Googling the ‘Big Lie’: How search engine algorithms determined exposure of the US 2020 presidential conspiracy

de León Williams, Ernesto Emiliano; Makhortykh, Mykola; Urman, Aleksandra; Ulloa, Roberto (2022). Googling the ‘Big Lie’: How search engine algorithms determined exposure of the US 2020 presidential conspiracy (Unpublished). In: Computational Communication Research in Central and Eastern Europe. Helsinki. June 27-29 2022.

Full text not available from this repository. (Request a copy)

Search engines have become a key venue for political information exposure. In this study, we focus on how they shaped access to information about the debunked conspiracy theories that the US 2020 election was 'stolen' from Donald Trump. Colloquially known as the 'Big Lie', these claims have become a key part of political discourse in the US. By algorithmically shaping information selection presented to internet users, search engines have profound impact on how citizens perceive these conspiracies, in particular consider their known tendency to present outputs of varying quality, in some cases aligning with specific partisan narratives (Kravets, & Toepfl, 2021; Urman et al., 2021). Such discrepancies in the provision of information can have dire effects on how citizens think of politics, especially if such differences occur across partisan lines.

While multiple variables affect search outputs, we argue that algorithm-driven exposure to conspiratorial information on search engines is primarily affected by three affordances: 1) search terms, 2) location, and 3) search engine used. We explore how variance in these factors increases likelihood of exposure to election conspiracy information through an agent-based auditing technique. This technique uses software (automated agents or ‘bots’) to simulate human browsing behavior (see, for instance, Pradel, 2021; Ulloa et al., 2021; Unkel & Haim, 2019), while keeping constant other confounding characteristics. This allows us to attribute variation of specific characteristics (search terms, location, and search engine) to changes in content presented.

Using a large sample of agents (n=180), we explore changes in results obtained by manipulating the ‘conspiracy-ness’ of the search terms used. Here we constructed a list of 10 terms ranging from the mild ‘voter fraud 2020’ to the more explicit ‘water leak Fulton County’. Past work has shown that location can impact results presented (Kliman-Silver et al., 2015) - because we are interested in the effect geographical partisanship can have, we systematically compared results from a Republican (Ohio) and Democratic (California) held state, as well as an English-speaking placebo - the United Kingdom. Tests were also conducted across the most popular search engines internationally (Google, Bing, Yahoo, DuckDuckGo, Yandex and Baidu). Qualitative content analysis is then used to determine the impact that these three factors can have in the content presented by search engines, with a specific focus placed on whether and how results promote or debunk conspiracies.

Kravets, D., & Toepfl, F. (2021). Gauging reference and source bias over time: how Russia’s partially state-controlled search engine Yandex mediated an anti-regime protest event. Information, Communication & Society, 1-17.

Pradel, F. (2021). Biased representation of politicians in Google and Wikipedia search? The joint effect of party identity, gender identity and elections. Political Communication, 38(4), 447-478.

Ulloa, R., Makhortykh, M., & Urman, A. (2021). Algorithm Auditing at a Large-Scale: Insights from Search Engine Audits. arXiv preprint arXiv:2106.05831.

Unkel, J., & Haim, M. (2019). Googling politics: Parties, sources, and issue ownerships on Google in the 2017 German federal election campaign. Social Science Computer Review, 0894439319881634.

Urman, A., Makhortykh, M., & Ulloa, R. (2021). The Matter of Chance: Auditing Web Search Results Related to the 2020 US Presidential Primary Elections Across Six Search Engines. Social science computer review, 08944393211006863.

Item Type:

Conference or Workshop Item (Paper)

Division/Institute:

03 Faculty of Business, Economics and Social Sciences > Social Sciences > Institute of Communication and Media Studies (ICMB)

UniBE Contributor:

de León Williams, Ernesto Emiliano, Makhortykh, Mykola, Urman, Aleksandra

Subjects:

000 Computer science, knowledge & systems
300 Social sciences, sociology & anthropology
300 Social sciences, sociology & anthropology > 320 Political science

Language:

English

Submitter:

Mykola Makhortykh

Date Deposited:

17 Aug 2022 08:24

Last Modified:

05 Dec 2022 16:22

Uncontrolled Keywords:

algorithm audit, US elections, bias, partisanship, conspiracy theory, Trump, Google

URI:

https://boris.unibe.ch/id/eprint/171973

Actions (login required)

Edit item Edit item
Provide Feedback