<p>The global proliferation of social media platforms has transformed political communication, making the study of online interactions between politicians and voters crucial for understanding contemporary political discourse. In this work, we present a methodological framework for modeling and examining the dynamics of political messaging and voter behavior in online social media. We propose a nuanced categorization of voters who changed their political preferences over time (i.e., swing voters) based on their shifting patterns, as well as a characterization of their susceptibility to propaganda messages. Drawing on the case of the 2022 Italian elections on Twitter, our analysis reveals that during election periods, politicians’ popularity increases, accompanied by notable variations in the use of persuasive language techniques, including doubt, loaded language, appeals to values, and slogans. Swing voters demonstrate higher vulnerability to these propaganda techniques compared to non-swing voters, with distinct vulnerability patterns across various types of political shifts. These findings highlight the nuanced impact of social media on political opinion formation, contributing to a deeper understanding of the interplay between digital political communication and online voter behavior in contemporary democracies.</p>