatures. Republicans are polarizing faster across more states than Democrats. Examples include Tennessee and Colorado. But in lots of states, the patterns are different, where Democrats are leading the charge to extremism (eg, Idaho, Mississippi, and California). And both parties are polarizing roughly equally and simultaneously in places like Texas, Missouri, and Nebraska.

Read the whole post but here’s the key plot:

party_chamber_years

I have a new post about polarization trends in state legislative chambers across the country in the sister blog. Go there for the full details.

Here’s a little peek at the key plot.

polarization_chamber_years

Most state legislative chambers are polarizing, but a number are stable and a few are even going the other way.

Over at American Legislatures, I have a new post about Governor Chris Cristie’s choices for a replacement nomination for US Senate in New Jersey. Three of the top choices from the state legislature–Tom Keane, Joe Kyrillos, and Jon Bramnick–are current state legislators who are near the center of their party in the state (with Bramnick slightly more to the left).

The punchline, though, is that New Jersey Republicans are amongst the most moderate in the country:

npat_boxplot_states_parties_nj

I estimate that they would probably vote like Maine Senators Olympia Snowe and Susan Collins once in Congress.

 

I just posted an updated visualization of state legislative polarization over at the Measuring American Legislatures blog.

Here’s a small version you can look at, but see the full post and explanation here. Look at California at the top with massive polarization, and Louisiana and Rhode Island at the bottom with relatively small amounts of partisan division.

state_polarization_mcmc_1996-2011

 

 

My coauthor (Nolan McCarty) and I are releasing a new version of our state and chamber-level aggregate data. We have focused on two major updates:

  1. In all, we have 140 chamber-years of new data. These now include party data for Nebraska thanks to friend and coauthor Seth Masket, who generously provided the informal but well-known partisan affiliations for Unicameral legislators.
  2. The individual level data underlying this release has been extensively cleaned to minimize the random noise inherent in acquiring roll call votes from printed journals.

You can find the data here.

I recently posted the graph of my estimates of the two parties’ congressional candidates. In that post, I wanted to emphasize that moderation still exists, even in this polarized age. To highlight that point and make the plots prettier, I smoothed out the distributions.

However, that smoothing hid another very interesting take-home point from the 2012 candidate scores. There appears to be evidence of bimodality (two peaks) not only across the parties—that’s good old polarization—but also within the parties. Here are the unsmoothed plots that make that clear:

cands_house2012

cands_senate2012

No, those aren’t Halloween ghosts. It looks like both parties have two distinct wings, a moderate one and an extreme one. This visual inspection is backed up by test statistics from the Hartigan dip test for unimodality.

Feel free to download the estimates for all the 2012 congressional candidates here. The explanation of how I generated them is here.

We haven’t seen this before in roll call-based ideal point estimates, and I don’t think I’ve seen it before in previous years’ survey estimates (this is something I need to go back and check). So this could be something new under the political sun.

What could be causing this? Perhaps new electoral forces like the Tea Party on the right and Occupy Wall Street on the left are forcing candidates to pay lip service to dogma in some new way. And what happens after the election? Will this internal schism go away? Or does this presage a new battle between liberal liberals and liberal moderates, and between conservative conservatives and conservative moderates?

Your guess is as good as mine, though. Any ideas?

 

Here are two graphs representing the distribution of 2012 US House and Senate congressional candidate ideological positions. Higher (more rightward) scores are more conservative, lower (more leftward) scores are more liberal. Click on the plots for higher resolution versions:

cands_house2012

cands_senate2012

A couple of things can be seen clearly from these two pictures:

  1. There are two distinct distributions of scores, representing the two political parties. They are distinct; or, in other words, the parties are ideologically polarized. Democrats are liberal, and Republicans are conservative.
  2. There is a significant amount of overlap between the party bell curves. That is, there are plenty of conservative Democrats who are more conservative than a number of liberal Republicans (and vice versa). Even in an age of polarization, the candidate pool is not completely divided, unlike Congress in recent years. This replicates a finding about the Congress of the mid 90s by Stephen Ansolabehere, Jim Snyder, and Charles Stewart from over a decade ago.
  3. On average, Senate candidates are slightly more centrist than House candidates. This makes sense given the larger, more heterogeneous states that they seek to represent, relative to the smaller and more extremist House districts.
  4. It appears the candidate pool of the parties in 2012 is roughly symmetrically polarized.

