31 May 2017
- New analysis reveals role of ‘junk news’ and Twitter bots as poll nears
Analysis by researchers at Oxford University’s Oxford Internet Institute shows that content about the Labour Party is dominating traffic on Twitter during the UK general election campaign.
Hashtags such as #VoteLabour and #JezzWeCan are outperforming the likes of #VoteTory and #StrongAndStable, according to the new study.
But the ‘junk news’ that blighted the US presidential race is also playing a role in the UK election, accounting for almost 13% of relevant content shared.
The researchers analysed more than 1.3 million tweets produced between 1 and 7 May 2017, using hashtags associated with the primary political parties in the UK, the major candidates, and the election itself.
- Labour Party hashtags appear the most often in political tweets (39.7% of party-specific tweets).
- Conservative Party hashtags account for 26% of the total.
- The SNP is ‘disproportionately’ represented in Twitter conversations (19%), given the party’s size.
- UKIP: 9.6%; Lib Dem: 5.7%.
- One in eight tweets (12.3%) about UK politics are generated by bots (highly automated accounts).
Labour is, by some margin, attracting the greatest interest from highly automated accounts (Labour: 21,661 tweets; SNP: 13,819 tweets; Conservative: 13,409 tweets).
- However, the analysis does not show whether automated accounts are run to reflect positively or negatively on a particular party or candidate, nor who is running these accounts.
- ‘Junk news’ accounts for 12.6% of relevant content shared (once non-political and spam content is removed).
- This can include fake, hyper-partisan or emotionally charged news content, much of which is deliberately produced false reporting.
- Social media users in the UK share four links to professionally produced news and information for every one link to junk news.
- UK users are sharing better-quality information than US users (an Oxford Internet Institute study showed a 1:1 professional to junk ratio in Michigan in the run-up to the US election) but lower quality than German or French users.
Senior researcher Professor Philip Howard, from the Oxford Internet Institute, said: ‘Social media is playing an increasingly important role in public life, and the combination of automation and propaganda is a powerful tool that can have a significant and pernicious impact on people’s opinions during important events such as elections and referenda. Indeed, the World Economic Forum recently identified the rapid spread of misinformation online as one of the top ten risks facing society.
‘This study shows that the situation in the UK isn’t as bad as it was during the US presidential campaign, and is more in line with recent findings from elections in France and Germany. But there is still an issue more generally over the quality of information being shared by social media users.’
Co-authors Monica Kaminska and John Gallacher, of the Centre for Doctoral Training in Cyber Security at Oxford University, added: ‘Young people regard social media as their primary source of political news, and Twitter in particular has become very popular in the UK. Analysing the content shared across the site therefore gives us an important snapshot into political life online.’
For further information or a copy of the paper, please contact Stuart Gillespie in the Oxford University press office at firstname.lastname@example.org or on +44 (0)1865 283877.
Monica Kaminska (study co-author): email@example.com
Notes to editors:
‘Junk News and Bots during the 2017 UK General Election: What Are UK Voters Sharing over Twitter?’ is a Data Memo produced by researchers at Oxford University’s Oxford Internet Institute.
Data Memos are designed to provide quick snapshots of analysis on current events in a short format. They reflect methodological experience and considered analysis but have not been peer reviewed.
The study is part of the Oxford Internet Institute’s Computational Propaganda Project, which involves international and interdisciplinary researchers in the investigation of the impact of automated scripts – computational propaganda – on public life.