Study: Facebook likes, comments, and shares from news outlets that routinely publish misinformation roughly tripled from Q3 2016 to Q3 2020 to 1.8B interactions
During the 2016 presidential election, Russian operatives used Facebook, Instagram, Twitter and other social media platforms …
Context & Ripple Effects
This finding walks back the optimism of the platform's earlier progress story. A 2018 study had shown Facebook engagement for fake-news sites falling more than half after the 2016 election (engagement dropped 50%+ by mid-2018), and observers credited Facebook with a decent 2018 midterm showing even while warning that false news still spread fast without outside monitoring (experts on the midterm performance). The tripling to 1.8B interactions by Q3 2020 says that decline reversed under election-scale demand.
The corpus also shows why platform-side fixes under-measure the problem: internal documents reveal Facebook's own researchers found users are four times more likely to hit misinformation when it arrives via a share of a share, a path ranking tweaks don't cleanly cover. The follow-on NYU work then found misinfo publishers drawing six times the engagement of sources like WHO into early 2021, suggesting Q3 2020 was not a spike but a baseline.
First-order effects
- Facebook enters the final weeks of the 2020 election with its 2016-era disinformation problem measurably back at roughly triple scale, undercutting the midterms narrative its defenders relied on.
- News outlets flagged as routine misinformation publishers get an engagement windfall — 1.8B quarterly interactions — that rewards the exact publishing behavior platforms claim to demote.
Second-order effects
- Independent researchers become the de facto auditors: the gap between the 2018 drop study and this tripling forces reliance on outside measurement like NYU's, since Facebook's own metrics missed the reversal.
- Resharing chains emerge as the pressure point — with internal data confirming share-of-a-share exposure multiplies misinformation reach, remediation shifts from publisher-level demotion to distribution-path interventions Facebook has so far left open.
Third-order effects
- If every claimed fix is followed by an engagement rebound, the durable structure is permanent external oversight of platform content systems rather than trust in self-reported progress — a governance burden that follows whichever interface carries the feed next.
The trend: Platform self-assessments of misinformation progress keep being contradicted by independent engagement studies, pushing accountability toward continuous outside auditing of social media distribution.