Notes on the Europeanization of Twitter – Part I

Social media play an increasingly important role in today’s politics. The micro-blogging platform Twitter, in particular, has become an autonomous political space, in which political issues are publicly discussed and where interactions between politicians, journalists, intellectuals and the general public take place on a continuous basis. However, whether or not there is such a thing as a genuinely “European Twittersphere”– an arena where European issues are debated within and across member states–is still an open question.

Although the State of the Union speech should represent, for EU citizens, a moment of public reflection about the Union’s achievements, priorities and challenges, the topic holds rather low salience outside the circle of insiders and connoisseurs.

In this and a forthcoming article we try to shed some light on this topic. We begin here with an analysis of Twitter discussions related to two recent high profile European events, namely the State of the European Union address (SOTEU) that Commission President Jean-Claude Juncker delivered before the European Parliament on September 14, and and the summit of EU heads of state and government–except the United Kingdom–held in Bratislava two days later.

Supranationalism is for connoisseurs

To isolate discussions around the two events, we collected all tweets featuring, respectively, the SOTEU official hashtag (#soteu) and the two hashtags #bratislavasummit and #bratislava. Each collection covered the five days before and after the respective event. We then processed these two sets of tweets through an algorithm that estimates the location of the sender by analysing tweet’s metadata. This allowed us to isolate 56,165 tweets related to the State of the Union and 17,632 tweets about the Bratislava summit as coming from one of the 28 member states.

Figure 1: Temporal distribution of SOTEU tweets

As Figures 1 and 2 show, the temporal distributions of tweets spiked–unsurprisingly–on the days in which the two events took place. As it has been noted elsewhere, the shape of the distribution of #soteu tweets (one almost without tails) is revealing of the nature of the event. Although the State of the Union speech should represent, for EU citizens, a moment of public reflection about the Union’s achievements, priorities and challenges, the hashtag #soteu was hardly at all used outside of the event. This seems to suggest that the topic holds rather low salience outside the circles of insiders and connoisseurs.

Figure 2: Temporal distribution of Bratislava tweets

On the other hand, conversations about the Bratislava summit (an intergovernmental meeting in which national leaders bargained on policy options and decisions), show a different temporal pattern. Although unappealing at first sight–these occasions are often depicted by Eurosceptics as opaque and useless exercises of intergovernmental coordination–this event triggered a more sustained reaction on Twitter. The summit was preceded by minor surges in conversations, and followed by a more persistent reaction, with specific country idiosyncrasies, probably indicating spillovers of the European event into the national political and media agendas.

Europeanization through national issues

A thorough analysis of the nature of the various national Twitter discussions is beyond the scope of this article. However, we can gain some insight into the content of the conversations by looking at the hashtags contained in them. Conversations about #soteu were focused on topics covered by Juncker in the speech, and hashtags seldom referred to political figures other than the Commission President. Tweets about the Bratislava summit, on the other hand, referred more often to key national political players (#Merkel, #Renzi, #Hollande were widely used) and specific political topics and issues, such as Brexit, which appears as a popular hashtag in every national set.

The only exception is the United Kingdom, whose share of conversations about SOTEU and Bratislava is significantly lower compared to the random sample. This is particularly surprising in the case of the Bratislava summit, which was largely devoted to the discussion of post-Brexit scenarios.

As an example, Figure 3 shows two distinct word clouds containing the two sets of hashtags for Italian tweets. As one can easily see, unlike SOTEU discussions, Bratislava tweets are largely a reverberation of domestic discussions and quarrels (e.g. #renzichiediscusa; #renziservo; #ventotene). Hashtags connected to a wider European discourse (e.g. #brexit, #juncker) are present in this set, but hold lower saliency.

Figure 3: Bratislava and SOTEU wordclouds

The Brussels bubble and the British hole

Figure 4 offers a glimpse of the relative popularity of EU-related issues across European countries by aggregating the total volume of tweets by country. Unsurprisingly, in both samples the highest share of conversations originated from Belgium, the country that hosts most EU institutions. Nonetheless, the relatively higher popularity of the #soteu event in this country further confirms the specialized nature of this event in particular.

Figure 4: Distribution of SOTEU and Bratislava tweets by country

To better understand how differently national Twitter spheres behave when it comes to European events, we compared the above two distributions with the distribution of general Twitter conversation across countries, which we estimated through a random sample of tweets collected and annotated in comparable conditions. The results are shown in Figure 5.

Figure 5: National distribution of SOTEU and Bratislava tweets vs. random sample.

The magnitude of the Twitter-related “Brussels bubble” is even more striking if we compare the portion of Belgian tweets on SOTEU and Bratislava with the random sample. For most other countries, on the other hand, the relative salience (proxied by share of tweets) of European issues is only slightly higher or lower than average, possibly indicating an evenly spread interest in European topics across the EU.

The only major exception to the foregoing is the United Kindgom, whose share of conversations about SOTEU and Bratislava is significantly lower compared to the random sample. This is particularly surprising in the case of the Bratislava summit, which did exclude Britain, but was largely devoted to the discussion of post-Brexit scenarios among the remaining 27 member states. Future analyses will tell us whether this is an exceptional pattern or, on the contrary, it signals that some sort of “Brexit from Twitter” has already happened.

In conclusion

Three main findings can be summed up from the foregoing analysis. In the first place, looking at Twitter conversations it seems that it is those intergovernmental events involving more directly the national level, like the Bratislava summit, that are more able to sustain a prolonged debate on European issues. At the same time, discussions revolving around supranational events such as the SOTEU seem to generate more “purely” Europeanized conversations, while reactions to intergovernmental events tend to be mediated by national issues and subjects. Finally, our results suggest that in most EU member states Twitter debates are Europeanized to a similar degree, with two big exceptions: Belgium and the UK. The former displays a much higher than average degree of Europeanization, most likely due to a “Brussels bubble” effect. The latter, conversely, seems quite aloof from conversations on European matters.

In the next instalment of this analysis we will look more closely at the extent to which European conversations are structured transnationally, and the interactions between masses and elites regarding European matters.

Photo Credits CC: European Council

Also published on Medium.