Voting In The New York State Senate

Introduction

New York's state Senate is unusual. The chamber is currently controlled by a coalition of Republicans and a few Democrats who caucus with them (the "Independent Democratic Conference", formerly including now-indicted Malcolm Smith, and Simcha Felder).

Typically, legislators caucusing across party lines would face near-certain defeat in the next primary, but it's not at all clear this will happen in New York.

Beyond that, New York's state government is infamously known as being controlled by "three men in a room": the Governor (Cuomo), the Senate Majority Leader (Republican Dean Skelos), and the Assembly Speaker (Democrat Sheldon Silver).

How does all this show up in legislative voting? To find out, I used the OpenStates API (and my own OpenStatesParser or variations thereof) to retrieve all contested roll calls this session.

Note: This currently includes votes from the 2013-2014 session through 03/31/2014, and doesn't include the handful of most recent votes.

A surprisingly agreeable chamber

Let's start with a chart of the Senators sorted by the percentage of contested votes where they agreed with Republican leader Dean Skelos.

Democrats who caucus with Democrats are in blue. Democrats who caucused with the Republicans (or who used to before they were indicted) are in green, and Republicans are in red. Hover over a Senator to enlarge the picture and show their name, district, and exact percentage.

Note that the Republican-caucusing Democrats (Valesky, Smith, Klein, Felder, Savino, Carlucci) all voted with Skelos at least 97.36% of the time. (The exception is Tony Avella, who only joined the IDC a few weeks ago, and who doesn't appear to have been voting similarly to them for most of the session.) John Sampson voted with Skelos the most of any other Democrat (97.58%).

For that matter, even the median Senator by this metric, Democrat Joseph Addabbo, voted with Skelos more than 96% of the time, and quite a few Democrats voted with Skelos at least 90% of the time, including Ruben Diaz (who has considered joining the IDC).

There might not be another state legislative chamber where this would happen. Remember, this is only counting votes that were contested by at least one Senator!

Ideological scores

For a somewhat more nuanced picture, let's look at the Senate's ideal points.

It's pretty consistent with the Skelos chart, especially in the first dimension (probably "liberal/conservative"), as ideal points are closely correlated with the principal component of the correlation matrix.

For example, Velmanette Montgomery has the "most liberal" ideal point and also agreed with Skelos the least. Republicans with "conservative" ideal points--Greg Ball, John Bonacic, Kathleen Marchione--agreed with Skelos rarely for Republicans.

As you'd expect, the parties are pretty "well-sorted", with Democrats basically to the left of Republican-caucusing Democrats, who are to the left of Republicans, with only Simcha Felder (technically not an IDC member) fully in the Republican cluster, and only Tony Avella (who, again, only joined a few weeks ago) fully in the Democratic cluster.

You can also use the drop-down menu to color the Senators by any of the votes, which are currently listed by OpenStates ID.

Except for the "Party" option, "No" votes are in yellow, "Yes" votes are in grey, and everyone else is hidden from the display.

Since there are so many votes, try using your arrow key to scroll through and see the different breakdowns.

Methodology

I used my OpenStatesParser on the OpenStates API to get their voting records for the current session, calculated their 2-dimensional ideal points using Simon Jackman's "pscl" package in R, and grabbed their pictures from OpenStates too.

I had to remove a few misclassified votes, though, and re-run everything.

Every contested vote in the OpenStates API is included--some apparent committee votes along with the full Senate.

Each circle is a CSS element. I think it's a fun visual, but there are problems (like people ending up on top of each other, even with the hovering).

I suppose I can justify it by saying that people on top of each other probably have very similar voting records anyway, but honestly, I just like the idea. I looked up a bunch of different how-tos, perhaps chiefly, but not limited to, this one (C. Bavota, on making the circular images) and this one (Scott Murray, on making the axes look nicer). Also, this NYT chart by Larry Buchanan, Tim Wallace, and Derek Watkins (to figure out how to make labels).