There is increasing debate over the ways in which search algorithms and patterns of user behavior around Internet search tools may support or undermine the quality and diversity of information accessed by citizens in ways that could effect the vitality of democratic processes for the better or worse. For example, does search help citizens to obtain better information about political issues, or does it bias what citizens know in ways that could distort democratic choice, such as through disinformation? Too often, answers to such questions are anchored in deterministic perspectives on technology and overly simplistic models of the democratic process. Moreover, there is limited empirical research on these issues, creating a need for theoretically informed empirical research focused on the actual role of search in political opinion formation in liberal democratic political systems of the digital age.
The Quello Search Project has focused on the role of online search in the context of other media and information technologies, such as social media, in shaping access to and the influence of political information. Data was gathered from six EU countries and the US, using Web-based online surveys as well as aggregated trace data from search engines. Quello’s MSU research team is managing the project, which included academics from the University of Oxford and the University of Ottawa as well as MSU.
The major source of data was an online survey of seven nations. The four-person project team designed the survey questionnaire to address a wide array of questions about when and how Internet users employ search and related media and information technologies to find general information as well as political information in particular. The survey questions included the following topics:
• how use of various media, including search engines, are related to political opinions;
• how users vary in their use of search and their interest in politics;
• how different types of Internet users perceive search in comparison to other media, such as social media, broadcast and print news;
• how search impacts the views of Internet users, such as whether there is any real or perceived bias in what users see online;
• whether search engines play a major or minor role in opinion formation;
• whether personalization of search is understood, and whether this is viewed as a benefit or harm by users.
The team worked with ICM, an opinion research firm in London, UK, to field the survey questionnaire online across Europe and the US. ICM is an experienced and well-regarded firm with links with survey research organizations across Europe and North America that have created online panels of survey respondents. These panels allowed us to obtain responses from individuals with known characteristics, such as age and gender, ensuring that the respondents more closely represent the online population in each nation.
The EU countries included Britain, France, Germany, Italy, Poland, and Spain. In each country, the survey obtained valid responses from about 2,000 Internet users, for a total sample of over 14,000 individuals.
The second source of data was aggregated trace data from search engines, such as Google Trends. This data enabled the team to gather information on how people search for political information in comparison with searches for other information. We were able to analyze these data by country for comparisons within the EU and with the US. We looked for global and nationally specific patterns in the kinds of terms that were most often searched for and differences in types of political searches. These findings complement and expand on the survey research results on the part played by search in shaping political opinions.
The project began in the fall of 2016. Surveys and a https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2960697″>final report were completed in early 2017. A series of academic analyses and publications are being conducted on the basis of this cross-national comparative data set along with the complementary trace data analyses.
The descriptive and multivariate findings cast doubt on deterministic perspectives on search. We find that technology matters – search indeed plays a major role in shaping opinion – but it is not deterministic. For example, the thesis of a filter bubble is overstated, as our pattern of findings counter this expectation. For instance, search engines are among an array of media consulted by those interested in politics. Another more sociotechnical deterministic narrative is around the concept of echo chambers, where users enabled by increased media choice and social media tend to surround themselves with the viewpoints of likeminded people. Our evidence contradicts this view as well. Most of those who search for political information expose themselves to a variety of viewpoints. Results also demonstrate that research has tended to underestimate the social shaping of technology. National cultures and media systems play an unrecognized role, as well as individual differences in political and Internet orientations. This study shows how overestimating technical determinants while underestimating social influences has led to disproportionate levels of concern over the bias of search.
The findings suggest that misinformation can fool some search engine users some of the time, suggesting that a sizeable group of users could benefit from more support and training in the use of search engines. Also, the findings should caution governments, business, and industry from over-reacting to panic over the potential bias of search in shaping political information and opinion.
This research project has been supported by Google, and conducted by a multidisciplinary team that has previously worked together on a number of projects at the University of Oxford and Michigan State University. The team was comprised of researchers from the fields of political science, sociology, Internet studies, and information & communication sciences. The research team also brought together strong expertise in the theories and methodologies used to analyze and discuss the research questions. The project team included:
• William Dutton, Quello Center, Michigan State University, USA
• Bianca Reisdorf, Quello Center, MSU, USA
• Elizabeth Dubois, University of Ottawa, Canada
• Grant Blank, University of Oxford, Internet Institute (OII), UK
Two research assistants supported the project, including:
• Craig Robertson, Information and Media Doctoral Program, MSU, USA
• Sabrina Ahmad, Joint Honours B.A. Communications and Sociology, University of Ottawa, Canada
Project Documents and Papers
Survey Questionnaire Instrument (English Version) is posted here: QGS161129 V2.2 FINAL 211216 b_w.
Dubois, E., and Blank, G. (2017), ‘The Echo Chamber is Overstated: The Moderating Effect of Political Interest and Diverse Media’, Information, Communication & Society, 21(5): https://www.tandfonline.com/doi/abs/10.1080/1369118X.2018.1428656
Dubois, E., and Blank, G. (2018), The Myth of the Echo Chamber, The Conversation, March: https://theconversation.com/the-myth-of-the-echo-chamber-92544
Dutton, W. H. (2017). Fake news, echo chambers and filter bubbles: Underresearched and overhyped. The Conversation, 5 May: https://theconversation.com/fake-news-echo-chambers-and-filter-bubbles-underresearched-and-overhyped-76688
Dutton, W.H., Reisdorf, B.C., Dubois, E., & Blank, G. (2017). Search and Politics: The Uses and Impacts of Search in Britain, France, Germany, Italy, Poland, Spain, and the United States. Quello Center Final Report for the project “The Part Played by Search in Shaping Political Opinion”, supported by Google, UK, in collaboration with Google Inc. East Lansing, MI: Michigan State University. Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2960697.