Get Out the Vote!

How to Increase Voter TurnoutBy Donald P. Green Alan S. Gerber

Brookings Institution Press

Copyright © 2004Brookings Institution Press
All right reserved.
ISBN: 0-8157-3268-6

Chapter One

Why Voter Mobilization Matters

The United States has the busiest election calendar on earth. Thanksto the many layers of federal, state, and local government, Americanshave more opportunities to vote each decade than Britons, Germans, orJapanese have in their lifetime. Thousands of Americans seek electiveoffice each year, running for legislative, judicial, and administrative posts.

Given the frequency with which elections occur and the mundanequality of most of the contests, those who write about elections tend tofocus exclusively on the high-visibility contests for president, senator, orgovernor. This focus gives a distorted impression of how election battlesare typically waged. First, high-profile races often involve professionalizedcampaigns, staffed by a coterie of media consultants, pollsters,speechwriters, and event coordinators. Second, in order to reach largeand geographically dispersed populations, these campaigns often placeenormous emphasis on mass communications, such as television advertising.Third, the importance of these races calls press attention to theissues at stake and the attributes of the candidates.

The typical election, by contrast, tends to be waged on a smaller scaleand at a more personal level. Few candidates for state representative orprobate judge have access to the financial resources needed to produceand air television commercials. Even long-standing incumbents in stateand municipal posts are often unknown to a majority of their constituents.The challenge that confronts candidates in low-salience electionsis to target potential supporters and get them to the polls, while livingwithin the constraints of a tight campaign budget.

A similar challenge confronts political and nonpartisan organizationsthat seek to mobilize voters for state and local elections. Making scarcecampaign dollars go as far as possible requires those who manage thesecampaigns to think hard about the trade-offs. Is it best to assemble alocal phone bank? Hire a telemarketing firm? Field a team of canvassersto contact voters door-to-door? Send direct mail and, if so, how manypieces of direct mail?

This book offers a guide for campaigns and organizations that seek toformulate cost-effective strategies for mobilizing voters. For each form ofvoter mobilization, we pose two basic questions: (1) What steps areneeded to put it into place, and (2) How many votes will be producedfor each dollar spent? After summarizing the “how to do it” aspects ofeach get-out-the-vote (GOTV) tactic, we provide an impartial, scientificallyrigorous assessment of whether it has been shown to produce votesin a cost-effective manner. The chapters that follow cover the staples ofstate and municipal election campaigns: door-to-door canvassing, leafleting,direct mail, and phone banks. We also discuss some newer campaigntactics, such as voter mobilization through e-mail. The concluding chapterdiscusses other forms of GOTV activity and the research that is currentlyunder way to evaluate their effectiveness.

Does Voter Mobilization Matter?

The sleepy quality of many state and local elections often conceals whatis at stake politically. Take, for example, the 1998 Kansas State Board ofEducation elections that created a six-to-four conservative majority. Thiselection featured a well-organized campaign that used personal contactwith voters to mobilize hundreds of churchgoers in low-turnout Republicanprimaries. This victory at the polls culminated a year later in a dramaticchange in policy. In August 1999, the Kansas State Board of Educationvoted six to four to drop evolution from science educationstandards, letting localities decide whether to teach creationism in additionto or instead of evolution. This move attracted national attentionand renewed debates about science curricula and religious conviction.But what occurred in Kansas is a story not only about clashing ideologiesbut also about how campaigns work to get voters to the polls. Very fewKansans changed their mind about the merits of evolution and creationismover the course of the election campaign. What changed in 1998-and in subsequent elections, as countermobilization campaigns causedconservatives to lose their majority-was who showed up to vote.

