Tsfati, Y., Boomgaarden, H. G., Strömbäck, J., Vliegenthart, R., Damstra, A., & Lindgren, E. (2020). Causes and consequences of mainstream media dissemination of fake news: literature review and synthesis. Annals of the International Communication Association, 44, 157–173. https://doi.org/10.1080/23808985.2020.1759443
Summary
This article synthesizes literature across political communication, journalism studies, and cognitive psychology to advance a counterintuitive argument: mainstream news media are a major disseminator of fake news, likely reaching more people than fake news websites or social media do directly — even though their coverage is usually intended to correct and debunk. The authors resolve an apparent puzzle: direct exposure to fake news is limited to a small, ideologically extreme slice of the public, yet large majorities know about and often believe prominent false stories. Their explanation is that most people encounter fake news secondhand, through mainstream coverage. The paper works through three questions — what role mainstream media play in disseminating fake news, why they cover it, and how audiences process that coverage.
Key Contributions
- Redirects attention to the underexamined, paradoxical role of mainstream media in spreading (not merely correcting) disinformation.
- Offers an integrated framework for why journalists cover fake news: truth-seeking role perceptions, high news value, the social psychology of news decisions, and existing content-monitoring infrastructure.
- Bridges disinformation research with the cognitive/social psychology of misinformation correction to theorize the audience effects of coverage.
- Sets a research agenda: content analysis of fake news coverage, newsroom ethnographies, and experimental/linkage studies on audience processing (including partisan audiences, media trust, and attitude certainty).
Methods
Narrative literature review and synthesis rather than original data collection. The authors integrate survey evidence (e.g., YouGov/Frankovic on Pizzagate, BuzzFeed studies), traffic and network analyses of exposure and sharing, and content analyses. They use illustrative LexisNexis searches across 12 U.S. outlets for terms like “fake news” and “Pizzagate,” draw on Allcott and Gentzkow’s dataset of 156 election-related fake articles to show how such stories satisfy news values, and apply meta-analytic findings on misinformation correction to theorize audience effects.
Findings
- Public belief in fake news is high: over a third of Americans thought Pizzagate was probably/definitely true; 75% of recalled fake headlines were rated somewhat/very accurate; similar patterns appear in Brazil and Israel.
- Direct exposure is highly concentrated: 80% of fake news on Twitter reached 1% of users; ~60% of fake news site visits came from the 10% most conservative online diets; real news sites had ~40x more monthly visitors than fake news sites.
- Sharing is rare and skewed (0.1% of Twitter users produced 80% of sharing), and fact-checks rarely reach fake news consumers due to selective diffusion.
- Fake news is highly newsworthy — negativity, elite actors, surprise, entertainment, celebrity — and appears designed to maximize spread and align with national agendas.
- Partisan media are more heavily influenced by fake news than mainstream media, partly due to ideological fit.
- Correction is undermined by repetition (mere-exposure and truth effects), lexical complexity, congruence with prior attitudes, and low media trust; the modal audience response to reported fake news is uncertainty rather than rejection.
Connections
This review is foundational for the recalibration of how we think about misinformation exposure, complementing empirical work showing that direct fake-news consumption is concentrated and limited — themes central to Guess2020-rr, Grinberg2019-ua, Guess2019-ym, and Allen2020-nj. Its focus on the amplification role of mainstream coverage speaks to broader debates over the “disinformation order” and networked propaganda in Benkler2018-lw and Bennett2018-ys, while its treatment of repetition, the truth effect, and correction resistance connects directly to Pennycook2021-jq, Lewandowsky2012-vn, Nyhan2023-gb, and Vosoughi2018-at.