This week I heard political scientists Melissa Michelson (CSU-East Bay), Lisa Garcia Bedolla (UC-Irvine) and Donald Green (Yale) present preliminary findings from their research on which methods used by community organizations among James Irvine Foundation grantees in the California Votes Initiative really worked to increase voter turnout. The research involved doing randomized trials of turnout techniques on low propensity voters using the community groups as campaign labor.
Folks interested in the mechanics of campaigns need to think about these findings through a reality filter -- maximizing turnout is seldom a campaign goal. Campaigns are about maximizing the right turnout, the turnout of voters on our side. (Remember what Karl Rove thought U.S. Attorneys ought to be doing.) Nonetheless, these social science observations can help us think about what we do.
The academics assured the audience that they had not found anything startlingly different from the findings in a previous book by Green and Alan Gerber which I have discussed here. But they brought some nuances and additional certainty about previous findings.
What doesn't work
Since resources are always scarce, this can be the most valuable insight. This project reinforced previous research that direct mail has almost no discernable positive effect on turnout. Considering the amount of money that campaigns spend on it, that's important. Mail may persuade a likely voter to choose your candidate as opposed to someone else, but it won't engage unlikely voters. If winning requires engaging those voters, you'll have to use other forms of contact.
Two other contact methods that do not create measurable turnout bumps are robocalls (even by trusted local spokespeople like pastors) and mass emails. Campaigns might believe these are worthwhile to reinforce existing support, but they aren't going raise vote tallies.
What does work
The gold standard remains the "at the door" personal canvass. Researchers provided a lot of useful information about how to raise the quality of these contacts. My takeaway from their findings follows (the academics are not responsible for every nuance I'm giving this.)
- At the very least, the canvasser needs to be trained to carry on an actual conversation with the voter. This needn't be deep; the researchers were unclear whether messages that were partisan were more effective than messages emphasizing the duty of citizenship. But the canvasser needs to be able to solicit a back and forth interaction with the target voter.
- Canvassers associated with a known and respected community group were relatively more effective than unconnected volunteers. Campaigns need those community groups in order to be heard by unlikely voters.
- Canvassers of the same ethnicity and/or language group are more effective with voters with the same demographics. This is intuitively obvious, but it is nice to have research begin to confirm it.
- Similarly, but not quite so obviously, the closer the canvasser lives to the voter, the higher the likelihood of turning the voter out. Researchers were surprised by how strong the effect was. A canvasser from the same precinct was measurably more successful than one from a few streets further away. For the researchers, the finding has prompted a desire to repeat and replicate the experiment because the finding seems so strong. For the campaign manager, this argues for using "precinct captains" or even "block captains" as much as possible to do the canvassing.
This one is sad news for campaigns: unlikely voters are more likely to be engaged by contacts in the last two weeks before election day than they are by earlier contacts. Again, we've always known that intuitively, but we're getting the research to prove it. This argues for firming up your base (and attracting and training your volunteers) in the period two months until two weeks out, then intensively canvassing the unlikely voters.
This also makes the issues raised by vote by mail and absentee campaigns more complex. Donald Green specifically reported that his research says that absentee voting merely moves likely voters from the election day category to the absentee category. It does not seem to attract new voters. Voting by mail does however lengthen that short two week window when contacts are most effective.
I've long argued that pushing absentee voting is the wrong way to go for low budget campaigns in low income, low propensity voter communities. Absentee voting is bureaucratic and expensive to track -- not what community campaigns do best. To draw people into the voting universe, the process should reinforce a sense of community: going to a community polling place on a particular day is more a collective activity than dropping an envelope in a mailbox individually when you get around to it.
All the research reported by these folks supports the premise that voting increases with the strength of social capital; that civic participation expresses and reinforces social cohesion. When we think of ourselves as isolated families and individuals, alone, protecting ourselves in a cruel world, there is little motivation for voting. Why bother? When we experience ourselves as members of a lively community that collectively is creating its own future, we vote and we get together in organizations small and large to influence our quality of life. Campaigns need to live inside of and enlarge the flow of ongoing community life. When they do, they win.
For many, many interesting reports of election research, take a look at GOTV.