Although Americans often take a cynical view of state and local elections,supposing that who fills a given office makes no difference, theKansas example is not as exceptional as it may seem. During the 1960s,the U.S. Supreme Court struck down many states’ system of legislativerepresentation as inconsistent with the principle of “one man, one vote.”Prior to the Supreme Court’s rulings, several states assigned equal representationto all counties, which meant that rural voters were heavilyover-represented in proportion to their share of the population. Oncestate legislatures were reorganized according to the “one man, one vote”principle, the share of government funds flowing to rural countiesdropped dramatically. Voting power matters. When groups such as conservativeChristians or elderly Americans vote in large numbers, policymakershave an incentive to take their concerns seriously. By the sametoken, elected officials can afford to disregard groups that vote at lowrates, such as southern blacks prior to the Voting Rights Act in 1965.Largely excluded from the electorate by racially biased voter registrationpractices, southern blacks saw their needs for schooling, transportation,and jobs go unheeded by state and local government.

The Kansas State Board of Education elections also illustrate thepower of small numbers in elections where turnout is low. The ability tomobilize a few hundred supporters can prove decisive when only a fewthousand votes are cast. Knowing what it takes to generate a few hundredvotes in a reliable way can therefore be extremely valuable. It canbe valuable not only for a specific candidate conducting the voter mobilizationcampaign but also for all of the candidates who share similarparty labels. Mobilizing 500 Republicans to support the GOP nomineein a state assembly race furnishes votes for Republican candidates upand down the ticket.

Getting Advice on Getting Out the Vote

Campaigns vary enormously in their goals: some are partisan, some nonpartisan;some focus on name recognition, some on persuasion, and someon mobilizing their base of loyal voters. Some campaigns seek to educatecitizens, some to register citizens, and some to motivate citizens. But variedas they are, campaigns have important and obvious commonalities.As election day approaches and campaigns move into GOTV mode, theiraims become quite similar and their purposes very narrow. By the weekbefore the election, they are all homing in on one simple task-to gettheir people to the polls. Each campaign struggles with the same basicquestion: How should remaining resources be allocated in order to turnout the largest number of targeted voters?

Ask around and you will receive plenty of advice on what is the bestway to mobilize voters in those final days or weeks. You may hear that itis one part mailings to three parts phone calls for an incumbent race. Youmay hear that, regardless of the office, it is two parts television andradio, if you can afford it, to two parts phone calls. You may even hearthat, for a nonpartisan GOTV campaign, it is four parts door-to-doorcanvassing, but you will never be able to get enough canvassers, so it isbest just to make phone calls. Almost all this advice is based on conjecture-conjecture drawn from experience perhaps, but conjecturenonetheless (see box 1-1).

What sets this book apart from the existing “how to win an election”canon is five years of rigorous scientific research. Every study reportedin this book used a randomized experimental design, which is a researchmethodology that produces a reliable way to gauge effects-in this case,the effects of GOTV interventions. In a nutshell, the experiments wereport divide lists of registered voters into a group that receives the interventionin question and a group that does not. After the election is over,researchers examine public records to see who voted and then tabulatethe results in order to determine whether those assigned to receive theGOTV treatment voted at higher rates than those assigned to the controlgroup. Although these field experiments still leave room for interpretation,they go a long way toward replacing speculation with evidence.

Another aspect of our work that contributes to our objectivity is thatwe are not in the business of selling campaign services. In the past, scanningfor truth about the effectiveness of various GOTV strategies waslike having to consult with salespeople about whether or not to purchasethe items they are selling. Many campaign consultants have financialinterests in direct mail companies, phone banks, or media consultancyservices. In this book, we make a concerted effort to incorporate theresults of every experimental study conducted since the mid-1990s, notjust the ones that are congenial to a particular point of view.

Two constraints of this work must be acknowledged at the outset.First, we have not yet looked at high-profile campaigns, such as U.S.Senate races or presidential races. Although we believe that the findingsdiscussed here are relevant to such large-scale campaigns insofar as theyrely on GOTV tactics such as phone banks or direct mail, we have yet toconduct experiments that speak directly to the effectiveness of massmedia, on which these large-scale campaigns rely heavily.

