Repertoires of digital action and types of party on Facebook. The case of the 2022 French presidential election
Elections
Political Parties
Political Sociology
Campaign
Internet
Methods
Social Media
Big Data
Abstract
French political parties are more and more present on social media, especially during political campaigns. Social media bring together wide circles of supporters from the different political parties (Pena and Gold, 2022). Their uses are strongly influenced by the political orientation of the partisan communities. Right-wing parties can adopt highly centralized online strategies due to their relationship to authority (Vittori, 2020). On the contrary, left-wing parties would have a much more horizontal and collegial presence on social media because of their political ideology (Ignazi, 2017). These differences in the online strategies of the partisan communities are reflected in their “repertoires of digital action” (Chadwick, 2007) and the way they react to online publications. These differences lead them to use specific repertoires in order to promote the mobilisation of the community members, their coordination, their information-sharing behavior or their political discussion (Magin et al., 2017; Theocharis et al., 2015).
The internet and social media favored over time a new structure of political parties, the “digital parties” (Gerbaudo, 2018), especially in Europe (Pirate parties, Podemos in Spain, the M5S in Italy, La France Insoumise in France, etc.). As these parties are recent, they often have fewer resources to communicate and organize their actions. Internet resources allow them to structure themselves (Deseriis, 2020). There is a lack of evidence on the way these digital parties use social media to promote specific repertoires of action and how they manage them to broadcast the themes of their political campaign. In this paper, we support the hypothesis (H1) (confirmatory) that the structure changes of the political parties (traditional party VS digital party) influence the repertoires of digital action of the partisan communities and the way they manage social media to promote specific topics during the political campaign as well as their reception (Blassnig et al., 2021). We also support the hypothesis (H2) that the political effects of these changes in the structure of political parties are amplified according to their political orientation.
Our dataset consists of 26 000 publications published by the administrators of the Facebook political parties pages (n = 852) during the 2022 French presidential campaign. To analyze the repertoires of digital action of the French parties, we used natural language processing machine learning algorithms. We manually labeled part of our dataset for the purpose of fine-tuning a large language model (camemBert). In doing so, we are able to classify the Facebook messages in a semi-supervised manner according to the labels we defined beforehand. First, our results show that the repertoires of digital action vary according to the structure of political parties and their political orientation. Digital parties use social media to coordinate their political campaign and promote their candidate ; Traditional parties use them to broadcast their political programs. Second, the topics broadcast on partisan communities vary depending on their political orientation. Third, the structure of the political parties and their political orientation influence the numbers and types of reactions (like, share, comments) to the Facebook posts.