Fact-checking: Is it effective on Facebook? Depends on

Since Facebook is not particularly well known for its transparency, little is known about the effectiveness of its own measures. “Certified” fact-checkers, i.e. those whose comments appear next to a Facebook post that have been found to be false or misleading, regularly request data so that they can gauge the scope of their work. Therefore, two French researchers from the Institute of Political Science in Paris attempted to draw on the “engagement” data of the groups or Facebook pages targeted by this verification, paying particular attention to the groups or pages that regularly shared disinformation in 2019 and 2020.

As a result, reach – the number of comments, likes or “shares” posted by these accounts – decreased significantly after a Facebook page was flagged as having shared at least two news stories. Error by any of the supported media. We are talking about a drop in range of 43% or 62%, depending on the database used by the researchers (one of the two databases is that of the verification site Science notes, one of the authors of the research is its founder).

But surprisingly, this encouraging finding does not apply to Facebook groups (pages are public, while groups can be public or private). Admittedly, as of June 2020, the average engagement rate per message for groups that spread fake news has been halved, but in return these groups doubled their number of posts between 2019 and summer 2020, so the net result is positive for them. We note that COVID is central to this analysis: the more the June 2020 changes can be linked to Facebook’s efforts – like other platforms – to reduce the massive amount of false news about the pandemic, the more the number grows. The posts may have been the work of the fake news producers who discovered the epidemic.

At most, the researchers wrote, “Facebook’s policy appears to be able to contain, rather than reduce, the growth of misinformation shared by ‘repeated criminals’.”