Study: Facebook engagement for sites identified as publishing fake news dropped by 50%+ between 2016 election and July 2018, while shares on Twitter kept rising
Facebook's efforts to reduce misinformation in its news feed since the 2016 election have opened the company to all manner of criticism …
Context & Ripple Effects
This study lands two years into Facebook's post-election cleanup arc. Back in November 2016, reporting said Facebook had shelved an update that would have flagged fake news stories, worried it would disproportionately hit right-wing sites; by April 2017 the company was publicly describing improved detection of coordinated amplification via fake accounts. The new data is the first hard readout of whether any of that moved the numbers.
It did — on one platform. Engagement for identified fake-news publishers fell by more than half on Facebook between the 2016 election and July 2018, while their shares on Twitter kept climbing, making the divergence itself the story: suppression is a per-platform product decision, not a property of the content.
First-order effects
- Fake-news publishers lose their largest distribution channel overnight — Facebook's ranking and detection changes translate directly into a 50%+ engagement cut for sites the study identifies, with no equivalent loss on Twitter.
- Twitter now stands out as the unmoderated alternative, giving the same publishers a growing audience precisely where Facebook clamped down.
Second-order effects
- Advertisers and regulators gain a clean A/B comparison between two platforms' moderation approaches, raising pressure on Twitter to explain why its trajectory runs opposite to Facebook's.
- Facebook's apparent success sets a benchmark rivals get measured against — but later research shows the win was fragile: interactions for routine misinformation outlets roughly tripled from Q3 2016 to Q3 2020, reaching 1.8B.
Third-order effects
- If the pattern holds, platform-level suppression proves reversible and demand-driven — the 2020 rebound and an NYU finding that misinfo publishers drew six times the engagement of sources like WHO from August 2020 to January 2021 suggest ranking fixes decay without sustained enforcement.
- The durable shift is toward treating misinformation reach as a measurable, auditable metric per platform, with researchers — not the platforms — holding the comparative scorecard across Facebook and Twitter.
The trend: Social platforms are diverging into individually accountable misinformation environments, where each company's enforcement choices — not the content itself — determine how far fake news travels.