(MIDDLETOWN, CT) June 24, 2025 – New research published from the Wesleyan Media Project (WMP) in collaboration with Privacy Tech Lab at Wesleyan University describes a new cross-platform digital advertising dataset. The data cover federal election activity from the 2022 midterm general election period for Meta and Google. Published in Scientific Data and funded in part by the National Science Foundation, “Comparable 2022 General Election Advertising Datasets from Meta and Google” (open access link) seeks to provide a starting point for scholars seeking to understand, describe, and assess the impact of the 2022 digital advertising landscape on the two largest digital platforms. The paper describes the extensive computational pipeline infrastructure – named Cross-Election Advertising Transparency Initiative (aka CREATIVE, which is a play on the word for advertising content as well) – that the authors built to gather, systematize, and label digital advertising related to federal elections in a consistent way across platforms.
“Creating these datasets took a considerable team effort to overcome the challenges of effectively using public ad libraries provided by platform companies,” said Meiqing Zhang, Wesleyan Media Project postdoctoral research fellow and lead author on the paper. “We hope that the resulting comparable datasets across platforms and our data pipelines facilitate quantitative studies of election campaigns, contribute to ad transparency, and complement election research based on controlled experiments or user-generated content on digital platforms.”
The collaborative project identified political figures from digital ads that are not keyword searchable, such as images and videos. It then extracted creative content from audiovisual information using state-of-art models accessed through proprietary APIs, making the results accessible to the public. The Project also combined expert knowledge and computational methods to design procedures for detecting political variables of interest, including ad tone, goals, issues, electoral races, and party leaning.
“The online digital ad libraries are very valuable; however, because they lack consistency across platforms, many researchers were spending a lot of time doing time consuming and expensive work in labeling. Our goal was to provide infrastructure and data to standardize and remove those obstacles so that all researchers can work from a more uniform set of federal election advertising with consistent and comparable labeling,” said Erika Franklin Fowler, co-director of the Wesleyan Media Project.
About the Project
The Cross-Platform Election Advertising Transparency Initiative (CREATIVE) project – a joint infrastructure project between the Wesleyan Media Project and Privacy Tech Lab – was funded in part by the National Science Foundation grants 2235006, 2235007, and 2235008. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.
