From Polarization to Pro-Sociality: Measuring Beneficence in Controversial Online Conversations

This study presents a novel computational approach to quantifying beneficence, defined as a pro-social attitude that positively influences others, in the polarized context of online debates on controversial topics. Starting from a dataset of conversations on Facebook pages on controversial and polarized topics such as vaccination, we used semantic proximity measures to analyze the linguistic landscape, such as confidence, normalized Google distance, and pointwise mutual information. We built an undirected weighted co-occurrence network in which two users are connected if they both comment on the same post. We analyzed polarization trends toward the semantics of beneficence from the point of view of comments, users, and the neighborhood. We found that the formation of echo chambers on the vaccination topic did not correspond to echo chambers on beneficence, with both groups of pro-vax and anti-vax users exhibiting similar levels of beneficence in their discourse. These findings highlight the challenges of bridging the communicative gaps in communities around controversial topics that form echo chambers, showing that opposing parties can share similar beneficence levels. Future research should explore the dynamics of opinion formation and the role of beneficence in preventing and managing hate speeches.