Dataset for the paper Emotional drivers of misinformation consumption and toxicity on YouTube (Santoro et al.)

Dataset for the paper Emotional drivers of misinformation consumption and toxicity on YouTube (Santoro et al.)<br><br>The dataset consists of YouTube comments in Italian that have been collected, pre-processed, and analyzed for emotional content and related features. The data is stored in both raw and processed formats for the compatibility with the code of the study.<br>Abstract<br>Emotions shape online discourse and user behavior, yet the emotional interplay between misinformation and toxicity remains understudied. This paper analyzes one million comments on Italian YouTube videos to examine Emotions shape online discourse and user behavior, yet the emotional interplay between misinformation and toxicity remains understudied. This paper analyzes one million comments on Italian YouTube videos to examine how users react to reliable and questionable information, focusing on emotional expression and comment toxicity. Drawing on Plutchik's psychoevolutionary framework and combining network science with human-centered AI, we identify distinct dynamics among users who predominantly engage with misinformation or mainstream sources. Emotions shape online discourse and user behavior, yet the emotional interplay between misinformation and toxicity remains understudied. This paper analyzes one million comments on Italian YouTube videos to examine how users react to reliable and questionable information, focusing on emotional expression and comment toxicity. Drawing on Plutchik's psychoevolutionary framework and combining network science with human-centered AI, we identify distinct dynamics among users who predominantly engage with misinformation or mainstream sources. Comments on questionable sources convey more fear and less joy, with misinformation-prone users especially prone to expressing fear-tinged anger. We introduce the concept of \emph{emotional arborescence}---capturing how secondary emotions emerge from basic ones---and find that mainstream-prone users exhibit 25\% greater emotional diversity. These findings offer new insights into the emotional dimensions of online misinformation and toxicity, and inform strategies to mitigate their impact on public discourse.Comments on questionable sources convey more fear and less joy, with misinformation-prone users especially prone to expressing fear-tinged anger. We introduce the concept of \emph{emotional arborescence}---capturing how secondary emotions emerge from basic ones---and find that mainstream-prone users exhibit 25\% greater emotional diversity. These findings offer new insights into the emotional dimensions of online misinformation and toxicity, and inform strategies to mitigate their impact on public discourse.how users react to reliable and questionable information, focusing on emotional expression and comment toxicity. Drawing on Plutchik's psychoevolutionary framework and combining network science with human-centered AI, we identify distinct dynamics among users who predominantly engage with misinformation or mainstream sources. Comments on questionable sources convey more fear and less joy, with misinformation-prone users especially prone to expressing fear-tinged anger. We introduce the concept of \emph{emotional arborescence}---capturing how secondary emotions emerge from basic ones---and find that mainstream-prone users exhibit 25\% greater emotional diversity. These findings offer new insights into the emotional dimensions of online misinformation and toxicity, and inform strategies to mitigate their impact on public discourse.