Do politicians try to set the agenda on social media? Are they selective on the topics they talk on Twitter? If so, to what extent? To answer these questions I conducted a study using curated tweets of Congress members.
The results were shocking: I showed that political groups of the monitored members of 113th Congress can be predicted with more than 95% accuracy just by looking at the topics they talk (and do not talk) about.
Data & Method: This study uses a unique dataset collected by the author, and effectively uses social network analysis to measure politicians’ agenda similarity. Our entire dataset consists of 7,376 news published between January 2013 and January 2015 on theplazz.com news site, as well as 156,480 commentary tweets of 1442 newsmakers on these news events. All the tweets are curated manually by theplazz.com editors. Using this dataset, I first created a co-commentation network where the vertices are congress people and the edge weights are the number of common newsworthy events (headlines) that they have commented during 113th Congress. Then, I detected the communities in this network by employing a weighted modularity clustering algorithm. Using this method, two groups are detected when the highest modularity achieved, as a result, 62 of the 65 monitored Congress members are found to be in the same group as their co-party members.
Suggested citation: Oz, Talha, 113th Congress as News Commentators on Twitter (June 17, 2015). American Political Science Association, Political Networks Workshops & Conference 2015. Available at SSRN: https://ssrn.com/abstract=2918958