Facebook labelling of state-controlled media working, says study

In 2020, Facebook started to label possible propaganda from state-affiliated outlets (Brett Jordan / Unsplash)
In 2020, Facebook started to label possible propaganda from state-affiliated outlets (Brett Jordan / Unsplash)

Back in 2020, Facebook quietly began to label potential propaganda from sites with links to governments.

Now a joint study from researchers at Carnegie Mellon University, Indiana University and the University of Texas has concluded that the labels are indeed reducing the engagement of such links on the platform.

The first experiment looked at how 1,200 American Facebook accounts interacted with posts with and without state-affiliated tags. The result was that engagement on labeled posts from China and Russia decreased: users were less likely to like, read, share, comment or — crucially — believe the content if they saw the label.

A second study wanted to look at how the state involved impacts user behaviour. Nearly 2,000 US Facebook accounts were shown posts, some with and some without labelling. It found that public sentiment corresponded to the country involved.

“For example, they responded positively toward content labelled as coming from Canadian state-controlled media and negatively toward content labelled as coming from Chinese and Russian state-controlled media,” Carnegie Mellon University explains.

The final study used field data to examine user engagement before and after the June 4, 2020 introduction of state-affiliated labels. It found that shares and likes on posts from Chinese and Russian state-controlled pages were down 34 per cent and 46 per cent respectively after the labels appeared — though this could also be partially down to other changes to the newsfeed algorithm from the time.

“Propaganda is a major concern on social media, but it has not received the same attention that mis- and disinformation have received, and it can be more insidious and even less obvious,” said Professor Avinash Collis, who co-authored the research.

“By understanding the impact of labelling propaganda, social media companies, news media companies, and users will be able to implement and respond to the labels more appropriately.”

Given some users missed the labels altogether, that’s one possible area of improvement. Indeed, the change could have been more impactful if Facebook had informed users of it, according to PhD student Nicholas Wolczynski, who co-authored the study. “Given that Facebook debuted the new labels quietly without informing users, many likely did not notice the labels, reducing their efficacy dramatically,” he said.

What is and isn’t “state-affiliated media” can be a touchy subject for publications, and there’s an irony in the label having possible positive connotations for publications from trusted countries, given a recent blowup on another social network.

Back in April, NPR, PBS and the Canadian Broadcasting Corporation announced they were leaving Twitter after the platform conflated “publically funded” with “state-affiliated” in its labelling.

The two sound similar, but have a key difference, with only the former guaranteeing editorial independence and no government involvement in what is and isn’t published.

The resulting scuffle reportedly led to owner Elon Musk threatening to give away NPR’s existing Twitter account if it didn’t return. That was three months ago, and at the time of writing, the account remains in NPR’s hands. It still remains dormant.