Did It Matter? The Effects of Voter Suppression Campaigns on Election Outcomes

• In Event: We Won't Get Fooled Again? Misinformation, Deception, & Political Communication

Abstract

The US Senate Select Committee on Intelligence (Russian Active Measures and Election Interference in the 2016 Elections, 2019) reported that the Internet Research Agency (IRA), a Kremlin-linked influence campaign operative, engaged in voter suppression campaigns on social media (including Facebook) especially targeting African American voters and likely Clinton supporters (see also, DiResta et al., December 2018; Howard et al., December 2018; Kim, September 2018). A study (Kim et al., 2018a) discovered that “suspicious” groups, many of them later turning out to be Kremlin-linked Russian groups, ran divisive issue campaigns targeting battleground states including Pennsylvania and Wisconsin. A follow-up study (Kim et al., 2018b) also found voter suppression campaigns sponsored by unidentifiable groups targeted nonwhite voters, Bernie supporters, and young females. Such targeted voter suppression campaigns were concentrated in minority-majority counties in battleground states.

However, little empirical research has been conducted to systematically assess the effects of voter suppression campaigns on election outcomes in 2016. While a recent empirical study (Bailey et al., 2019) that tracked those who were exposed to the IRA on Twitter with a longitudinal survey found little persuasive effect (at best, reinforcement effects), the study is limited by a very small number of cases that identified Twitter exposure in the first place. Furthermore, the study was conducted post-election, when most of IRA’s Twitter activities heavily focused on organizing rallies and counter-protests on both sides. Scholars indeed have been divided in speculating the effects of the IRA influence campaigns on election outcomes. For instance, while Alan Abramowitz argues (August, 2019, Rasmussen Report) no single factor other than party identification explained the vote shares in the 2016 presidential election, Kathleen Hall Jamieson (2018) maintains in her book that the IRA influence campaigns built up negative public discourses on Hillary Clinton, hence, helped Donald Trump win the election.

The proposed study strives to assess the effects of the voter suppression campaigns by unidentifiable sponsors on voter turnout and candidate preferences. The study utilizes 5 million Facebook ads exposed to nearly 10,000 individuals who represented the US voting age population. We first track the individuals’ turnout records in 2016. By employing a quasi-experimental design, we isolate and estimate the effects of voter suppression campaigns on turnout at the individual level. We also utilize our four-wave panel study to track the changes in individuals’ candidate preferences over time for six weeks up until Election Day.

A preliminary analysis (an analysis of a small subset of the sample) indicates that although marginal, individuals’ exposure to voter suppression ads indeed significantly decreased their turnout, especially that of nonwhite voters. It also significantly decreased Clintonpreferences, even when compared to Democrats and nonwhites in general. A full scope of analysis of the actual turnout tracking as well as the panel data will be included.

Author

• Young Mie Kim, University of Wisconsin-Madison