The Art and Science of GOTV

Like most truisms, "it's all about turnout" is overused and undervalued. In the 29th, this is especially true. Since GOTV is so important in the 29th, I want to take an in-depth look at what's known about GOTV, beginning with "the book" on GOTV, Get Out the Vote! - How to Increase Voter Turnout.

GOTV! summarizes major recent research of the late '90s and early 2000s. It looks at partisan and non-partisan GOTV efforts, and it tries to assign a price per vote on common GOTV efforts.

The first interesting fact about GOTV studies is that there aren't many of them. Most campaigns don't want to be part of an experiment where half of the potential voters who serve as a control aren't contacted. The other interesting fact is that campaigns spend a lot of money on technologies, like robo-calls, that aren't very effective.

The main conclusion of this book is twofold: Personal approaches are more effective than impersonal ones, but personal approaches are harder to replicate on a large scale.

So, not surprisingly, the studies reviewed in the book find that the most effective approach is door-to-door canvassing. It's interesting that it really doesn't seem to matter what the canvassers say, as long as they meet a minimum level of competence. It's all in the personal contact, which puts a face on the campaign of one of the candidates.

The downside of knocking on doors is that eligible voters are hard to find, and that it's hard to find the teams of volunteers, or hard to manage paid canvassers. Canvassing is also hard to scale - you can contact a few thousand voters this way, but getting to hundreds of thousands of voters is much more difficult.

Leaving leaflets (such as door hangers), the poor cousin of canvassing, is less effective than direct personal contact, but partisan leaflets have been shown to increase turnout by a small but statistically significant amount.

Methods that don't involve feet on the street still have some effect, but that effect is again proportional to the personal contact. The effectiveness of phone banks seems to be directly related to how "conversational" the contact is. If the script is carefully constructed to solicit interaction with the person being called, and if the volunteer or professional making the call takes time to slowly and carefully go through the script, phone banks can have some effect.

Robo-calling, which is impersonal and non-conversational, was shown to have little or no effect. Repeated calling doesn't seem to work well either, and calls are more effective if they're closer to the election.

Direct mail is even less effective than phone banks, and it is the least effective with marginal voters. The partisan base can be motivated by direct mail, but even so, it is expensive.

Perhaps the most interesting part of the book was the price-per-vote summary. Canvassing costs approximately $19/vote, using contract labor. Obviously, free labor is cheaper, making volunteer canvassing the cheapest known way for a campaign to increase its vote tally. Leafleting may be a bit cheaper ($14/vote) to motivate a partisan base, but further study is needed to confirm that as fact.

Volunteer telephone banks cost about $35/vote, assuming that the volunteers are well-supervised and use a good script. Professional calls with long scripts and good supervision cost about $45/vote, but low-quality professional calls can cost upwards of $200/vote. Robo calls don't work, so cost per vote can't be calculated.

Direct mail costs $59/vote when addressed to base voters, but $200/vote when directed to sporatic voters. Email doesn't work (no surprise there).

What I take home from this study is that it isn't money alone that gets out the vote. Campaigns can spend a lot of money with little or no effect. To get out the vote effectively, campaigns must plan well in advance and have enough staff to coordinate expensive GOTV efforts.

With the "scientific" base of GOTV covered, the next GOTV post will examine how the campaigns in the 29th will turn out voters next month.