The successful use of online material in political campaigns over the past two decades has motived the inclusion of social network websites—such as Twitter—as an integral part of the political apparatus. Indeed, after the growth of candidate websites in 1996, Jesse Ventura’s e-mail campaign in 1998, the online fund-raising for John McCain’s campaign in 2000, the use of blogs for Howard Dean’s campaign in 2004, and the pivotal role of social media in the Obama ‘Yes We Can’ campaign in 2008, Twitter and social media in general have established themselves as a key communication channel in the political arena.
In the same way in which social media has become a primary source of news and a forum for journalists to publish, political analysts are increasingly turning to Twitter as an indicator of political opinion. On Twitter we can “see” what is actually being said without having to explicitly draw attention to it through questions as in the case of current methodologies for polling, which tend to consist of questions over the phone, online or face-to-face. Through Twitter posts, we can observe influencing topics, people, and events without influencing them ourselves.
At the School of Computing and Digital Media at Robert Gordon University (Aberdeen), a group of researchers led by Professor Ayse Göker are addressing research questions at the cutting edge of news, political science and social media. Currently, they are investigating whether Twitter can provide a platform for political deliberation online, and how accurately Twitter can inform us about the electorate’s political sentiment. This work builds on the large European (FP7) project, SocialSensor, which ran from 2011 to the beginning of 2015 and focused on aspects of news, elections and political engagement.
Göker’s team work on the Scottish Independence Referendum (2014-2015) has sparked off a unique collaboration among computer scientists, journalists and academics in political science and the humanities, who followed up the political chatter in social media for almost one year, starting a few months prior to the Scottish Referendum in September 2014 and continuing on to the General Election in May 2015. Nearly 28 million tweets were collected through hundreds of hashtags where people and organisations voiced their views around these events and in between. The team created a list of 300 Twitter accounts, consisting of Scottish journalists, politicians, bloggers and public bodies interested in the debate. All public tweets sent to and from those accounts during the debate were collected, and also all tweets containing the hashtags #indyref, #bettertogether, and #yesscotland along with every tweet geo-tagged as being published within Scotland. Throughout this, the team identified and monitored trending topics.
Additional post-referendum work has been done to carry out a qualitative investigation of engagement among young voters—aged 16-19 at the time of the Referendum. Working in association with the Scottish Youth Parliament, the team investigated whether the high levels of political engagement during the Referendum translated into broader engagement with politics. Further work has also fostered collaboration with colleagues from the Department of Communication, Marketing and Media at Robert Gordon University, who were interested in the response to televised debates associated with the 2015 General Election.
The algorithms developed by the RGU team have been ranked as top performing in benchmarking initiatives such as the SNOW Challenge. The team maintains and aims to extend a platform for real-time indexing and search of text and multimedia information in the social Web, which also focuses on context-aware retrieval, and data analysis of social media content.
Social media is now central to how people consume and produce election news, and this ground-breaking research will provide us with deeper insights into how people are developing their own understanding of elections, which goes beyond the traditional news media agenda and might allow us to spot early trends which could affect the overall outcome. For some authors, the predictive power of Twitter for real-world outcomes has been exaggerated. For others, social media can be construed as a form of collective wisdom that can indeed be used to make quantitative predictions; yet, work is still ongoing and critical questions lie ahead.
This article was written by Marco Palomino, Ayse Göker and Michael Heron.
The authors acknowledge the contribution made by Sarah Pedersen (media), Simon Burnett (information management) and Graeme Baxter on TV debates during the elections. When working on Scottish youth and political engagement, the authors acknowledge the collaboration with Elizabeth Tait (information management), Iain MacLeod (political science), Graeme Baxter and Peter McLaverty (political science).
Ayse Göker’s team has been composed of Carlos Martin, Michael Heron, David Corney, Malcolm Clark and Marco Palomino.