Lazer, D. M. J., Baum, M. A., Benkler, Y., Berinsky, A. J., Greenhill, K. M., Menczer, F., Metzger, M. J., Nyhan, B., Pennycook, G., Rothschild, D., Schudson, M., Sloman, S. A., Sunstein, C. R., Thorson, E. A., Watts, D. J., & Zittrain, J. L. (2018). The science of fake news. Science, 359, 1094–1096. https://doi.org/10.1126/science.aao2998
Summary
This influential Policy Forum article stakes out a foundational, agenda-setting position for the emerging field of fake news research. The authors define fake news by the intent and process of its publishers—fabricated content that mimics the form of news media without its editorial norms—rather than by the accuracy of any single story. They argue that politically oriented fake news thrives because the internet has eroded the institutional safeguards (journalistic gatekeeping, sustainable business models, public trust) that once contained misinformation, while rising geographic, attitudinal, and affective polarization has produced homogeneous networks receptive to ideologically compatible falsehoods. Crucially, they emphasize how little rigorous evidence exists about fake news’ prevalence and long-term effects, and issue a normative call for interdisciplinary research and platform–academic collaboration to build a genuine “science of fake news.”
Key Contributions
- A scientific definition of fake news grounded in publisher intent and process, distinguishing it from broader misinformation and disinformation.
- A historical framing that situates fake news within the rise and decline of journalistic norms and within trends of political polarization.
- A two-part taxonomy of interventions—empowering individuals (fact checking, media literacy) versus structural/platform changes—with critical assessment of each.
- An agenda-setting call for interdisciplinary research and for platforms to share unique data with independent academics.
Methods
Interdisciplinary synthesis drawing on communication studies, political science, psychology, and computer science. The authors review empirical literature (exposure estimates before the 2016 U.S. election, diffusion of false information on Twitter, bot-prevalence studies), develop conceptual definitions and a taxonomy of interventions, and assess platform practices and legal/regulatory frameworks (e.g., the Communications Decency Act, the European “right to be forgotten”).
Findings
- The average American encountered roughly one to three known fake-news stories in the month before the 2016 election—likely a conservative estimate.
- False information on Twitter spreads faster and reaches more people than true information, especially on political topics.
- Bots are prevalent (9–15% of active Twitter accounts; Facebook estimated up to 60 million) and amplified political content in the 2016 U.S. and 2017 French campaigns.
- Trust in mass media hit historic lows in 2016, with a stark partisan gap (51% of Democrats vs. 14% of Republicans).
- Scientific support for fact checking is at best mixed; cognitive tendencies (selective exposure, confirmation bias, familiarity/fluency effects) can render corrections ineffective or counterproductive.
- No comprehensive data-collection system exists to track the evolving fake-news ecosystem.
Connections
This piece is a canonical starting point that many subsequent studies build on or empirically test. Its claims about faster spread of falsehoods draw directly on Vosoughi2018-at, and its skepticism about the actual prevalence and impact of fake news anticipates later measurement work such as Grinberg2019-ua, Allcott2017-yz, Guess2019-ym, Guess2020-rr, and Guess2021-ym. Its treatment of correction and cognitive resistance connects to Lewandowsky2012-vn and to accuracy-focused interventions in Pennycook2021-jq, while its institutional and polarization framing dovetails with Benkler2018-lw, Del-Vicario2016-uj, and Flaxman2016-lm.