GOTV - Part 5 - Need for a Longer Time Horizon

by: Grebner

Fri Nov 26, 2010 at 05:49:59 AM EST

In order to encourage someone to vote, we stick a leaflet in her door, if she lives in a particular class of precincts. If her name appears on a particular list, maybe she gets a call from a volunteer phonebank.  If she says something in response, unless it fits into the handful of specific categories the volunteer has been told to record, the voter's comments are simply ignored. 

Without thinking, we have constructed a GOTV system that deals with individual voters almost solely in brief stand-alone contacts that are each nearly anonymous.  Where did we lose sight of the fact that voting, or not voting, is a decision made by each voter in the complicated context of their entire life?  Why do we believe the most cost-effective interventions can be broken into tiny, atomic, fragments?

[Note - My GOTV essays, plus many more, are collected under Technical Politics]  

Maybe the question how we got here is rhetorical, but we need to realize we've made a choice, and the approach we have unthinkingly taken limits our impact to a very few percent increase in turnout.

Instead of thinking of GOTV as a program that runs the last four days before November elections, we ought to look at it as continuous, running from year to year.  There would still be periods of greater or lesser activity, but we need to understand that the most efficient way to turn out voters in 2012 requires work spread over a much longer period.

One aspect of this expansion is the need to move much of our volunteer labor AWAY from a handful of days in November, spreading it over the months of September and October.  As I'll argue in a subsequent piece (only eight more to write!) we need to conduct voter registration in an aggressively partisan manner, both registering new voters and updating previously registered ones.  That work needs to be done close to the election, but before the end of registration, which is roughly Columbus Day.  After that, we need to shift our resources and attention to absentee voting, concentrating on seniors and college-age kids.  Finally, at about the time the absentee program is wrapping up, we should shift to Election Day activities.

Another aspect of stretching the time is that the individual voter should have much more of a sense that WE KNOW WHO THEY ARE.  Each contact should build on previous contacts, mentioning information we obtained previously and adding to it. To the extent possible, we should try to dedicate the same volunteer to make follow-up contacts, recognizing doing so will increase the cost and reduce the apparent efficiency of our program.  In reality, if we measure "efficiency" by the number of additional votes we turn out, rather than the number of contacts made, we're probably better off, not worse.  Both the impact on the voter, and the effect on the volunteer's morale, will be far more positive when each of them perceives a continuing relationship, rather than isolated messages.

To some extent, we already maintain long-term databases of likely Democratic voters, which we dust off and feed into the GOTV process as the two-year cycle comes around.  We may know that a voter is a Democrat, but we don't know that her mother is in long-term care and needs help voting.  We don't know that the son who was 17 two years ago is now 19, and needs to register.  We have no idea that the 24-year-old daughter who was living with her is now 26 and living on her own, and not likely to vote unless Mom helps us track her down. 

Finally, we need to think of GOTV as depending on long-term gathering and preservation of information about INSTITUTIONS and ORGANIZATIONS.  Under our current approach, we re-invent the wheel before every election - and the wheel we invent isn't very round.  We have virtually no stored information that tells us who owns a particular trailer park, or management policies regarding politics at a senior complex.  If we happen to stumble onto such information, we don't preserve it so it can help two years from now.  We need to pull together a list of contact information for every "location" larger than (say) 20 residential units, including management names, phone numbers, policies, attitudes, number of units, successful tactics, and the volunteers who are familiar with the site.  Statewide, we may be talking about 5000 sites, which sounds like a huge number.  But if we think about specific regions, the numbers are less daunting - perhaps 100 in Kalamazoo County for example.

Even Green and Gerber have tended to reinforce this focus on short-term tactics and results, because of the pressure on researchers to create manageable projects that can be proposed, funded, executed, and analyzed in time for publication deadlines.  Working with them, I find myself struggling to shift their interest toward studying richer, longer-term contacts.  Almost all their work has focused on the impact on voting of a single, specific contact like a phone call or a mailed flyer.  As long as we think that way, the most we can hope for is to increase turnout by a percent or two - when our goal ought to be ten times that.