The data hereby presented refers to the Twitter activity of the candidates to the European Parliament and of the most influential party figures/accounts from the main coalitions of France, Germany, Ireland, Italy, Spain and the United Kingdom.
During the electoral campaign we tracked the Twitter activity of all the candidates in #GE2017 and measured the selective salience that candidates accorded to the Brexit theme. Results showed that tweeting “too much” about Brexit may have counterfired.
Election campaigns are complicated and messy in nature, and understanding what candidates are talking about, and why, can turn out to be a complicated task as well. Political parties and candidates often differ substantially in terms of strategy, agenda and incentive structure; moreover, the willingness to emphasise (or understate) a specific issue might depend crucially on other players’ behaviour. This is even more true when we come to electoral campaigns on Twitter. The social network platform provides two-way interaction between candidates and voters: over the course of the campaign, politicians can evaluate the public’s reaction to their messages, voters can share their opinions and campaign strategists can follow voters’ evolving opinions and react accordingly.
On June 8, 2017, roughly one year after the Brexit referendum, British voters went back to the polls for a snap general election. The election was unexpectedly called by Prime Minister and Conservative leader Theresa May: with EU-UK negotiations over Brexit about to commence, May’s surprising move has been read as an effort to consolidate the Tories’ majority in Parliament and May’s own leadership of her party.
Brexit could easily be expected to be a key theme of the electoral campaign. As it turned out, the complicated relationship between the UK and the EU never strayed from the thoughts of British voters, accompanying them throughout the election. We looked at Twitter data to understand how Brexit influenced the agenda, and how it shaped candidates’ strategies.
During the electoral campaign we tracked the Twitter activity of all the candidates in #GE2017. We followed the five major parties according to the polls—Conservative, Labour, Liberal Democrats (Lib Dems), SNP, UKIP—collecting all messages sent by their candidates as well as citizens’ reactions to such messages, expressed in the form of RT retweets and/or replies. As a second step, we used both thematic dictionaries and machine learning classifiers to identify a sub-sample of ‘EU-related’ conversations—that is, all tweets relating to the EU and, most notably, to Brexit.
We measured the selective salience that candidates accorded to the Brexit theme as the share of ‘EU-related tweets’ in overall campaign conversation. As we can see from Figure 1, the salience of Brexit is far from being homogeneous across parties. The Liberal Democrats made Brexit the focus of their campaign, in line with their leader Tim Farron who promised a second EU referendum in his manifesto. On Twitter, Lib Dem candidates dedicated one tweet out of eight (12.26%) to discussing Brexit and the upcoming negotiations with the EU. Theresa May’s Conservatives also appeared to be strongly engaged in this topic (8.53%), as were Ukippers, who were expected to be rather vocal on the subject (11.83%). Strikingly enough though, neither the SNP nor the Labour Party made it above the threshold of 3%. Sensible reasons can be found to account for the low salience accorded to the topic by SNP candidates, who might have been more focused on campaigning for a second Scottish referendum. On the other hand, Labour candidates turned out to be the least interested overall: Labour’s absence from the Brexit battlefield really stands out, especially when compared with the Conservatives.
Fig. 1. Percentage of EU-related tweets by party
Today’s election campaigns—and online campaigns even more so—are highly personalised, and focus to a great extent on the personality of the party leaders. With this in mind, we compared the attention given to the relationship with the EU by the leaders of the two main parties, as reflected by their tweets. As Figure 2 shows, the difference is striking: Theresa May appears to be highly involved, as half of her communication on Twitter (51.92%) revolves around Brexit to some extent. Jeremy Corbyn on the other hand, dedicates about 2% of his tweets to the topic, which is less than that of the average Labour candidate.
