Replication data for "Emergent structures of attention on social media are driven by amplification and triad transitivity"

As they evolve, social networks tend to form transitive triads more often than random chance and structural constraints would suggest. However, the mechanisms by which triads in these networks become transitive are largely unexplored. We leverage a unique combination of data and methods to demonstrate a causal link between amplification and triad transitivity in a directed social network. Additionally, we develop the concept of the "attention broker," an extension of the previously theorized tertius iungens (or "third who joins").
<br><br>
We use a novel technique to identify time-bounded Twitter/X following events, and then use difference-in-differences to show that attention brokers cause triad transitivity by amplifying content. Attention brokers intervene in the evolution of any sociotechnical system where individuals can amplify content while referencing its originator.
<br><br>
The full dataset consists of the time-bounded follower counts, and their associated timings, for each retweeted account and attention broker followers and non-followers. All retweeted accounts' usernames and followers' user IDs are hashed using a non-reversible hash function for privacy.