Levin, S. H., Wang, R., Seo, J., Leong, Y. Y., Bennett, E., Yang, D., Yang, Z., Chen, M., Zhang, Y., & Suk, J. (2026). From #StayWoke to “culture wars”: How social justice discourse is separately and synergistically politicized on twitter and YouTube. Political Communication, 1–25. https://doi.org/10.1080/10584609.2026.2717428

View paper

Summary

This paper traces the decade-long politicization of the term “woke” — from its origins in Black communities as “stay woke,” a call to anti-racist consciousness, into an anti-woke “culture war” epithet weaponized by partisan actors. Drawing on identity politics and polarization theory, the authors combine large-scale user- and content-level computational analysis of Twitter and YouTube data (2012–2022) to show how partisan opinion leaders and their followers co-opted social justice discourse. They argue this hijacking happened both separately — shaped by each platform’s distinct affordances and user bases — and synergistically, through coordinated cross-platform amplification. The central finding is that culture-war framing came to dominate “woke” discourse across every user group, including liberals and Black opinion leaders, with conservatives building an asymmetric cross-platform ecosystem to promote right-wing content.

Key Contributions

  • Documents the longitudinal politicization of a social justice term (from #StayWoke to anti-woke culture wars) via large-scale computational analysis spanning a decade.
  • Advances comparative/cross-platform social media research by distinguishing the separate and combined roles of Twitter and YouTube in politicizing social identity.
  • Identifies an asymmetric right-wing cross-platform information ecosystem in which YouTube “supplies” content and Twitter “demands”/amplifies it.
  • Provides empirical support for theories of identity-politics-driven polarization and the co-optation of marginalized communities’ discourse.

Methods

The authors collected 18.4M tweets (5.97M users) via the Twitter API v2 and 35,445 videos (17,398 channels) via the tuber R package, using “woke”-related search terms, segmented into three periods (2012–2015, 2016–May 2020, May 2020–2022). User-level analysis used Vintage Sparse Principal (VSP) component analysis to partition the Twitter retweet network into 74 clusters and identified the top YouTube channels via a long-tail distribution, followed by manual coding of opinion-leader groups. Content-level analysis applied BERTopic (validated with LDA) to video titles and a 10% tweet sample to derive topic categories, then predicted topics across the full corpus. Embedded URLs were extracted and unshortened to trace cross-platform link-sharing, with Z-tests and Chi-square tests (standardized residuals) assessing shifts across platforms and time.

Findings

  • Political opinion leaders were more prevalent on YouTube overall (conservative news/commentary drew 65.41% of views), but liberals dominated Twitter while conservatives dominated YouTube.
  • Politicized topics (Culture Wars + Politics) made up 67.72% of tweets but only 23.46% of YouTube videos, where lifestyle/entertainment dominated (56.99%).
  • Discourse shifted toward Culture Wars and away from Stay Woke across all Twitter groups and phases — even Black opinion leaders’ culture-war engagement reached 69.22% by T3.
  • On YouTube, Culture Wars grew across nearly all channel categories, and Politics rose sharply among conservative channels (0% at T1 to 16.95% at T3).
  • Conservatives on Twitter shared YouTube content more than liberals, often to redefine or ridicule “woke”; right-wing outlets (Fox News, Breitbart, NYPost, Daily Mail) appeared prominently in Twitter’s top domains, indicating coordinated promotion.
  • The Black community’s declining use of “stay woke” suggests the term became so politicized that its originators grew reluctant to use it.

Connections

This paper’s account of coordinated cross-platform amplification of partisan content connects to work on link-sharing coordination and its detection, such as Giglietto2017-4375de2f and Giglietto2019-882f1900. Its focus on the co-optation and reframing of discourse by partisan actors resonates with research on conspiratorial and identity-driven media narratives like Marwick2026-qd and Grusauskaite2026-po, while its asymmetric right-wing ecosystem finding speaks to broader studies of platform-level political asymmetry such as Falkenberg2026-ka.

Podcast

A research-radio episode discusses this paper: 🎧 MP3 · Spotify