The impact of social media discussion threads on public sentiment

This project conducted research on the impact of widely accessed social media discussion threads about people’s sentiments and opinions on Scottish independence and devolution, political leaders and their parties. The project was particularly focused on uncovering who is affected, and when and how they are affected. It also explored changes over time as the debate was building ahead of the referendum and it juxtaposed web sentiment alongside existing poll data and political context.

The data collection included approximately 20 separate threads from prominent front-page Scottish independence news stories between January 2011 and December 2013 covered by, and accessible from, the widely accessed BBC news on-line site. The data was analysed by both content analysis and sentiment and opinion mining of text using both NVivo and the R programme (both of which can identify patterns in complex text).

This project was part of the ESRC-funded Centre on Constitutional Change


Findings from this research were published in the article Social media and the 2014 Scottish independence referendum: flaming keyboards or forums for deliberation? published in the Electoral Studies Journal, June 2015, Volume 38, Pages 192-205

Findings were published on the Centre on Constitutional Change blog.

Findings were also presented in a TedX talk ‘The Impact of Social Media in Political Debate’.

Dr Mark Shephard also produced a policy briefing in 2015 which draws on the findings from this project – Improving on line political engagement for effective public engagement


There was very extensive media coverage of the results, notably at key points such as:
7-9 November 2013: Festival of Social Science event attracted coverage in The Guardian and Mail on Sunday.
1 June 2014: Sunday Herald news story on things to avoid doing on-line with social media.
12 June 2014: Scotland 2014 (BBC2 Scotland): Studio discussion by a former adviser to the First Minister refers to the research as evidence that no side is culpable more than the other on negative postings.
14 July 2014: Scotland Tonight (STV): Item on the use of social media by both sides of the independence referendum campaign. Includes discussion involving Mark Shephard


The project had five main impacts:
We provided extensive and rigorously analysed evidence that, despite frequent claims in public, the on-line debate was on the whole being conducted fairly and with reasonable respect, though as robustly as any political debate.

Nevertheless, the on-line debate also did not use evidence and there is little sign that the ideals of deliberative democracy were being achieved, despite the possibilities offered by some on-line forums.

Our work in producing teaching materials for schools (along with the two other AQMeN Scottish Independence projects under the ‘Future of the UK and Scotland’ scheme) have been designed to help students engage critically with social media (for example, spotting fallacious posts).

We have also taken these materials to a wider audience in the UK by sharing our best practice guidelines with the Beyond the Youth Citizenship Commission Re-port: Young People and Politics (Political Studies Association) and beyond the UK, by broadening the applicability and reach of the guidelines through the TedX talk (mentioned above).

We have produced materials and run training sessions for AQMeN – giving others hands-on experience of the mixed-methods techniques we deployed using our own data examples.

In September 2017, Dr Shephard submitted evidence to 3 UK parliamentary committees – Standards in Public Life; Citizenship and Civic Engagement and the Political Polling and Digital Media committee. These submissions drew on findings from this project.

Policy relevance:
Public engagement is an underpinning principle of the Scottish Parliament, and politicians and SPICe (the Scottish Parliament Information Centre) will have much to learn from the project’s findings on what undermines (as well contributes to) attempts at rational engagement in the independence debate. The findings should assist with guidelines for public engagement practices with politicians.

Mark Shephard
Lindsay Paterson