“Fake Media Facebook will tell you this article is disputed.” One of the first times we saw a Fakebook fake news alert in the wild was around St. Patrick’s Day, for a debunked story on the “Irish slave trade.” (Better known as the Irish slaves myth.) The Guardian’s Sam Levin reported this week that Facebook’s “throttling” had the opposite effect: “A bunch of conservative groups grabbed this and said, ‘Hey, they are trying to silence this blog — share, share share,'” claimed Christian Winthrop, the editor of Newport Buzz, the site that published the article.

It’s unclear how often this is happening. Facebook’s recognition of the Irish slave trade article, in particular, as fake was written up by a number of news outlets, so it was probably easier to identify it as something to run a re-promotion campaign around. Some editors and writers at fake news sites told Levin that it wasn’t clear if Facebook’s campaign was having an effect on their traffic. One said the site had “definitely seen a drop in traffic since Facebook started relying on outside fact-checkers.”

Individual anecdotes really aren’t the best way to gauge whether this is working, though. This full tweetstorm from MIT Media Lab’s J. Nathan Matias (whose team, earlier this year, collaborated with Reddit to encourage users to fact-check sensational or potentially misleading stories) is an interesting read; a few highlights:

It's very unlikely that the fact checking organizations would develop clear intuitions on the effects across so many people

— J. Nathan Matias (@natematias) May 18, 2017

In the meantime, we should avoid judging the effectiveness or failure of fact checking based on single examples

— J. Nathan Matias (@natematias) May 18, 2017

During our study w/ worldnews on reddit, if I had pulled a random link to unreliable news, I could easily have concluded it wasn't working.

— J. Nathan Matias (@natematias) May 18, 2017

Effects can be surprising. Hypothesis: by promoting fact-checked articles further, unreliable news sources are undermining their credibility

— J. Nathan Matias (@natematias) May 18, 2017

In all, it’s early and there’s not enough evidence to declare the effort a success or failure. Facebook, of course, isn’t releasing data.

“A very successful troll plays with ambiguity.” “Media Manipulation and Disinformation Online,” a big report by the nonprofit research institute Data & Society’s Alice Marwick and Rebecca Lewis, is out this week. The dense report — with a shorter and more “fun” summary version at New York magazine’s Select All — looks at how “various internet subcultures — sometimes summarized as the ‘alt-right,’ but more accurately an amalgam of conspiracy theorists, techno-libertarians, white nationalists, Men’s Rights advocates, trolls, anti-feminists, anti-immigration activists, and bored young people — leverage both the techniques of participatory culture and the affordances of social media to spread their various beliefs.”

In addition to 100-percent fake sites, the report calls out sites that “spread information that falls on a continuum between true and false”:

Publications with highly ideological agendas, such as Breitbart News or Occupy Democrats, often deliberately manipulate information to fit into a specific worldview. For instance, in 2014 the Washington Free Beacon published an article claiming that the U.S. government was funding a research effort to track and surveil conservative statements on social media, which was picked up by Breitbart and eventually Fox News. While the project did exist, its findings were mischaracterized. Similarly, liberal activists publicized a story from the Conservative Daily Post claiming that the Trump administration would charge political protesters as terrorists. This was based on an informal proposal by Senator Doug Erickson to charge protesters who blocked businesses with economic terrorism. In both cases, the news stories were a combination of facts and misinformation.

“I do not expect to hit 100 percent in terms of accuracy.” Another Guardian article this week, by Jon Swaine, further explores the mixing of true and false on the parts of bloggers like the anti-Trump Louise Mensch and pro-Trump Mike Cernovich, “whose track records of faulty reporting are occasionally interrupted by stories that are actually true.” Mensch told Swaine: “I believe in a free press. People both can and should write whatever they want.” Claude Taylor: “Speaking only for myself, I do not expect to hit 100 percent in terms of accuracy but I do hope, when all is said and done, to be over 80 percent — roughly stated.”

BuzzFeed’s Charlie Warzel wrote about a related question this week: “What Happens When The Pro-Trump Media Get Actual Scoops?” Warzel writes that real news stories broken by Cernovich and others have “complicated the once black-and-white characterization of the pro-Trump media as purveyors of fake news. In recent weeks especially, the pro-Trump media has frequently seized control of the political news cycle via an unexpected tactic: real and, at times, well-sourced reporting.”