NYU study: from August 2020 to January 2021, publishers known for putting out misinfo got six times more engagement than trustworthy news sources like WHO
Right-leaning pages also produce more misinformation, the forthcoming study found. — A new study of user behavior on Facebook around … Tweets: @kevinroose , @lizzadwoskin , @justinhendrix , @froomkin , and @caseynewton Tweets: Kevin Roose / @kevinroose : @lizzadwoskin Sorry to be That Guy, but aren't engagements and “clicks” different metrics? (They're different from reach/impressions, as you note in the piece, but also from each other!) Elizabeth Dwoskin / @lizzadwoskin : NEW: Facebook blocked NYU researchers. They just published a study finding that misinfo during the election got 6x more clicks than factual news on FB. It's one of the first comprehensive studies that examines how algos may boost misinfo. https://www.washingtonpost.com/ ... Justin Hendrix / @justinhendrix : Peer-reviewed study by researchers at NYU and Université Grenoble Alpes finds that that from August 2020 to January 2021, news publishers known for putting out misinfo got six times the amount of likes, shares, and interactions on Facebook- @lizzadwoskin: https://www.washingtonpost.com/ ... Dan Froomkin / @froomkin : This strikes me as a gross exaggeration of a potentially flawed study. So people who visited Breitbart's Facebook page were more likely to click on an item than people who visited the World Health Organization's page? This tells us what? https://www.washingtonpost.com/ ... Casey Newton / @caseynewton : “From August 2020 to January 2021, news publishers known for putting out misinformation got six times the amount of likes, shares, and interactions on the platform as did trustworthy news sources, such as CNN or the World Health Organization.” https://www.washingtonpost.com/ ...
Context & Ripple Effects
The NYU findings land amid a dispute over how Facebook’s engagement data should be interpreted: the company blocked the researchers’ data access, while its News Feed team had previously challenged CrowdTangle-based rankings of its highest-engagement stories as inaccurate.
The newer study also complicates a simple progress narrative. A prior analysis found Facebook engagement with identified fake-news sites had fallen sharply through mid-2018, but NYU’s later election-period sample shows a pronounced engagement advantage for publishers known for misinformation.
First-order effects
- Publishers known for misinformation received a far stronger engagement signal on Facebook than trustworthy sources including the World Health Organization, while the study identifies right-leaning pages as producing more misinformation.
- Facebook’s decision to block NYU researchers restricts independent examination of the distribution patterns the study reports.
Second-order effects
- The World Health Organization and other trustworthy news providers compete on Facebook against publishers whose content, according to the study, generates substantially more engagement.
- Facebook cannot rely on earlier declines in fake-news-site engagement as evidence that its current distribution incentives are resolved, given the later NYU findings.
Third-order effects
- If the engagement pattern persists, Facebook’s misinformation problem becomes less about individual false posts than about whether engagement-led distribution systematically rewards publishers associated with misinformation.
- Independent researcher access becomes a core governance issue: without it, Facebook’s claims about feed outcomes are harder to assess against external measurement.
The trend: Platform misinformation governance is shifting from content takedowns toward scrutiny of the engagement incentives and research access that reveal distribution outcomes.