Second, although they are of obvious importance, GOTV strategiesare not the only factors at play in an election. When we speak of theeffectiveness of GOTV techniques, we have in mind the percentageincrease in voter turnout that can be attributed to professional phonecallers or direct mail, for instance. Using the most effective get-out-the-votestrategy will not guarantee victory. All the other factors that shape theelectoral fortunes of a candidate-persona, platform, party, and campaignmanagement-are relevant as well. A spectacularly successful GOTVcampaign might lift an overmatched candidate from 28 to 38 percent ora competitive candidate from 48 to 58 percent. Often, winning electionsis possible only when voter mobilization strategies are combined withmessages that persuade voters to vote in a particular way (see box 1-2).

GOTV Research and Larger Questionsabout Why People Do Not Vote

Political observers often turn to broad-gauge explanations for why so fewAmericans vote: alienation from public life, the lack of a proportionalrepresentation system, the failings of civic education, the geographicmobility of the population. We might call these long-term-very long-term-GOTV considerations. Many books written by academics focusexclusively on these explanations.

This book, in contrast, is concerned with GOTV considerations in theshort term. We do not discuss the ways in which political participationis shaped by fundamental features of our political, social, and economicsystem, although we agree that structural and psychological barriers tovoting are worthy of study and that large-scale reforms might well bebeneficial. In the concluding chapter, we describe research that might beuseful to those interested in learning more about how voter turnoutrelates to these broader features of society. The focus of this book is quitedifferent. Our aim is to look closely at how GOTV campaigns are structuredand to figure out how various GOTV tactics affect voter participation.This close-to-the-ground approach is designed to provide campaignswith useful information on the effectiveness of common GOTVtechniques. With six weeks until an election, even the most dedicatedcampaign team will not be able to reshape the country’s basic constitutionalframework or the political culture of American society. What a campaigncan do, however, is make informed choices about its GOTV plans,ensuring that its resources are being used efficiently to produce votes.

Evidence versus War Stories

Before delving into the research findings, we want to call attention to acluster of assumptions that often hinder informed GOTV decisionmaking.One is the belief that the experts know what works: that knowledgeis, after all, what makes them the experts. Someone with a lot of campaignexperience must know which tactics work and which do not underassorted circumstances. On the other end of the spectrum is the idea thatno one really knows what works because no one can adequately measurewhat works. There is no way to rerun an election using different GOTVmethods, no parallel universe in which to watch the very same campaignfocusing its efforts on mass mailings, then on phone banks, and then ontelevision ads. The final assumption is that if everybody is doing it, it mustbe useful: 5,000 campaigners can’t be wrong about prerecorded calls!

The following six chapters respond to these misguided assumptions.In short,

• Experts, be they consultants, seasoned campaigners, or purveyorsof GOTV technology, rarely, if ever, measure effectiveness. Hal Malchow,one of the few consultants to embrace experimentation, reports that hiscalls for rigorous evaluation repeatedly go unheeded. Notwithstandingthe large quantities of money at stake, Malchow observes that “no onereally knows how much difference mail and phone GOTV programsmake.”
  • Experts may report speculations in the guise of “findings,” butwithout a rigorous research design, those “findings” are suspect. Thosewho manage campaigns and sell campaign services have a wealth ofexperience in deploying campaign resources, formulating campaign messages,and supervising campaign staff. But lacking a background inresearch design or statistical inference, they frequently misrepresent(innocently in many cases) correlation as causation. They might claim,for instance, that a radio GOTV campaign is responsible for increasingthe Latino vote in a particular media market. In support of this assertion,they might point to the lack of change in the Latino vote in a neighboringmedia market. Because it is difficult to know whether the two mediamarkets are truly comparable, we find this form of proof-by-anecdoteunpersuasive.
    • There is an accurate way to measure the effectiveness of GOTVtechniques, namely, through experimental research. Randomly assigninga set of precincts or media markets to different campaign tactics makesmeaningful causal inferences possible.
      • Lastly, our results may surprise you. Just because everybody isdoing it does not necessarily mean that it works. It appears that largesums of money are routinely wasted on ineffective GOTV tactics.

        Continues…
        Excerpted from Get Out the Vote!by Donald P. Green Alan S. Gerber Copyright © 2004 by Brookings Institution Press . Excerpted by permission.
        All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
        Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.