Networked Partisanship: Using Framing to Measure Polarisation on Social Media

The use of Facebook and Twitter (collectively referred to as "social media" hereafter) for political
discourse, has been argued to directly encourage the partisan polarisation of online communities. If
verifiable, this development has the potential to redefine public debate in the UK's political arena.
To measure this perceived phenomenon, an examination is required of social media's transformative
effect on the discursive relationship between political actors, news media and the public, which defines
political issues in the public sphere.
Firstly, news and opinion is produced by an ideologically-diverse array of political and news media
sources. Secondly, social media users experience news interactively, and collectively respond to political
information through shared subjective expressions of opinion and emotion. These changes to the way
political information is communicated suggest that public opinion on social media is established through
a reciprocal dialogue between news producers and a respondent public.
However, this implied connection is not cohesively studied in current academic research. The influence of
"elite" news sources (the social media profiles of politicians, parties, journalists and news organisations)
and public responses to political content (including discursive polarisation) are primarily studied
independently of one other.
To address this absence, this study presents a pioneering methodology to examine how perceptions of
political issues resonate across news sources and public conversation on social media. By holistically
considering the proliferation of partisan opinions in discourse between political, news media and public
actors, it will reveal the extent of political polarisation on a national level. The scope, length and unique
design of this project will make a significant academic contribution, addressing a scarcity of UK based
studies.
News articles and social media posts from elite sources, and posts/comments by public users concerning
the three most prominent issues in British politics (defined by opinion polls) will be collected over an 18-
month period and analysed for shared political evaluations, i.e. "frames". Frames are understandings of a
societal issue communicated through media discourse. In political terms, they refer to a combination of a
perceived problem, its cause and a judgement on the situation.
Using software that finds commonly used words and the context in which they appear, identification of
prominent "problems" and "causes" expressed in elite sources will be made into data points ("coded").
Text analysis software will be used to measure the positive and negative associations ("sentiment")
between these problems and causes. This will statistically indicate an evaluative judgement, completing
the frame. Influential frames across media sources will be found using advanced quantitative methods
(e.g. hierarchical cluster analysis).
Once the lexical and semantic structure of common political arguments are found using this method,
social media posts by public users will then be compared to see which frames are reciprocated and are
therefore resonant in social media conversation.
The results will indicate whether partisan polarisation is present and influential on British social media. A
hypothesis will also be assessed: that ideological news production by elite sources, and collective public
responses united by shared opinion and emotion, will encourage the wider diffusion of partisan
4 / 23
perspectives. A deeper knowledge of how partisan opinions are spread through social media will find
application in civil society, informing constructive responses to societal concerns, including political
extremism, discursive isolation and the perceived decline of agreeable "truth" in political communication.