Why not just give it a shot? How the Russian COVID-19 vaccines are framed by web search engines

Kuznetsova, Elizaveta; Makhortykh, Mykola; Urman, Aleksandra; Ulloa, Roberto (2022). Why not just give it a shot? How the Russian COVID-19 vaccines are framed by web search engines (Unpublished). In: Computational Communication Research in Central and Eastern Europe. Helsinki. June 27-29 2022.

Full text not available from this repository.

Web search engines have become trusted sources of news that shape representations of social reality worldwide. However, specific issues are represented differently across various engines leading to information inequalities between their users, in particular, when individual engines prioritize information sources of various types and quality. Therefore, understanding how search engines frame societally relevant phenomena, in particular when these phenomena are epistemically contested in the context of the geopolitics, is essential for ensuring a sustainable information ecosystem.

In this paper, we examine how various search engines in different locations select news about the Russian COVID-19 vaccines. Coronavirus pandemic is characterized by competing media narratives and increased misinformation, which is heavily influenced by ideological factors. The issue is particularly relevant due to strong effects of vaccine coverage on the individual choices about vaccination and the use of vaccines as a geopolitical tool. Amid these concerns, we set out to explore the role of search engines in prioritizing certain representations of the Russian vaccines and discuss how it may influence their global perception.

Firstly, in order to understand how the topic was represented we conceptually focus on frame setting. We understand it as a mechanism of shaping an interpretation of given issues that may affect public opinion. To do it, we focus on issue-specific frames that are unique to certain contexts and, thus, have more potential for investigating how framing can generate new meanings or amplify biases compared to more generic frames. We pose the following research question:

RQ1: How were the Russian vaccines framed in the journalistic content returned by search engines?

Secondly, we build upon previous research on algorithm-driven framing on search engines to conduct a comparative analysis of news selection in relation to the Russian vaccines. While the majority of existing studies focus on a single search engine, a few comparative inquiries (e.g., Kravets and Toeplf, 2021; Makhortykh et al., 2020; Urman et al., 2021) identify major differences between individual engines in terms of outputs and biases they might promote. Furthermore, the majority of above-mentioned studies focus on political elections, where search engines assumingly serve a “mainstreaming” function, whereas in the case of health-related matters, there is evidence of more diverse framing. Hence, we ask the following research question:

RQ2: How were frames on the Russian vaccines distributed across different search engines?

To answer these questions we use a mixed-method research design that combines frame analysis and algorithmic auditing. We particularly focus on the ‘News’ search on search engines which bears closest resemblance to how news stories are packaged in journalistic outlets. To collect data for the study, we conduct an agent-based auditing that relies on software simulating human browsing behavior (e.g., page scrolling or entering queries) to produce inputs for the algorithmic system and then record its outputs. Using custom software designed for the project, we deployed 360 automated agents via Amazon Elastic Compute Cloud in three regions - London, Ohio, and Northern California - to collect data on five search engines: Google, Bing, Yahoo!, DuckDuckGo (DDG), and Baidu.

The agents were distributed equally between the regions (120 agents per each) and search engines within the region (20 agents per each). We used two rounds of data collection - on March 12 and March 17, 2021 to examine how our observations concerning search engine performance change over time. On each of these dates, our agents opened the News search page of the respective search engines and entered the "russian covid vaccine" query. Then, the html content of search pages with top 10 search results from each agent was sent to the remote server, where we extracted links to individual search results which were then analyzed using frame analysis.

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:	Makhortykh, Mykola, Urman, Aleksandra
Subjects:	300 Social sciences, sociology & anthropology 000 Computer science, knowledge & systems 000 Computer science, knowledge & systems > 070 News media, journalism & publishing
Language:	English
Submitter:	Mykola Makhortykh
Date Deposited:	17 Aug 2022 08:21
Last Modified:	05 Dec 2022 16:22
Uncontrolled Keywords:	COVID, algorithm audit, search engines, Russia, vaccine, propaganda, disinformation, framing
URI:	https://boris.unibe.ch/id/eprint/171972