Why are we analyzing the advertisements ?
Every election cycle, the use of digital advertising by US political campaigns becomes more commonplace. In the 2018 midterms, Congressional candidates and political parties spent a combined $623 million on digital ads alone . For the 2020 elections, forecasters anticipate that digital political ad spending will cross the $1.8 billion mark . Yet despite these year-by-year increases in digital ad spending by political campaigns, the bulk of analysis on political advertising remains focused on television ads.
For much of the internet’s history, digital political ads have proved challenging to study. Digital ads are highly targeted, meaning only specific individuals will ever see them, and highly transient, meaning they disappear once an individual has navigated away from the webpage. With no long-lasting record of digital ads, researchers have often relied on other modes of investigation, such as analyzing limited, non-random samples (e.g. Ballard, Hillygus, & Konitzer, 2016), interviews with social media platforms (e.g. Kreiss & McGregor, 2018; 2019), or ethnographic research on campaign operations (e.g. Nielsen, 2011).
Recently, in an effort to improve transparency following recent public controversies, companies like Facebook, Twitter, and Google have begun establishing archives of political ads run on their platforms. This has created new opportunities for researchers to examine how political campaigns use social media ads to connect with voters. This project focuses on political advertising on the social media giant Facebook. The Facebook Ad Library is an online archive of all political ads run on the platform since May 2018. Using the Ad Library API, we have downloaded a complete universe of campaign ads from political candidates who ran for office in the 2018 US Congress elections
Scholars have speculated that the internet and social media have shifted the focus of campaigns from traditional persuasion
of undecided voters to mobilization of existing supporters. Concerningly, this may be creating an environment in which
polarizing rhetoric and partisan combat are normalized, with limited dialogue over issues and policy (Bimber, 1998; Kaid, 2004;
Motta & Fowler, 2016; Allcott & Gentzkow, 2017; Fowler, Franz, & Ridout, 2018). We attack this speculation head on, using a
complete population of political ads from a major social media platform to help get accurate estimations of the rates and
correlates of such content.
Using this dataset, we set out to analyze the content, tone, and targets of digital communications in an effort to better
understand the kinds of messages that candidates highlight to voters online. Specifically, we aim to answer three research
questions: