Policies put in place by Facebook to tackle covid-19 vaccine misinformation had a meaningful impact on the amount of negative conversation around jabs 鈥� but only temporarily.
David Broniatowski at George Washington University in Washington DC and his colleagues used CrowdTangle, a tool belonging to Facebook鈥檚 owner Meta, to analyse anti-vaccine conversations on Facebook before and after the platform made policy changes. These were introduced between November 2020 and February 2021 and included removing posts that could lead to聽harm, such as false claims that聽covid-19 vaccines contain microchips.
The researchers identified 216 Facebook pages and 99 Facebook groups conducted in English where discussions about vaccines regularly took place. A Facebook page is often run by an individual or organisation to predominately broadcast to users, while a group more often hosts discussions between users. Two-thirds of the聽pages and groups the team identified were opposed to covid-19 vaccines, based on descriptions written by moderators.
Advertisement
Every public post made on those groups and pages in the 12 months prior to November 2020 was extracted, along with the levels of engagement with them. The same pages and groups were revisited in聽August 2021 鈥� after policy changes聽鈥� and all public posts made in the previous six months were extracted and analysed. In total, more than 350,000 posts were gathered from these groups and pages. A further 266,000 posts were analysed from groups and pages that appeared between November 2020 and August 2021.
After revisiting, the team found that 13 per cent of pages and 24 per cent of groups in the original sample disappeared. A further 5聽per cent of groups changed their聽privacy setting to private, preventing further analysis.
In those groups and pages that remained, the volume of posts and interactions decreased, indicating Facebook had some success in tackling vaccine disinformation. On anti-vaccine pages and groups, posts decreased to 29 per cent and 60 per cent of pre-policy levels and engagements dropped to 23 per cent and 66 per cent of pre-intervention levels, respectively. However, the number of posts and engagements on pages promoting vaccines dropped, too, though not by as much. The team says this content may have been caught up in the dragnet.
The effect was short-lived. After six months, new posts on pages returned to 82 per cent and engagements to 80 per cent of their pre-policy level, while in groups, posts were 8 per cent higher than before the policy and engagements were 312 per cent higher. While the team is still analysing why this happened, Broniatowksi believes that users making anti-vaccine posts may simply have found a way around the policy.
A spokesperson for Meta says: 鈥淭here were many other changes happening during this period which could have played a role in these results, which the research doesn鈥檛 take into account.鈥� Meta says this could include the roll-out of covid-19 vaccines, and how the tone of discussions around them changed over time.
Reference:
