This paper aims to clarify the dynamics of EU politicisation during the 2017 election campaign in three EU countries: UK, France and Germany. The aim of the paper is twofold. On the one hand, it assesses the extent of EU politicization within the three considered countries (RQ1), explored in terms of (a) saliency, (b) contentiousness and (c) public resonance of EU related issues. On the other hand, the paper investigates on the political determinants of EU politicization (RQ2). According to literature, EU politicization is the result of parties’ strategies, in particular during election campaigns. Extreme parties within the ideological spectrum are more incentivized to take a critical stance towards the EU (Hooghe & Marks 2009). Relying on this, we expect candidates from parties located at the extremes of the political spectrum (H1) to be more likely to enhance EU-related issues in their tweets; (H2) to be more likely to make use of critical tweets on EU-related issues; (H3) to be more effective in boosting public resonance on EU-related issues.
Twitter represents nowadays an ideal environment for testing these hypotheses (Nulty et al, 2016). Our methodological approach combines the use of supervised machine-learning (based on Natural Language Processing technology) and qualitative data analysis of the electoral debate on Twitter. We employ an original dataset containing over 3 million tweets. For each of the general election considered, our data features all tweets sent by all candidates of parties achieving at least 3% in pre-electoral polls: we collected 177,372 tweets by UK candidates, 148,758 tweets by French candidates and 59,256 tweets by German candidates, together with all retweets and replies to such tweets. Data were collected during the 14 days prior the election day.
