With just one tap. Network study of tweets dissemination during the war in Ukraine

Polarization of public opinion and the resulting social desensitization may support the Russian invasion of Ukraine. New media, being particularly potent in news diffusion, play a crucial role in amplifying these phenomena. In this paper, we study social networks that emerged on X during selected war events. Community detection algorithms were applied to investigate echo-chambers. To analyze network properties associated with social desensitization exponential random graph models were used. The research revealed that initial factors, such as established connections and shared acquaintances, continued to be influential in predicting social connections on Twitter. However, over time, communities tended to form around more contentious influencers. The observed polarization extended across all networks, displaying recently an increased level of radicalization. Additionally, we detected a decline in engagement that correlated with a reduction in the likelihood of establishing connections based on specific characteristics in the estimated models. Furthermore, likelihood of forming ties for certain homophily factors diminished. These results indicate that exposure to violence is linked to reduced affective responses. While these processes are disturbing in terms of their social consequences, some solutions already available in scientific contributions could be effectively used to counteract them.