Emergent structures of attention on social media are driven by amplification and triad transitivity

Summary

This paper proposes a new brokerage concept — the tertius amplificans, or “attention broker” — to capture how social media users with large audiences close triads at scale by resharing others’ content with attribution. Extending Obstfeld’s tertius iungens to platforms with many-to-many amplification affordances, the authors argue that retweets by high-degree nodes are a local causal mechanism producing the well-known macro-level tendency toward triad transitivity in directed networks. Using a novel exploitation of the Twitter/X V1 API cursor to obtain time-stamped follower events, they run a difference-in-differences event study on two ideologically divergent cases (Jorts the Cat and J.K. Rowling) and show that amplification causally accelerates follower-to-amplified-account tie formation.

Key Contributions

• Theoretical: Introduces attention brokerage / tertius amplificans as an amplification-based extension of tertius iungens, suited to sociotechnical systems with attributed resharing.
• Methodological: Documents a Twitter/X V1 API cursor technique (modified Unix nanosecond timestamps) that recovers time-bounded follower events, enabling precise temporal analysis of tie formation.
• Empirical: Provides causal (DiD) evidence that retweets by influential accounts generate transitive triads, across two structurally and ideologically divergent cases.
• Conceptual bridge: Links a local micro-mechanism to a macro-level network regularity (transitivity), addressing brokerage literatures that emphasize structural position without identifying causal mechanisms of closure.
• Open resources: Releases an anonymized dataset (SOMAR/ICPSR) and code documenting the cursor-based collection method.

Methods

• Two-case comparative design: Jorts the Cat (~200K followers, pro-union, Dec 2021–Mar 2022) vs. J.K. Rowling (~14M followers, TERF advocacy, Jun 2018–Mar 2023).
• Broker timelines collected via the focalevents package, restricted to simple retweets (excluding quote tweets to avoid “dunking”).
• Hand-coding of 646 (Jorts) and 534 (Rowling) retweeted accounts along cause-alignment and interest-actor dimensions, with adjudication by a third coder.
• Treatment motif: transitive triad (follower–broker–retweeted); control motif: open triad (nonfollower–broker–retweeted), measured in 2-week pre/post windows.
• Attentive population sizes estimated via the POPAN Jolly-Seber mark-recapture model in Project MARK.
• Two-stage DiD event study (Gardner) with account and time fixed effects, plus Rambachan–Roth sensitivity analysis for parallel-trends violations.

Findings

• Day-0 treatment effects are positive and significant for both brokers: followers form ties to amplified accounts at substantially higher rates than nonfollowers.
• Heterogeneity by account type: Jorts’s effect is strongest for union-related accounts; Rowling’s effect is significant across types but largest for TERF interest actors.
• Small positive pre-retweet effects suggest incidental prior exposure also contributes; a post-spike decline indicates amplification accelerates ties that would have eventually formed, depleting latent followers.
• Rambachan–Roth sensitivity tests show robustness: parallel-trends violations would need to be more than 4× larger post-retweet than pre-retweet to overturn key results.
• Attentive populations differ markedly across groups (e.g., ~164K Jorts followers vs. ~17.9M nonfollowers; ~841K Rowling followers vs. ~2.68M nonfollowers in the sampled set).

Connections

This paper sits in a growing computational social science literature on amplification dynamics and influencer effects on platforms; it complements work on virality cascades and superspreader-style accounts such as Bak-Coleman2026-mk and Luceri2025-tr, and shares thematic ground with Green2025-ap on elite signaling on Twitter/X. Its methodological focus on platform affordances and API-derived behavioral traces also connects it to broader debates about data access and platform research infrastructure represented by Freelon2024-sc and Murtfeldt2025-wu. Substantively, by tying micro-level brokerage to macro network structure, it offers a causal complement to descriptive accounts of follower-network formation in works like Lai2024-to.