Notes:

  1. These scores are based on candidate positions expressed in survey responses, campaign statements, web sites, etc., as compiled by Project Vote Smart.
  2. They represent 722 House candidates from 419 districts and 64 Senate candidates from 33 states with elections this year. Not all candidates were scored because of a lack of data, but it’s a small number in that position.
  3. I have jointly classified all candidates into a common space, which simply means that House and Senate scores are comparable.
  4. More details about how I generated these scores can be found in a companion post that I wrote to keep this one more lean.
  5. The underlying scores are preliminary and subject to change, but I’m making them available to anyone interested in the name of transparency in another companion post here.
  6. You can find out more about my research on legislative ideology here.

Political scientists have been trying to summarize politicians’ ideological preferences for a long time. The most well accepted version of these are called ideal point estimates. These are measures of inherently unobservable preferences that are estimated from observed behavior. I see you voting in favor of a higher minimum age and regulating carbon dioxide as a pollutant, and I infer you’re probably a liberal. Or maybe you vote in favor of the Canadian oil pipeline as well as against “Obamacare” and I think you’re probably a conservative. As a sign they’ve hit the (nerdy) big time there’s now even a great XKCD comic about Keith Poole’s and Howard Rosenthal’s DW-NOMINATE scores.

The observed behavior that is most commonly employed are the Yea and Nay votes taken on roll calls in legislatures like Congress. These are very attractive to use as the raw data for ideal points for many reasons, one of which is that there is almost always an embarrassment of data. I’ve used them extensively in my research; here is a paper I recently published with Nolan McCarty on state legislative roll calls.

But they’re not perfect, for two reasons. First, a candidate at election time may present a different platform to voters than he actually uses as a guide to voting on roll calls once he achieves office. Second, by definition, they are only available after an election. This means we can’t get information on the losing candidate in state or district. This is a much more serious problem than the first.

An attractive alternative observable data is the candidate survey. In my opinion, the best candidate survey these days is administered by Project Vote Smart. It has been in the business of surveying tens of thousands of federal and state candidates for office since the mid 1990s. The questions it asks are numerous, well-phrased, and stretch across nearly all of the contentious political terrain you’d want them to. The results of their survey, which used to be called the National Political Awareness Test (NPAT) and is now the Political Courage Test (PCT), is published in a variety of formats for voters to use. The idea is that this makes it easier for voters to find out information on the policy preferences of candidates of whom they might otherwise know very little. The organization appears to be without a hint of partisan bias, as a nice bonus.

There’s another problem, one you might have guessed. Not every candidate answers the survey; in fact, fewer and fewer candidates do as time goes on. Many obviously feel that doing so could be an electoral liability now or in the future; better instead to refuse to be pinned down on many questions of policy specifics.

So Project Vote Smart figured out a solution in 2010 and now again in 2012. It would research answers to a subset of their candidate survey using good old fashioned research brawn. So nearly all of the congressional candidates in 2012 for nearly all of the congressional districts and all the states that are having elections to the House of Representatives and the Senate are represented in their 2012 Vote Easy tool. The tradeoff for this broad coverage is that only a small subset of policy stances could be researched for the many hundreds of candidates this year.

I’ve built on their work by merging their deeper but narrower NPAT with the smaller but broader Vote Easy. This gives us the best of both worlds. And the most important step is to estimate ideal points from this merged survey data. I’ve done this using a Bayesian two-parameter, one-dimensional item response model, implemented in the R statistical environment with Simon Jackman’s invaluable pscl package and visualized with Hadley Wickham’s powerful ggplot2 package.

How valid are these scores? One way to assess their external validity is to assess their convergence with measures taken from unrelated data. Luckily for me, just such an external data source exists in the form of Adam Bonica’s candidate scores for 2012. Bonica’s candidate scores correlate with my own at a level of r=0.88, which is quite high, especially as both of our measures are measured using no data in common and different estimators. The advantage of my method, though, is that it allows me to jointly classify candidates and voters, something I’ll be returning to in my blog in the coming days before the election.

For more technical details, you can consult a paper I cowrote on congressional voting with Jon Rogowski in part by using this data amalgam. You can find out more about my research on legislative ideology here.

Normally, I write something in 2012 for publication in 2013-2014 about what happened back in 2008 or 2010. Interesting, but not as much fun as it could (should) be. So, without further ado, here are the results of my exercise. Here are the plots of the two parties in 2012, and here are the underlying scores.

Big thanks to Chad Levinson, a political science PhD candidate at the University of Chicago, for helping me gather the survey data from Project Vote Smart.

Next Page »

Follow

Get every new post delivered to your Inbox.