I’ve just “linked a paper”:http://www.henryfarrell.net/polsci/2007/07/middleton_and_green_on_whether.html at my political science papers weblog that seems potentially pretty interesting to CT readers and others. Joel Middleton and Donald Green have written a “piece”:http://www.yale.edu/csap/seminars/middleton.pdf that tries to gauge the effects of MoveOn’s canvassers on voting turn-out. They estimate this by looking at situations where (a) the boundary of two voting precincts bisects a street so that people on one side of the street are in one precinct, and those on the other side are in the other precinct, and (b) where one of these districts had MoveOn organizers canvassing prospective voters, and the other didn’t. They then look at the voters on each side of the relevant street – this should eliminate potential bias (there’s pretty good reason to think that voters on different sides of the same street aren’t going to be measurably different from each other). When they do this, they find that people contacted by MoveOn canvasers seems to be 7% more likely to vote than the control population, which is a pretty significant difference. As Gerber and Green say, this is a nice result for the election turnout literature. But it has wider implications, as they note in an aside. There’s a lot of talk from Robert Putnam and others about how people are losing ‘social capital’ as the Internet becomes ever more important. Regardless of the underlying merits of the social capital thesis (personally, I’m pretty skeptical of it, at least in Putnam’s formulation), this provides smoking gun evidence that the Internet can enable greater participation than would otherwise have been possible, through allowing decentralized movements such as MoveOn to get the vote out more effectively.