We're two nights into the Democratic National Convention, and the themes could not be more distinct from those championed at the RNC last week. Whereas the RNC heavily emphasized the role of personal initiative in economic success, the DNC's speakers have focused on the many barriers that keep success away from even determined, hard-working Americans.
It's a sign, the New Republic's Alec MacGillis argues, that the parties are getting more ideologically coherent — that is, they have more sharply defined, and sharply distinct, viewpoints than they once did. Democrats and Republicans are now, he writes, "ideologically coherent to the point where they make even Europe's parliamentary parties look muddled by comparison."
But is this showing up in actual congressional votes? To find out, I looked at Keith Poole and Howard Rosenthal's "DW-NOMINATE" database, which uses congressional votes to measure the ideological position of members of the House and Senate.
Steve Smith, a political scientist at Washington University in St. Louis, and Sarah Binder, who's at George Washington, computed the standard deviation — which measures variation in scores — for each chamber and each year to see how widely scores ranged within each party, and graciously passed the results on to me. If the standard deviation is bigger, the party is less unified. If it's smaller, the party is more unified. So how do the standard deviation figures change from Reconstruction to today? Here's the Senate:
Poole and Rosenthal also put out "party unity" scores which measure how frequently members vote with their parties on key issues. Interestingly, these numbers show much more similar behavior between the parties. These numbers measures the percentage of votes that members of each party voted with their party on: