In “Auditing Google’s Search Headlines as a Potential Gateway to Misleading Content: Evidence from the 2020 US Election,” a new paper published in the peer-reviewed Journal of Online Trust & Safety, a team of University of Washington Center for an Informed Public researchers present new data related to an effort to scrape more than 800,000 headlines from Google’s Search Engine Results Pages (SERP) from the 2020 U.S. elections to better understand how content with the potential to delegitimize or undermine trust in the election may have reached voters.
Himanshu Zade and Morgan Wack, doctoral candidates in the Department of Human Centered Design & Engineering and Department of Political Science respectively, and their five UW co-authors present results from qualitative coding of 5,600 headlines focused on the prevalence of delegitimizing information. Results reveal that videos — as compared to stories, search results, and advertisements — “are the most problematic in terms of exposing users to delegitimizing headlines.” The co-authors also “illustrate how headline content varies when searching from a swing state, adopting a conspiratorial search keyword, or reading from media domains with higher political bias.” The paper also offers policy recommendations on data transparency that allow researchers to continue to monitor search engines during elections.
The co-authors on this paper include Yuanri Zhang, a former CIP research assistant and UW Information School data science and informatics student (now studying computer science at the University of Illinois Urbana-Champaign); Kate Starbird, CIP director and HCDE associate professor; Ryan Calo, a CIP co-founder and UW School of Law professor; Jason Young, a UW Information School senior research scientist and CIP research fellow; and Jevin West, a CIP co-founder and iSchool associate professor.
As part of this work scraping Google’s search engine results pages, the team looked at 20 election-related keywords — 10 general (e.g., “ballots”) and 10 conspiratorial (e.g., “voter fraud”) — when searched from 20 cities across 16 states.
