Internal Research from Facebook Reveals that Re-Shares Can Significantly Amplify Misinformation

What if Facebook removed post shares entirely, as a means to limit the spread of misinformation in its apps? What impact would that have on Facebook engagement and interaction?

The question comes following the release of new insights from Facebook’s internal research, released as part of the broader ‘Facebook Files’ leak, which shows that Facebook’s own reporting found that post shares play a key role in amplifying misinformation, and spreading harm among the Facebook community.

As reported by Alex Kantrowitz in his newsletter Big Technology:

“The report noted that people are four times more likely to see misinformation when they encounter a post via shares – kind of like a retweet of a retweet – compared to a typical photo or link on Facebook. Add a few more shares to the chain, and people are five to ten times more likely to see misinformation. It gets worse in certain countries. In India, people who encounter “deep reshares,” as the researchers call them, are twenty times more likely to see misinformation.”

So it’s not direct shares, as such, but re-amplified shares, which are more likely to be the kinds of controversial, divisive, shocking or surprising reports that gain viral traction in the app. Content that generates emotional response sees more shares in this respect, so it makes sense that the more radical the claim, the more re-shares it’ll likely see, particularly as users look to either refute or reiterate their personal stance on issues via third party reports.

And there’s more:

“The study found that 38% of all [views] of link posts with misinformation take place after two reshares. For photos, the numbers increase – 65% of views of photo misinformation take place after two reshares. Facebook Pages, meanwhile, don’t rely on deep reshares for distribution. About 20% of page content is viewed at a reshare depth of two or higher.”

So again, the data shows that those more spicy, controversial claims and posts see significant viral traction through continued shares, as users amplify and re-amplify these posts throughout Facebook’s network, often without adding their own thoughts or opinions on such.

So what if Facebook eliminated shares entirely, and forced people to either create their own posts to share content, or to comment on the original post, which would slow the rapid amplification of such by simply tapping a button?

Interestingly, Facebook has made changes on this front, potentially linked to this research. Last year, Facebook-owned (now Meta-owned) WhatsApp implemented new limits on message forwarding to stop the spread of misinformation through message chains, with sharing restricted to 5x per message.

This, WhatsApp says, has been effective:

“Since putting into place the new limit, globally, there has been a 70% reduction in the number of highly forwarded messages sent on WhatsApp. This change is helping keep WhatsApp a place for personal and private conversations.”

This is a positive outcome, and shows that there is likely value to such limits. But the newly revealed research looked at Facebook specifically, and thus far, Facebook hasn’t done anything to change the viral shares process within its main app, the core focus of concern in this report.

The company’s lack of action on this front now forms part of Facebook whistleblower Frances Haugen’s legal push against the company, with Haugen’s lawyer calling for Facebook to be removed from the App Store if it fails to implement limits on re-shares.

Facebook hasn’t responded to these new claims as yet, but it is interesting to note this research in the context of other Facebook experiments, which seemingly both support and contradict the core focus of the claims.

In August 2018, Facebook actually did experiment with removing the Shares button from posts, replacing it with a ‘Message’ prompt instead.

That seemed to be inspired by the increased discussion of content within messaging streams, as opposed to in the Facebook app – but given the timing of the experiment, in relation to the study, it seems now that Facebook was looking to see what impact the removal of the shares button could have on in-app engagement.

On another front, however, Facebook’s actually tested expanded sharing, with a new option spotted in testing that enables users to share a post into multiple Facebook groups at once.

Facebook shares to groups prompt

That’s seemingly focused on direct post sharing, as opposed to re-shares, which were the focus of its 2019 study. But even so, providing more ways to amplify content, potentially dangerous or harmful posts, more easily, seems to run counter to the findings outlined in the report.

Again, we don’t have full oversight, because Facebook hasn’t commented on the reports, but it does seem like there could be a benefit to removing post shares entirely as an option, as a means to limit the rapid re-circulation of harmful claims.

But then again, maybe that just hurts Facebook engagement too much – maybe, through these various experiments, Facebook found that people engaged less, and spent less time in the app, which is why it abandoned the idea.

This is the core question that Haugen raises in her criticism of the platform, that Facebook, at least perceptually, is hesitant to take action on elements that potentially cause harm if that also means that it could hurt its business interests.

This, at Facebook’s scale and influence, is an important consideration, and one which we need more transparency on.

Facebook claims that it conducts such research with the distinct intent of improving its systems, as CEO Mark Zuckerberg explains:

“If we wanted to ignore research, why would we create an industry-leading research program to understand these important issues in the first place? If we didn’t care about fighting harmful content, then why would we employ so many more people dedicated to this than any other company in our space – even ones larger than us? If we wanted to hide our results, why would we have established an industry-leading standard for transparency and reporting on what we’re doing?”

This makes sense, but that doesn’t then explain whether business considerations factor into any subsequent decisions as a result, when a level of potential harm is detected by its examinations.

That’s the crux of the issue. Facebook’s influence is clear, its significance as a connection and information distribution channel is evident. But what plays into its decisions in regards to what to take action on, and what to leave, as it assesses such concerns?

There’s evidence to suggest that Facebook has avoided pushing too hard on such, even when its own data highlights problems, as seemingly shown in this case. And while Facebook should have a right to reply, and its day in court to respond to Haugen’s accusations, this is what we really need answers on, particularly as the company looks to make even more immersive, more all-encompassing connection tools for the future.