Fig. 2. Percentage of EU-related tweets by candidate: Jeremy Corbyn vs. Theresa May
Although things in May’s campaign often seemed to be going in unexpected (and unwanted) directions—the ‘manifesto debacle’, above all—the Conservatives tried to stick to their ‘strong and stable’ mantra, a core message aimed at talking about leadership and, as a consequence, Brexit. On the other hand, Corbyn’s inertia on the subject appears consistent with his previous decision to direct his party to vote in favour of Article 50, and more generally with the idea of accepting the referendum result while advocating a ‘jobs-first Brexit’. This may have led Corbyn to downplay Brexit while campaigning, thus gaining space for issues and topics neglected by the other parties.
A quick look at the word clouds in Figure 3 provides support to this view. For both Conservatives and Labour, we used the text of all tweets sent by the candidates during the campaign. The clouds in Figure 3 give greater prominence to words that appear more frequently, offering a visual summary of the two campaigns. The word ‘Brexit’ is clearly visible in the Conservative discourse, whereas in the Labour campaign the most prominent substantive issues appear to be the ‘NHS’, education and welfare (as in ‘school’, ‘care’, ‘cuts’, etc.).
Fig. 3. Conservative vs. Labour word clouds. Most frequently occurring words in the election campaign.
We have further explored the differences between the two strategies. We used an algorithm to compare the two sets of words, and to highlight the most distinctive features of each set, namely those words that compare frequently in one set but are less prominent (or absent) in the other. We have plotted the results in Figure 4, which allows us to get a better idea of the main issues in the discourse of each of the two parties. The centrality of the debate about a ‘Brexit’ ‘deal’ amongst Conservatives is confirmed, whilst on the Labour side the issue disappears completely.
Fig. 4. Conservative vs. Labour word clouds. Most distinctive words in the election campaign.
But how did citizens react to their candidates’ strategies? To answer this, we tried to measure how candidates’ EU-related messages ‘resonated’ with the public on Twitter. We checked for the volume of retweets each message received, and on this basis compared tweets talking about Brexit and the EU with other conversations. We then built a standardised ‘index of reaction’, indicating how much more (or less) EU-related tweets compared with generic ones were retweeted: a value of 1 means that EU tweets resonated to the same extent as generic ones, while values above 1 indicate a higher resonance of EU-related tweets. Our results are shown in Figure 5.
Fig. 5. Resonance of EU-related tweets vs. other (1= same resonance)
Interestingly enough, and this is true across the parties, Twitter users tended to retweet EU-related content with more enthusiasm than they did other topics. Nonetheless, differences between the parties were again significant. While Labour ‘supporters’ showed only mild interest, amongst Conservatives a Brexit-related tweet was retweeted on average four times more than a general tweet.
The 2017 UK snap election has been billed by many commentators as a ‘Brexit election’. Indeed, the United Kingdom probably never had a general election before where the EU or relations with Europe have been so relevant.
While by no means exhaustive of the online – let alone offline – electoral debate, our analysis provides some insight into how the topic has been discussed on Twitter by British candidates in the run-up to the elections. We found that the EU/Brexit issue is given relatively high salience and priority in the candidates discourse. But there are some caveats: first and foremost, candidates of different parties faced the issue in varying ways and to different extents. The Conservative campaign stuck to the line of the United Kingdom needing a strong leader – May – to take Britain through the Brexit process. The Labour party on the other hand sought to prioritise core issues including schools, social care and the NHS.
Also, reaction towards EU-related messages vary greatly depending on the party. On the one hand, Labour’s strategy to focus its campaign on domestic issues such as health and living standards appears to have paid off. Labour candidates didn’t talk much about Brexit or the EU, and Twitter users who retweeted Labour content didn’t seem particularly engaged by the topic. On the other hand, the greater emphasis accorded to Brexit by the Conservatives generated a more widespread reaction, though maybe at the cost of neglecting other important issues.
In the next instalment of this analysis we will we will look more closely at the voters’ side of the campaign, trying to shed light on how they reacted to politicians’ messages.
Photo Credits CC: TaylorHerring