Great article on Micro-targeting and the importance of CANVASSING (the New Yorker)
Of course, even if you find the right voters, and make the right appeals to them, theres no guarantee that theyll show up at the polls. Campaigns have always worked hard to get out the vote on Election Day, but they didnt always work smart. As Issenberg shows, that has changed, thanks largely to the efforts of the political scientists Alan Gerber and Donald Green. Randomized, controlled experiments, common in academia, were virtually unknown in politics. But, starting with the 1998 midterm elections, Gerber and Green experimented with various get-out-the-vote techniques. The first experiment compared the effectiveness of calls from a paid call center, a piece of direct mail, and a home visit. Having people knock on doors, it emerged, boosted turnout by almost nine per centfar more than anyone had predicted, and easily sufficient to swing an election. Impersonal techniquesgeneric direct mail, phone calls from paid, script-reading workersdid little to encourage voting, but chattier, more inquisitive phone calls from volunteers were effective. Later experiments showed that social pressure could also boost turnout. For instance, Gerber and Green sent people postcards telling them that their neighbors would be informed after the election of whether or not they had votedwhich made turnout soar. And positive reinforcement works, too. Just thanking people for having voted in the past significantly increases the chance that theyll vote again.
From one angle, these successful tactics seem quite mundane. But thats whats striking: political campaigns long neglected useful strategies, because they never did the research to find out what really worked, and because they were obsessed with, as Issenberg puts it, changing minds through mass media. And of course the homely labor of knocking on doors and sending postcards wasnt as valuable before the egghead work of micro-targeting could tell you which doors to knock on. In that sense, a well-run ground game is a perfect blend of the technological and the concrete, of the future and the past.
A great ground game cant make up for dismal fundamentals or a terrible candidate. Rick Perry was a fervent believer in victory-lab techniques, but that couldnt help him string coherent sentences together. Yet such techniques can make a huge difference on the margin, and in close elections thats all the difference that matters. This doesnt seem to bode well for Mitt Romney. Although he pioneered the use of micro-targeting in his Massachusetts gubernatorial campaigntargeting premium-cable subscribers, who were thought likely to favor his technocratic approachhes now running a very TV-intensive campaign. The Obama campaign has put far more into the ground game, and it has the advantage of the experience and the data it gathered last time around. Romney will likely spend more and buy more ads, but that may not matter if Obama ends up with more knocks on the door
http://www.newyorker.com/talk/financial/2012/09/24/120924ta_talk_surowiecki