Online social platforms are targeted by centrally coordinated disinformation campaigns. Even though a large spectrum of coordination can be observed in social media, usually literature focuses exclusively on highly coordinated activity. Coordination networks are often built by linking accounts sharing a minimum number of times the same (or similar) content within a small timeframe. However, these multiple thresholds implicitly assume a scale in the system and ignore the full range of coordination present in social media. We propose a new measure of coordination, the co-retweet speed, between pairs of accounts that allows instead this characterisation and we show that it can be used to identify centrally coordinated information campaigns without the need of multiple filtering procedures. We test our approach by studying the spread on Twitter of messages supporting Ukraine or Russia at the beginning of the Russian invasion of Ukraine in 2022. We use a retweet network to characterise the flow of information between accounts, where we observe two very polarised communities (pro-Russia and pro-Ukraine). We then compare it with a coordination network built using the co-retweet speed as link weight. The expected wide range of coordination is present in the pro-Ukraine community; however, the pro-Russia community shows instead a peaked behaviour. This is paired with higher coordination levels and less information flow. Even though pro-Russia authors retweet more than pro-Ukraine authors, pro-Ukraine authors are more viral. Taken all together, our results indicate the presence of a pro-Russia centrally coordinated information campaign at the beginning of the invasion of Ukraine.
Join at: imt.lu/seminar