Back To Schedule
Sunday, September 15 • 3:41pm - 4:00pm
“Identifying the opinion leader: Influence, Twitter, and Canadian politics”

Sign up or log in to save this to your schedule, view media, leave feedback and see who's attending!

Studies have attempted to identify influential users on Twitter and within givenTwitter communities. Studies tend to identify influentials as those users who are themost popular (Cha & Gummadi, 2010), the most talked about (Bakshy, Hofman,Watts, & Mason, 2011), or, the instigators of the longest cascades of information(Lerman & Ghosh, 2010). While this has proven effective for identifying what Rogers(2010) would call “influentials” in the diffusion of innovations process, and may sufficefor market researchers looking to send a message to the widest possible public (Watts& Dodds, 2007; Hill, Provost, & Volinsky, 2006), these techniques do little to identifyinfluence on a more local/personal level.

Katz and Lazarsfeld (2006) describe the “opinion leader” as a person able tochange the opinions, attitudes, and/or behaviors of their “everyday associates” (Katz,1957). Opinion leaders use social pressure and social support to exert personal in-fluence. These locally influential individuals are important because they help guidepolitical discussion, integral to the strength of democracy (Dillard, Segrin, & Harden,1989; Mutz, 2006). Further, from word-of-mouth advertising (Li & Du, 2011; Bakshyet al., 2011) to get-out-the-vote campaigns (Middleton, 2006), it has been shown thatpersonal connection increases the chances of a target buying a product or going tovote.

Given the importance of these opinion leaders and relative lack of literature onthe topic this study aims to answer two main questions:

Who are the “opinion leaders” within two Canadian political Twitter communi-ties?Which methods most effectively identify and distinguish users from being either“influentials”, “opinion leaders”, or “followers”?

Responding to these questions, we collected all tweets containing the hashtags#CPC (Conservative Party of Canada, Government) and #NDP (New DemocraticParty of Canada, Official Opposition) over a two week period. From this set, weemployed a qualitative analysis to eliminate users not discussing Canadian politics. We selected the Canadian political Twittersphere given the lack of current researchon influencers within this community, and because, while active, the community issmall enough to conduct meaningful qualitative analysis which serves as our baselinefor comparison.

In order to identify opinion leaders, we first classify users as “influentials,” “opin-ion leaders,” “followers,” or “not political” based on their profiles and most recenttweets. We then use other methods of identifying opinion leaders in order to illustratesimilarities and differences and to assess which methods are most effective.

Self-identification is the traditional mode of identifying opinion leaders (Katz &Lazarsfeld, 2006), and continues to be used in recent work (Norris & Curtice, 2008).We implement this by sending links to a short online survey to all users withinthe network. Measures of indegree (Cha & Gummadi, 2010; Java, Song, Finin, &Tseng, 2007; Romero & Kleinberg, 2010), eigenvector centrality (Weitzel, Quaresma,& Oliveira, 2012; Bigonha, Cardoso, Moro, Almeida, & Gon¸calves, 2010) (socialnetwork analysis), user interaction, and user information sharing with others (Lotanet al., 2011)(content analysis) are methods adapted from studies of network wideinfluential identification. Finally, the clustering coefficient (Lerman & Ghosh, 2010;Java et al., 2007) (social network analysis) is used as an operationalization of thenotion of being socially embedded in ones local community.

We find that each measure has its strengths and weaknesses both in terms ofaccuracy and practical feasibility. Our findings suggest that there are multiple kindsof opinion leaders.


Bakshy, E., Hofman, J. M., Watts, D. J., & Mason, W. A. (2011). Everyone’s an Influencer:Quantifying Influence on Twitter Categories and Subject Descriptors. In Wsdm ’11proceedings of the fourth acm international conference on web search and data mining(pp. 65–74).
Bigonha, C., Cardoso, T. N., Moro, M. M., Almeida, V. A., & Gon¸calves, M. A. (2010).Detecting evangelists and detractors on twitter. In 18th brazilian symposium onmultimedia and the web (pp. 107–114).
Cha, M., & Gummadi, K. P. (2010). Measuring User Influence in Twitter : The MillionFollower Fallacy. , 10–17.
Dillard, J., Segrin, C., & Harden, J. (1989). Primary and secondary goals in the productionof interpersonal influence messages. Communication Monographs, 56 , 19–38.
Hill, S., Provost, F., & Volinsky, C. (2006). Network-based marketing: Identifying likelyadopters via consumer networks. Statistical Science, 21 (2), 256–276.
Java, A., Song, X., Finin, T., & Tseng, B. (2007). Why we twitter: understandingmicroblogging usage and communities. In Proceedings of the 9th webkdd and 1stsna-kdd 2007 workshop on web mining and social network analysis (pp. 56–65). NewYork, NY, USA: ACM.
Katz, E. (1957, March). The Two-Step Flow of Communication: An Up-To-Date Reporton an Hypothesis. Public Opinion Quarterly, 21 (1), 61–78.
Katz, E., & Lazarsfeld, P. F. (2006). Personal Influence: The Part Played by People inthe Flow of Mass Communications. Transaction Publishers.
Lerman, K., & Ghosh, R. (2010). Information contagion: An empirical study of the spreadof news on digg and twitter social networks. In Proceedings of 4th internationalconference on weblogs and social media (icwsm).
Li, F., & Du, T. C. (2011, April). Who is talking? An ontology-based opinion leaderidentification framework for word-of-mouth marketing in online social blogs. DecisionSupport Systems, 51 (1), 190–197. Available from http://www.sciencedirect.com/science/article/pii/S016792361000240X
Lotan, G., Graeff, E., Ananny, M., Gaffney, D., Pearce, I., & boyd danah. (2011). The Rev-olutions Were Tweeted: Information Flows during the 2011 Tunisian and EgyptianRevolutions. International Journal of Communications, 5 , 1375-1405.
Middleton, J. (2006). Middleton 2006 - MoveOn and Voter Mobilization in 2004 — GetOut The Vote. Available from http://gotv.research.yale.edu/?q=node/50
Mutz, D. C. (2006). Hearing the other side: Deliberative versus participatory democracy.New York, NY: Cambridge University Press.
Norris, P., & Curtice, J. (2008, May). Getting the Message Out: A two-step model ofthe role of the Internet in campaign communication flows during the 2005 BritishGeneral Election. Journal of Information Technology & Politics, 4 (4), 3–13.
Rogers, E. M. (2010). Diffusion of Innovations, 4th Edition. Free Press.
Romero, D. M., & Kleinberg, J. (2010). The directed closure process in hybrid social-information networks, with an analysis of link formation on twitter. In In icwsm.
Watts, D., & Dodds, P. (2007, December). Influentials, Networks, and Public OpinionFormation. Journal of Consumer Research, 34 (4), 441–458.
Weitzel, L., Quaresma, P., & Oliveira, J. P. M. de. (2012). Measuring node importanceon twitter microblogging. In Proceedings of the 2nd international conference on webintelligence, mining and semantics (pp. 11:1–11:7). New York, NY, USA: ACM.

avatar for Elizabeth Dubois

Elizabeth Dubois

DPhil (PhD) candidate, Oxford Internet Institute
University of Oxford, United Kingdom

Devin Gaffney

Little Bird, United States

Sunday September 15, 2013 3:41pm - 4:00pm EDT
ROWE 1020

Attendees (0)