Calling All Neighbors: Mobilizing Turnout for a Local Housing Referendum

Abstract

Existing political science and marketing scholarship suggests that a get-out-the-vote (GOTV) message focused on a neighborhood or not-in-my-backyard (NIMBY) issue will be more powerful than a basic GOTV script. However, randomized field experiments testing different scripts have found either minimal script effects or that neighbor scripts are less effective at moving voters to the polls. Here, we take advantage of a local vote on a NIMBY issue—approval of a zoning exemption for a low-income senior housing development—to test the effectiveness of reaching out to registered voters as members of a neighborhood. Consistent with prior work, we find that a basic GOTV message is more effective in increasing turnout; however, we also find that the neighbor message is more effective among some subgroups of voters, including those living farther away from the proposed development.

KEYWORDS:

• field experiment
• GOTV
• local politics
• NIMBY
• voter mobilization

Notes

These hypotheses were developed prior to the fielding of the study based on the results and remaining questions from previous studies, but there was no registered pre-analysis plan.

Given time and labor constraints, we were unable to attempt calls to everyone assigned to the three conditions. A total of 8,130 calls were attempted. The allocation done at the time of randomization was equal for the three groups, but different script lengths generated different numbers of attempted calls for each condition.

Following Nickerson (2005), the baseline is not included in our analysis.

While the timing of each call was not randomly assigned, the call sheets were arranged in a random order, thus randomizing call timing.

A call is considered complete if we were able to complete the script with the targeted registered voter. A call is considered as contact if we were able to start the script with the targeted registered voter.

The results are identical if a difference in means test is used instead of a difference in proportions test as well as when confining the analysis to permanent absentee voters. For the latter analysis, see Appendix Table B5.

There was some missing information in our data for date of contact. We used mean-substitution to eliminate dropped cases. Ordinary least squares analysis was also conducted producing similar results with only slight differences in coefficient values but no changes in significance in variables. These results are presented in Appendix Table B4.

We also tested whether the timing of the calls mattered in terms of the effectiveness of the treatments and find that the basic and neighbor calls made 2 weeks prior to the election were effective in increasing turnout, while they had no effect for the calls made the weekend before (see Appendix Table B3).

Additional information

Notes on contributors

Lisa Pringle

Lisa Pringle is a graduate student at Claremont Graduate University whose research focus is on Latino elected officials.

Melissa R. Michelson

Melissa R. Michelson (PhD Yale University 1994) is Professor of Political Science at Menlo College. Her current research projects explore voter registration and mobilization in minority communities and persuasive communication on LGBT rights.

Jennifer L. Merolla

Jennifer L. Merolla (PhD Duke University 2003) is Professor of Political Science at the University of California, Riverside and is co-author of Democracy at Risk: How Terrorist Threats Affect the Public (2009) and Framing Immigrants: News Coverage, Public Opinion, and Policy (2016).

Deborah Brown Mccabe

Deborah Brown McCabe is a Professor of Marketing at Menlo College whose research interests center on consumer decision-making. Her research has been published in the Journal of Consumer Research, Journal of Public Policy and Marketing, and Journal of Consumer Psychology, among others.