Washblog

According to Green and Gerber, these consultants and vendors promote the practices that are the most lucrative and convenient for them. Consequently, current campaign strategy is, to a large degree, molded by the profit motives of this industry. The authors contend that politics as an art - and political involvement as a community value - have declined as a result. Green and Gerber intend this work as not only a source of statistical data on what works best for getting out the vote (GOTV), but also as a persuasive text on the need for evidence-based practice in campaigning.

The source of the data in this book is meta-analysis, the statistical technique used in Evidence Based Medicine. Below, are some examples of conclusions reached by Green and Gerber through this technique, which combined the data of multiple controlled political experiments conducted from 1998 through 2003. These results reflect increased voter turnout only; they do not indicate whether the efforts were successful in turning out votes for a particular candidate or cause.

1. Door-to-door conversations: Voter turnout increased, on the average, by one vote for every 14 contacts (actually talking to someone) for an average cost of $19 per vote (assuming that workers are paid $16 per hour).
2. Literature drops: Voter turnout increased, on the average, by one vote for every 66 contacts for an average cost of $14 per vote (assuming that the leaflets each cost ten cents and that the volunteers are paid $10 per hour). This is for partisan literature; non-partisan is much less effective.
3. Direct mail: Voter turnout increased by one vote per every 177 pieces mailed, at an average cost of $59 per vote. This is for partisan mail; non-partisan was slightly less effective.
4. Volunteer phone calls: Voter turnout increased by 1 vote per 35 contacts, at an average of $35 per vote.
5. Commercial phone bank calls: Voter turnout increased by one vote per 400 calls at an average cost of $200 per vote.
6. Robo calls and email outreach: These methods did not appear to increase voter turnout.
7. The content of GOTV communications with potential voters seems to have little effect. Results are similar with experienced or new outreach workers. It is the way that the content is delivered that matters. "Canned speeches" delivered by commercial phone banks are expensive and have low effectiveness; relaxed and friendly conversations work best.

This is surprisingly specific information and it's natural to question its applicability to local campaigns. But a persistent pattern emerges. In study after study it was the in-person conversation that, by far, yielded the greatest result. The costs per vote for this were much lower than the next most effective method, volunteer phone banks, and comparable to the cost of literature drops (which had a much lower effectiveness). Evidently, costs rise as effectiveness declines.

The Precinct System: Leafleting vs door-to-door
The major parties use the precinct system, employing volunteer elected officers within small neighborhood boundaries to establish relationships with local voters and help get out the vote. Many neighborhoods now have no precinct officers and in many of these neighborhoods, the PCOs rely on literature drops rather than personal conversations.

Clearly, a major lesson from Green and Gerber is that the parties would do well to fill all precinct officer slots and encourage these volunteers to establish a community presence. It is also worthwhile taking a closer look at what these findings tell us about the relative effectiveness of leafleting and door-to-door conversations.

Using the data reported in this book, we can imagine a precinct with 400 houses, where approximately three-quarters of them are deemed to be likely Democrats or Independents. A PCO who drops off literature to these 300 houses is likely to help get out an additional 5 votes (1 in 66 leaflets dropped).

A PCO in this precinct who visits each of these 300 houses would be likely to find someone home in approximately 150 of them - and would drop off literature at the other half. If the Green and Gerber analysis holds, with approximately one in 14 of these contacts yielding an additional voter, this second PCO would help get out an additional 12 or 13 voters. (Ten to 11 for the personal contacts and another 2 or 3 votes for the additional 150 houses at which literature was merely dropped.)

This may seem a slim return for many hours of work. However, the raw numbers do not reflect the cumulative results across all precincts in a legislative district. Elections can hinge on several hundred votes, or less - and leafleting or doorbelling in every precinct would result in increased turnout that could decide these close elections.

One of the gems of information in this book is that voter mobilization campaigns have enduring effects. The authors find that, out of 100 new voters mobilized in a given season, an average of 33 additional voters participate in subsequent elections. This effect would hold for both visits and literature drops. Visits have additional value: personal conversations yield information about neighbors in a precinct so that they can be more effectively visited - or not visited - in future outreach. And they provide the opportunity for voters to share their concerns or ask questions - to actually be involved in the political process.

Winning Campaigns and Losing the Community?
There's another factor that Green and Gerber don't discuss, but that is implied in their findings. A focus on winning campaigns - without investing in personal community connections - may, paradoxically, result in the loss of community support - even if those campaigns are successful in the short run.

Political campaigns must use limited outreach funds to communicate with those who are most likely to respond to that outreach. People who have reliably voted in the past - and those who have identified themselves as supporting the party of the candidate - are typically the most responsive to this outreach. Targeting them year after year is likely to reinforce and even amplify existing voting trends.

If older, white, middle-class people have the highest percentage of voting turnout in a community - a typical pattern - and if outreach always targets these voters and seldom targets others in the community, this demographic is likely to form an ever-greater percentage of the voting public. Without an on-the-ground system of local voter contact that operates between elections, that does not rely on lavish funding, and that does not discriminate according to past voting practices, the voter base can become increasingly narrow and the party may lose connection with a wide spectrum of the community. Our current political situation suggests that we may be in the midst of this process.

The Democratic Party would be wise to seriously increase its investment in its PCO outreach system. Grassroots organizers in every neighborhood who are encouraged to visit or call every house - or even leaflet at every house - would not only result in dramatically increased Democratic turnout - but could also prevent the kind of amplification of existing political trends that has resulted in a voter base skewed toward older, white, middle class participants. Getting back to basics, having neighbors talk with neighbors, appears a necessary practice for maintaining - and perhaps re-establishing - community relevance.