Registering the Hard-to-Reach: Using Online Ads to Register Young Women to Vote

• In Event: Political Campaigning, Elite Discourse and Political Engagement

Abstract

Young American adults are almost entirely--and often extensively--online; these same young adults also often need to register to vote and update their registration addresses. Online advertisements offer a promising way to encourage young people to register, especially in states where they can easily do so online with a few clicks. However, demonstrating the effectiveness and calculating the cost efficiency of online advertisements to promote overall and specifically online voter registration (OVR) has proven difficult due to a number of technological and methodological challenges. This paper presents the results of a large-scale field experiment conducted across 5 states in 2022 with a national non-profit organization that introduces a cluster-based approach to measuring the effectiveness and efficiency of online ads that promote voter registration among young women aged 18-24, and offers guidance on implementing this technology to specifically reach hard-to-register populations.

While online ads promoting voter turnout can be targeted to individual registrants using cookies or custom audiences, it is not possible to target a list of the unregistered for voter registration since organizations do not know who they are or where they now live. Instead, ads are targeted generally to households thought to contain young, eligible citizens, while possibly including those already registered or in older age brackets, which lowers the cost effectiveness of the tactic. Methodologically, it is very hard to prevent spillover effects between geographic areas that are randomly assigned to be targeted with ads or held in a control group due to limitations on platforms' and data vendors' abilities to ascertain where someone actually resides.

In this paper we introduce a measurement approach that uses geographic clusters within a field experiment based on ZIP code targeting that randomly assigns specific areas with ads. Our study demonstrates that online voter registration is effective at increasing registration rates among young people by 0.82pp overall; our cluster analysis finds a 5.2pp effect in areas with high concentrations of young people and a 1.58pp effect in areas with high shares of POC residents. We find no effects in rural or urban areas with high shares of older white residents. A subsequent analysis looking at census data helps to illuminate the types of living situations in which our advertisements are most effective, with an eye to registrants who cannot easily be reached through other methods. These results suggest that online advertising can be an effective and cost-efficient way to register young women to vote.

Author

• Katherine Haenschen, Northeastern University