Facebook Ads: What Are They Good For?

• In Event: Election Campaigns: Advertising, Propaganda, and Media

Abstract

Each political cycle, journalists and scholars turn their attention to online political ads. They question the ethics of online advertising practices, concerned that the public has been deceived and manipulated (Barocas, 2012; Bennett & Segerberg, 2013; Bonifacic, 2019; Griffin, 2016; Kim et al., 2018; Rosenberg et al., 2018; Tufekci, 2014). While journalists typically frame the scale of ads and ad spending within the political horserace (Baldwin-Philippi, 2020; Lapowskey, 2017; Stampler, 2012), academics tend to focus on how these ads may – or may not – influences vote choice, turn out, and attitudes towards candidates (Aldrich et al., 2016; Bond et al., 2012; Broockman & Green, 2014). Overall, recent studies find that campaign ads have a “near zero” effect on vote choice (c.f. Coppock, Hill & Vavreck, 2020). On the other hand, campaigns spend billions of dollars on digital advertisements, seeing returns in not only fundraising and campaign involvement, but also on the metrics various platforms measure, such as sharing or commenting. At the heart of this disagreement, we argue, is that academics aren’t measuring the “success” of digital ads on their own terms.

Conceptually, we present a framework for understanding digital political ads on their own terms. We situate this within the political advertising literature, as well as the literature on affordances of social media. Our overarching question is: how, if at all, are practitioner’s outcomes related to social science outcomes?

To answer this, we take a two-pronged approach. First, in late October 2020, we conducted a national, online survey experiment of American adults, drawn from a Qualtrics panel (N=XXX). We exposed partisans to real Facebook ads from the Biden and Trump campaigns. The ads selected had clear goals: persuasion, list-building, or fundraising. We test the differential effect of those ads on traditional measures of interest to social scientists such as vote choice and favorability, and we test their effect on novel measures we developed to capture practitioner goals like email list sign-ups and donations. Early results suggest almost no main effects, but identify a significant, substantial relationship between participants’ Facebook use and their willingness to donate, sign up for email lists, or answer surveys. Second, we plan to compare our survey experiment results to the actual goals and ad performance from the campaigns. We are in talks with both the Biden and Trump campaigns which have both agreed to provide us access to their data on the ads we used as stimuli: the campaign's goal for the ad, how it performed towards that goal, targeting parameters, and platform engagement metrics.

In the field of political communication, this study offers a unique contribution. Political advertising content and effects studies have historically been based on television advertising and the goals associated with that medium. Ultimately, understanding how practitioner and academic measures of effectiveness relate to each other is necessary to interpreting, predicting, and understanding the online political advertising ecosystem and its effects on voters and elections.

Authors

• Shannon C McGregor, University of North Carolina

• Bridget Barrett