Starbird, K., Arif, A., & Wilson, T. (2019). Disinformation as Collaborative Work: Surfacing the Participatory Nature of Strategic Information Operations. Proc. ACM Hum.-Comput. Interact., 3, 1–26.
Summary
Starbird, Arif, and Wilson reframe strategic information operations—disinformation, propaganda, and online manipulation—as a form of collaborative work that unfolds within and through online crowds rather than something merely done to a passive audience. Drawing on Computer-Supported Cooperative Work (CSCW) and a sociotechnical lens attentive to the mutual shaping of technology, social structure, and human agency, the authors argue against reductive “bots and trolls” accounts. Through three long-term case studies, they develop a typology of operations spanning orchestrated, cultivated, and emergent forms, and highlight the persistent difficulty of distinguishing coordinated campaigns from organic crowd behavior—a difficulty they argue undermines platform policies built on “coordinated inauthentic behavior.”
Key Contributions
- Reframes information operations as collaborative “work,” integrating organizational (big-W) and distributed crowd (little-w) work through a sociotechnical lens.
- Offers a typology distinguishing orchestrated, cultivated, and emergent/self-sustaining operations, each grounded in a case study.
- Redefines disinformation as a campaign—a persistent collection of information-actions built around a plausible “rational core”—rather than a property of any single piece of content.
- Critiques “coordinated inauthentic behavior” policy framings as inadequate for authentic actors and self-sustaining narratives; argues for campaign-level, provenance-based, substantive policy.
- Situates the phenomenon historically (Soviet active measures, dezinformatsiya, “unwitting agents”) and demonstrates CSCW’s relevance to disinformation scholarship and platform policy.
Methods
A grounded, interpretivist, mixed-method approach combining qualitative, quantitative, and visual analysis of digital trace data, adapting grounded theory alongside trace and network ethnography. The work is grounded in crisis-informatics techniques for rapid collection and analysis of social media data, iterating between macro views (network graphs, descriptive statistics) and close reading of tweets, accounts, and articles. The authors emphasize researcher reflexivity—multiple investigators from diverse backgrounds, journaling, and reporting how their values shaped interpretation—across three case studies each spanning a year or more.
Findings
- Case 1 (Russian IRA): Over 3,000 coordinated persona accounts; ~300 impersonating African American activists and white conservatives produced about a third of English tweets but drew 85% of English retweets (18.5M). The operation was top-down (from St. Petersburg) yet strategically embedded into pre-existing organic movements, including domestic conservative networks like PJNET.
- Case 2 (White Helmets): Critical accounts outnumbered supporters (~40% more accounts, ~3× the content) and were largely authentic “information activists” and Western “journalists,” amplified by state media (RT, Sputnik). A tightly connected “echo-system” of sites copy-pasted identical content across varied ideological wrappers—best described as cultivated rather than fully orchestrated.
- Case 3 (crisis conspiracy theorizing): “Crisis actor” and “false flag” rumors emerged organically from communities of “true believers” sharing an anti-institutional epistemology, supported by an alternative media ecosystem (RT, Sputnik, PressTV)—largely emergent and self-sustaining, partly shaped by long-standing operations.
- Significant overlap in supporting media domains across Cases 2 and 3 reveals a shared disinformation infrastructure.
- Organic public participation can grow in relative importance as ordinary users adopt disinformation-aligned narratives as expressions of identity.
Connections
This is a foundational CSCW framing that the authors’ later work extends, notably Starbird2025-jj. Its emphasis on distinguishing orchestrated from organic behavior connects directly to the coordinated-inauthentic-behavior detection literature, including the Giglietto group’s work on coordinated link sharing (Giglietto2020-9d8acdd7, Giglietto2022-0e951ac5, Giglietto2017-4375de2f) and computational detection of coordination such as Nizzoli2020-cf and Keller2019-nk. Its participatory, crowd-centered account of disinformation as identity-expressive work resonates with structural analyses of partisan media ecosystems (Benkler2018-lw, Freelon2020-yp) and with broader debates on the science of misinformation (Lazer2018-mm, Vosoughi2018-at).