« Fighting Bob La Follette | Main | using Movable Type »

November 19, 2003

The GOP and voter mobilization

Racine, WI: I've been told that the Republican Party has conducted more than 50 randomized experiments to test which methods and messages most efficiently persuade people to vote. My organization, CIRCLE, has also funded and collected such randomized field tests, although we are a nonprofit organization, so we can only test completely neutral, non-partisan messages ("Vote for someone this fall").

In a true experiment, you don't just ask people to vote, check whether they do, and count each vote as a success. That would be a flawed methodology, since many people would have voted even if you hadn't asked them. Instead, in a true experiment, you divide the population randomly into two groups, ask one group to vote, and leave the other group alone. Your success rate is the difference in turnout between the two groups. CIRCLE-sponsored experiments have found that some strategies cause many young people to vote; some are ineffective; and some promising approaches actually reduce turnout. I find it fascinating that the GOP is now using this method for their own planning purposes. It means, first of all, that a sophisticated academic methodology seems valuable to hard-nosed political operatives. And second, it means that Republicans are likely to try to mobilize people through face-to-face contact in 2004. That is a form of campaigning that increases participation (in contrast to TV advertising, which is sometimes intended to reduce the opponent's turnout). Thus it is is a very beneficial development, although it would be unfortunate if the Democrats failed to imitate the GOP.

Tuesday, Nov. 18

November 19, 2003 6:01 AM | category: none

Comments

Peter,

Have you been able to obtain copies of any of the Republican studies?

Clayton

May 26, 2005 5:25 PM | Comments (2) | posted by Clayton Nall

No--as far as I know, all the partisan studies (both Democratic and Republican) are proprietary.

May 26, 2005 7:52 PM | Comments (2) | posted by Peter Levine

Site Meter