A Voter Mobilization Field Experiment with a Human Touch

• In Event: Effective Campaigns

Abstract

Numerous field experiments have been conducted to determine the effects of contacts on voter turnout. These experiments have utilized a variety of contact methods including phone calls, mailings, texts, emails, face-to-face visits and have demonstrated mostly small positive effects on turnout. What the studies mostly share in common is that the contacts (1) are relatively impersonal, administered as they are on a mass scale, and (2) deliver messages that encourage turnout generally rather than support for a particular candidate. To complement the deep array of findings from these studies, we study a different kind of field experiment that (1) uses personalized handwritten messages from one voter to another (2) and specifically names a candidate to support. The question is whether appeals that are more tailored to the recipients’ situations will have larger effects on turnout than have been estimated in previous studies.

The data we analyze result from a postcard-writing effort organized by a group of volunteers in a midwestern swing state. In the weeks leading up to the November 2018 midterm elections, a progressive group of politically active citizens developed a plan to hand-write personalized postcards to left-leaning registered voters. The recipients resided in two competitive state legislative districts in which the Democratic candidate was female. In addition, all of the recipients were also female and organizers hoped that the common gender bond would help get the attention of the voter and motivate them to participate. Each of the approximately 3,500 postcards was handwritten, stamped, and signed with the first name of the volunteer. This unusual form of contact should help to distinguish the cards from other election-related mailings that the subjects might be exposed to during the campaign season. To determine the effectiveness of the mailings on turnout, we identify a comparable control group selected in the same way as the postcard recipients and then match the individuals in both groups to the state’s official voter file. Intent-to-treat effects are estimated following standard protocols in the field experiments literature so that the effect sizes may be assessed relative to prior studies. We will also examine whether effects are conditional on respondent characteristics, especially age and prior voting history.

There are theoretical reasons to believe the experiment will have either larger or smaller effects in previous studies. On the one hand, meta-analysis suggests that field experiments featuring advocacy for a particular choice, in contrast with nonpartisan efforts, have no discernable effect on turnout (Green, McGrath, and Aronow 2013). Because the mailings in our study are advocating for specific candidates, we might see no meaningful effect on turnout. In addition, the races being targeted were not the most prominent ones on the ballot in 2018. It is possible that voter turnout was responsive almost entirely to the statewide offices at the top of the ticket and that an intervention highlighting more local contests will have little to add. On the other hand, the novelty of the treatment might attract more attention than other mobilization contacts. The message has a personal touch that is easily visible on the face of the postcard displaying the hand-written message and name of the volunteer who wrote it. Such a personal appeal could carry more weight than impersonal mobilization messages. The results of the study will generate implications for both campaign practitioners and scholars interested in the relative effectiveness of various messages and contact methods.

Authors

• Barry C. Burden, University of Wisconsin

• Levi Bankston, University of Wisconsin - Madison