Abstract
Internet advertisements have come under tremendous scrutiny recently for their potential to impact electoral outcomes. However, academic research has yet to determine if they have an effect on turnout. This article presents the results of a preregistered field experiment conducted in Dallas, Texas, in partnership with The Dallas Morning News in which individually targeted banner ads were able to generate a statistically significant increase in turnout among Millennial voters in a municipal election. The results show that a combination of information and voting reminder ads was effective, but only for voters in competitive districts. Estimated treatment effects were on par with a telephone mobilization campaign using live callers. These findings contribute to theoretical knowledge about the role of political knowledge and electoral competitiveness in voter mobilization, and offer a new method for testing online advertisements used by political campaigns.
Disclosure statement
No potential conflict of interest was reported by the authors.
Supplemental material
Supplemental data for this article can be accessed on the publisher’s website at https://doi.org/10.1080/10584609.2018.1548530.
Notes
1. The Federal Trade Commission describes Internet cookies as “information saved by your web browser. When you visit a website, the site may place a cookie on your web browser so it can recognize your device in the future” (Federal Trade Commission, Citation2018).
2. This study was reviewed by the University of Texas at Austin institutional review board (IRB) and approved as exempt, and preregistered with the Center for Open Science in April 2017 before voting records were released. All hypotheses tested and statistical models presented in the main article were included in the study pre-registration. The pre-registration is available at https://osf.io/35k4m/
3. Targeting is based on the Knight Foundation (Citation2015) report featuring Millennial voters who participated in presidential cycles but failed to vote locally. The age range of 23–35 was chosen to exclude Millennials ages 20–22 who may be registered to vote at home but attend college elsewhere.
4. The three ads produced turnout guesstimates ranging from 52.57% to 57.15%; the placebo guesstimate was 45.22%.
5. An article in the Dallas Observer broke down each race in terms of competitiveness (Young, Citation2017). After acknowledging the uncontested races, Young wrote, “Some of these contested elections aren’t really contests,” stating that the incumbents in Dallas single-member districts 1, 2, 3, and 9 had only token challengers who had done little to campaign. Young categorized the races for Dallas districts 4, 5, 6, 7, 8, 11, and 14 as meaningful contests. In the races that Young categorized as contested, no candidate received more than 70% of the vote (and in all but one, the winning candidate received less than 60% of the vote).
6. In Texas, city boundaries can cross county and school district boundaries. While most of the City of Dallas overlaps with Dallas County and Dallas ISD, a small portion overlaps with Collin County and Plano ISD. Of our 74,102 subjects, 2,891 were in Collin County and 2,866 were in the Plano ISD.
7. Only 1,208 voters out of the 73,909 in this study had prior municipal voting history, and were too small a group for an adequately powered subgroup analysis. Adding a control for prior municipal voting history to the primary interaction model reported in the article did not change results.
8. Calculated by dividing the effect of assignment to treatment, 0.52%, by the rate of treatment, 57.34%.
Additional information
Funding
Notes on contributors
Katherine Haenschen
Dr. Haenschen is an Assistant Professor in Virginia Tech's Department of Communication, VA, USA. Dr. Jennings is a Postdoctoral Fellow at the Annette Strauss Institute for Civic Life at the University of Texas at Austin, TX, USA.