Agenda-setters and “framing contestants” have been fighting in a new hot battlefield, Twitter, to promote their agenda and make others see the events through the lenses they offer. This study exploits this new agenda-setting avenue by analyzing two years of public newsworthy commentaries (tweets) of selected opinion-shapers and newsmakers residing in the U.S. and in Turkey. Tweets are manually curated by the editors of theplazz.com and nediyor.com news sites, and commentaries are placed below the related news contents. 156,480 curated tweets of 1,400 influential Americans who have commentated on 7,464 headlines between January 2013 and January 2015 are collected from theplazz.com. For the same period of time, 169,849 commentary tweets of 1,296 influential Turks on 12,000 events that have made to the headlines as well as the corresponding news contents published at nediyor.com are scraped. Using these datasets, first, a novel method to quantify the importance of news for a country is introduced based upon the elitist theory of democracy by analyzing the dynamics of the responses to important events that have made to the headlines in these two countries. Then as case studies, overall distributions of commentaries on different topics are examined, in particular, characteristics of the most appealing news to US newsmakers and that of Turks in the years of 2013 and 2014 are discussed. Moreover, coherence and diversity within groups and inter-groups are measured as a wide range of professionals are covered in the datasets, including journalists, academicians, politicians, businessmen and the artists, by asking questions such as “do members of a political party (or journalists of a news group) flock together, i.e. commentate on the same news? How can we measure the distance of a news group to a political party?” Furthermore, as an attempt to recontextualize the agenda-setting theory in this new avenue, I took a network theoretic approach and created weighted co-commentation networks for different groups, where the nodes of a graph are the members of group(s) examined, and the weighted links are the number of common events that they have commentated on Twitter. In support of these theories, application of a modularity based community detection algorithm on such a network consisting of members of 113th Congress reveals that among monitored 36 democratic and 30 republican Congress members, 95% of them are found to be in the same group as their co-party members. Furthermore, these networks enable us to scale the partisanship of individuals as well as groups. Last but not least, qualitative content analysis on the most descriptive events (politically polarized news) shed light onto the peculiar sensitivities of the groups.
Publications: Presented at the 1st Computational Social Science Summit (CSSS15), Evanston, IL May 2015
The source code and the datasets are available on GitHub