More on seating arrangements and voting

Just to expand on my previous post, a few years ago I wrote an article investigating the influence of legislative seating arrangements on roll call votes. We know there are all sorts of influences on the way legislators vote, including party leaders, the executive branch, constituents, etc. Did legislators also follow the votes of the people sitting near them?

Seating chart for the 1949
California Assembly

I investigated this using the California Assembly, where legislators sit in pairs. Usually, member-pairs are of the same party, but that wasn't always the case. The diagram at left shows the seating arrangements for the 1949 Assembly, with Republicans in gray and Democrats in white.

The paper finds that deskmates tended to vote together, even controlling for party, constituency preferences, and many other influences. Just sitting together made any given pair of legislators anywhere from two to six percent more likely to vote the same way.

All this is to say that Sen. Mark Udall isn't nuts when he claims that having senators sit together might change the way they behave. And he's not the first -- California Assembly Speaker Willie Brown used seating assignments to enforce party-line voting, to pair freshmen up with veterans for socialization purposes, and even to separate and punish those who conspired against him. I can't believe that sitting together for one 90-minute speech, as Udall is proposing, will make much of a difference, but the idea that neighbors can influence each other has some support.