Euroscepticism, or negative attitudes towards the European Union (EU), has significantly increased since the early 1990s, particularly among far-right anti-establishment parties. This surge coincides with major shifts in mass media and political communication technologies, including the rise of social media. While economic factors such as the 2008 financial crisis and rising inequality have also contributed to the growth of Euroscepticism, this paper focuses on the role of online political activity, particularly through social media. The authors hypothesize that exposure to online political information reinforces Eurosceptic views primarily among less-educated individuals who may be more susceptible to divisive messages and less equipped to critically evaluate information. Education is viewed as a crucial factor, as it fosters social capital, trust in institutions, and the ability to navigate complex political landscapes. The study aims to examine this interaction between online political engagement, education, and the development of Eurosceptic beliefs.

The paper reviews existing literature on Euroscepticism, its relationship with populism, and the impact of online political communication. It defines Euroscepticism as encompassing both soft (lack of trust in EU institutions) and hard (opposition to European integration) forms. While noting the association between populism and Euroscepticism, the authors emphasize that not all Eurosceptic views are populist. The literature highlights the role of social media algorithms in amplifying divisive content, influencing user engagement, and potentially increasing polarization. Studies cited show correlations between educational levels and attitudes towards European integration, with less-educated individuals tending to be more Eurosceptic. The authors argue that this is because highly educated individuals possess more cognitive, creative, and functional skills to adapt to the complexities of internationalized environments and are better at discerning credible from false news. They posit that less-educated individuals are more vulnerable to the negative effects of online political propaganda, which they are less equipped to analyze critically.

The study utilizes data from two primary sources: the 8th and 9th rounds of the European Social Survey (ESS) and several iterations of the Italian Multipurpose Household Survey (MHS). The ESS data provide information on trust in the European Parliament, preferences for leaving the EU, years of education, and exposure to online political activity. The MHS data allow for a more nuanced analysis by distinguishing between general online political activity and activity specifically through social media. The authors employ ordered and binary probit models for the ESS data and ordered probit regressions for the MHS data to investigate the relationships between Euroscepticism, education, and online political engagement. They control for various demographic and socioeconomic factors. To address potential endogeneity concerns—that is, the possibility that individuals predisposed to Euroscepticism are more likely to engage with online political information—the authors instrument the exposure to online politics and social media variables using indicators of internet connection speed and accessibility. This helps assess the causal relationship between these factors and Eurosceptic attitudes.

Analysis of the ESS data revealed that exposure to online political information is significantly correlated with Eurosceptic attitudes (lower trust in the European Parliament and preference for leaving the EU) only among individuals with less than 12 years of education. The interaction effect between online political engagement and education was significant, indicating that the impact of online exposure on Euroscepticism is strongest among the less educated. The MHS data further showed that the use of social media for obtaining political information, as opposed to general internet use, is specifically linked to lower trust in the European Parliament among less-educated individuals. The negative correlation between social media use and trust in EU institutions was robust to the inclusion of controls for self-assessed computer skills, and it remained significant even after instrumenting the exposure to online political activity using proxies for internet connection speed and availability. The results indicate that the average marginal effects of online political activity via social media are clearly negative for values of trust reaching at least the middle-scale position, especially for low-educated individuals. The instrumental variable approach suggests that standard estimates may underestimate the true effect.

The findings confirm the hypothesis that the impact of online political engagement on Euroscepticism is significantly moderated by education levels. The effect is particularly pronounced among those with lower formal education, who may be more susceptible to divisive messages and propaganda spread through social media. The results highlight the differential impact of various forms of online political activity; social media's role in fostering Euroscepticism appears more substantial than general internet usage. These findings underscore the importance of considering how the design of social media platforms can unintentionally amplify political polarization and extremism, particularly among vulnerable segments of the population. This has implications for understanding the rise of populist movements and the erosion of trust in democratic institutions.

This study demonstrates the significant interaction between education, social media use, and the formation of Eurosceptic attitudes. Low education amplifies the negative effect of social media exposure on trust in European institutions. Future research could investigate the direct link between social media use and voting behavior in elections and referenda to better understand the impact of these findings on actual political outcomes. Further exploration of the relationship between social media use and other socially relevant behaviors, such as vaccination rates, is also warranted. The authors conclude by emphasizing the need for policy interventions aimed at improving media literacy and promoting responsible online engagement to mitigate political polarization and maintain the stability of democratic institutions in the digital age.

The study acknowledges limitations regarding causality due to the observational nature of the data and the potential for endogeneity in the exposure variables. While instrumental variable techniques are used to mitigate this, the authors acknowledge that residual endogeneity might still exist. The data used primarily focuses on attitudes rather than concrete political actions, and the focus is limited to Euroscepticism in relation to the EU. Generalizability of the findings to other contexts may be limited given that the MHS data used are primarily from Italy.