Influence of web platform practives and algorithmic decisions on public access to climate information

The spread of misinformation on online platforms is a growing concern in society, and experts have increasingly highlighted platforms’ responsibility in addressing the issue. This research project aimed at quantifying the impact of platforms’ actions to fight the spread of misinformation. We chose to focus on Facebook, the biggest social media platform with more than two billion users worldwide. Facebook has announced several policies and interventions over the past few years, one of which is the downranking of accounts and websites that repeatedly share misinformation.

In this research project, we developed a methodology to identify such accounts and track the engagement they generate to quantify how much Facebook reduces the visibility of their posts. We find a significant, short-term reduction in engagement per post (by about 50%) following the publication of 2 or more ‘false’ information within a 90 days period by a website or a Facebook Page.

We show that this ‘repeat offenders’ penalty initially did not apply to Facebook groups that repeatedly share misinformation up until the end of 2020. Instead, groups had been affected in a different way with a sudden drop in their average engagement per post around June 9, 2020.

While the penalty did not apply to Facebook groups initially, we detected that Facebook started implementing it on groups in late February 2021. 

When integrated over time, Facebook’s policy against websites, Pages and Groups does not result in a reduction of the total engagement generated by their posts and articles. Overall, the policy appears to contain the increase in engagement on misinformation shared by ‘repeat offenders’ rather than to decrease it.

With the advent of the European Digital Services Act and regulations by individual countries, it becomes critical to develop additional methodologies to audit platforms’ actions to tackle misinformation