Beyond Direct Retweets: Multi-Step Pathways in Italian COVID-19 Twitter

Retweet networks are often analyzed through direct interactions, but online attention may also be shaped by short multi-step pathways connecting users and communities indirectly. In this work, we study the Italian Twitter debate during the first phase of the COVID-19 pandemic by combining a validated reconstruction of discursive communities with a random-walk framework on a weighted directed retweet network. Rather than reconstructing empirical cascades of individual tweets, we use path motifs as a structural tool to compare direct community-to-community connectivity with the endpoints reached by higher-order pathways. The results show that attention is strongly community-centered at the level of direct retweets, but this concentration weakens as path length increases. Longer pathways reveal non-uniform redistribution across communities: some groups become more prominent as endpoints, while others lose relative prominence. These patterns are not fully explained by community size or first-order connectivity alone, and they display relevant directional asymmetries when the network is analyzed in the reversed orientation. Overall, the work shows that looking beyond direct retweets reveals higher-order structures of attention that remain hidden in standard network analyses.

Edoardo Maggioni – PhD Scuola IMT Alti Studi Lucca