Individual Submission Summary
Share...

Direct link:

Understanding Digital Advertising across Platforms in the 2022 Midterms

Fri, September 6, 8:00 to 9:30am, Marriott Philadelphia Downtown, Franklin 6

Abstract

The arrival of the digital ad libraries by major platform companies in 2018 was heralded as a huge step toward providing transparency about digital political advertising. But large-scale, systematic analyses of such advertising has been slow to develop in large part due to 1) the large cost associated with gathering, processing, and standardizing digital data across platforms, 2) discrepancies in the universe of ads covered by platform databases, and 3) challenges in uncovering and systematically identifying group sponsors in particular. In this research, we unveil the results of a new Cross-Platform Election Advertising Transparency Initiative (CREATIVE), which provides a method, scripts, and standardized results of processed election advertising data from the Meta (including Facebook and Instagram) and Google (including search and third party advertising along with YouTube) advertising libraries in a centralized repository for further analysis. We describe how we standardized the universe of federal advertising and processed the advertising content to make comparable inferences about the content, spending, and distribution of digital advertising on the major platforms. We present a novel dataset that comprehensively tabulates the creative content extracted from text and audiovisual media, spending and audience targeting information, as well as political entities, persuasive goals, tone and sentiment inferred from the creatives. We compare how digital political advertising varies across platforms based on this dataset, and we further compare these findings to television. Our research is innovative methodologically, enabling other researchers to examine the scope of digital advertising. It also reveals important insights about patterns of digital ad spending–and differences in content–across both sponsors and platforms.

Authors