Airing Grievances Online: Search Engine Algorithms and the Fate of Minorities at Risk

ABSTRACT

This paper examines the role search engine algorithms have in ethnopolitical mobilization. Offline, the news media are vital in a minority at risk (MAR) group's message dissemination. As Tarrow (1994) explains, they are a valuable “external resource.” We argue search engine algorithms perform a similar external resource function mediating what sites get reported as top-ranked search results. The paper addresses whether an association exists between the number of MAR advocacy Web sites protesting a MAR's plight within a search engines' (Google) top 30 search results and the level of grievance a MAR group experiences or the available open opportunity structures. We hypothesize and find that a MAR's grievance level—rather than the availability of open opportunity structures—is vital in explaining the role of search engines in ethnopolitical mobilization.

KEYWORDS:

• Algorithms
• grievance
• minorities at risk
• mobilization
• search engines

Acknowledgments

The authors wish to acknowledge Mack C. Shelley, Iowa State University, for his invaluable editorial assistance. We would also like to acknowledge the support of the University at Albany, State University of New York for the support received for this project through the Faculty Research Awards Program.

Notes

1. It is important to note here that while 320 million pages may be added daily, it is estimated that only 20% will still be available within 1 year (Ntoulas et al., 2004, p.2). The Web's content is constantly changing, and search engine algorithms ensure search results keep apace. An in-depth discussion of search engines is given later in this paper.

2. For example, Web users spend some 13 million hours per month on Google alone (Ntoulas et al., 2004).

3. For a discussion of the Internet's glocalizing ability, see Wellman (2001).

4. An alternative to PageRank is the HITS algorithm; however research finds that it too produces nearly identical results (Ding et al., 2002; Kleinberg, 1999). Other variables that affect the result rankings are site design- the vocabulary it uses, how it uses the vocabulary, whether key terms are in text of in graphics. There are many factors involved, but the greatest emphasis is given to PageRank.

5. The importance placed on providing results that are from authoritative sites is, in part, because of the proliferation of spam. For example, according to Ntoulas, Najork, Manasse, & Fetlerly (2006), some 13.8% of English language Web pages are spam pages. Algorithms have evolved to counter such pages, ranking sites that have reliable information and authority. The countering of spam pages is obviously, therefore, beneficial to MAR protest sites.

6. Not all searches produced useable results. Our total number of groups included is 276.

7. We have, therefore, already replicated online analysis for future years, and we await the availability of our independent variables for those years.

8. The Minorities At Risk project monitors and analyzes the status of politically active ethnic groups, tracking groups' political, economic, and cultural dimensions. For more information about the project, see http://www.cidcm.umd.edu/mar/

9. In future iterations of the project, multiple query constructions, such as group name and main issue of grievance of a group (i.e., land rights, etc.) will be tested. In this initial project, we analyzed one condition to all group searches.

10. We would like to thank one of our anonymous reviewers for this extremely helpful suggestion.