Engagement with “fake news” on Facebook is declining

In recent years, there has been widespread concern that misinformation on social media is damaging societies and democratic institutions. In response, social media platforms have announced actions to limit the spread of false content. We measure trends in the diffusion of content from 569 fake news websites and 9,540 fake news stories on Facebook and Twitter between January 2015 and July 2018. User interactions with false content rose steadily on both Facebook and Twitter through the end of 2016. Since then, however, interactions with false content have fallen sharply on Facebook while continuing to rise on Twitter, with the ratio of Facebook engagements to Twitter shares decreasing by 60 percent. In comparison, interactions with other news, business, or culture sites have followed similar trends on both platforms. Our results suggest that the relative magnitude of the misinformation problem on Facebook has declined since its peak.

That is from a new NBER working paper by Allcott, Gentzkow, and Yu.

Comments

When might the relative magnitude of the misinformation problem on Twitter decline from its peak? --or should we be enamored that Twitter's universe of meonic content thrives at the expense of Facegook's?

I think anyone once taken in, on any platform, will be more cautious for a time. But that doesn't mean that there are not also long cycle false narratives.

With all of our accumulated overlapping and concurrent false narratives, everyone's beloved Internet has become in scant decades our global Potemkin village, concealing and prettifying as many of the world's horrid realities day-by-day and quarter-hour by quarter-hour as possible (along with the further contributions being perpetrated by our lying and spying Tech Establishment).

The Internet was born of DARPA, a military-industrial project just waiting to astound us with all the flattering sludge it can shove through: what might the value of Dwight Eisenhower's farewell warning to us be today?

Thank from Jayna Viar

Does the Jussie Smollett Lynching count as Real News?

It counts as real news for those that believe (and need to believe...) what he said happened happened and fake news for the rest of us with more than 2 brain cells to rub together.

"I realize that as a person on Twitter this seems hypocritical but you don’t have to speak out loud everything that comes into your brain"

https://twitter.com/AdamSerwer/status/1092449589905575937?s=19

MSM still taking the story on its face, but Twitter is leaning toward "fake news".

Maybe a correlation with there being a democrat president. If Warren or Booker are elected president next, won't we see another spike?

Fake news about fake news. Fake news for most of these people is determined by a "fake" chyron header, placed, not place, or removed and added at a whim.

Facebook and Twitter are purposefully or inadvertently in the misinformation business (I lean purposefully). Their entire model - good or bad intentioned (I lean bad intentioned) - is about the rapid dissemination of single-source information while providing algorithmic weight to some vs. others (blue check mark/followers).

This will continue and intensify.

Twitter is only algorithmic within the world of those you "follow." There is no way for left-field news stories to be promoted into your "home" timeline, unless one of your "follows" interacts with them. So some care and maintenance of sources is possible.

YouTube is much worse in this regard, shotgunning everyone with rising memes.

https://twitter.com/smartereveryday/status/1091833011262423040?s=19

Twitter jumps its boundaries all the time. There are millions (possibly billions) of twitter posts views coming to you from outside Twitter and its algorithm all the time. This is absolutely intentional as it allows Twitter additional impact beyond those views you intend to see as well as for the huge cohort of humanity which are not even Twitter users.

The relevance of their entire platform (Facebook's too) is "You will be made to care about what goes on here." They have been very successful at this. The leap from said directive to "You will be made to care about this thing." is so linear it almost doesn't bear mentioning.

This strikes me as a criticism of the information age as a whole - with or without any specific platform or player.

It is a criticism of a dearth of critical thinking - not specific to any age.

What a misfortunate world we occupy, that yields such phenomena:

Thought: the chief impediment to speech.

Talk: the movement of tongue and jaw that precedes inaction.

Science experiment you can conduct at home: Open a brand-new Twitter account with a brand-new email address. Follow no one. View your "home" timeline. If it's empty, you are correct. If it's not empty, then that must mean that some set of business rules have been set up to determine feed content.

That may be different, as they want you to get started. But with 100 < X < 200 follows, I don't see anything unconnected. Indeed the latest Android app tells me how each item is connected (perhaps someone I follow liked it, or perhaps a critical mass of (listed) people I follow saw it).

It establishes that business rules are driving user content at least some of the time. It would be odd if those rules just stopped working after a while, although I don't doubt that who a user follows and what those people like influences a user's own content.

This is just stupid. I've told you how the platform works when you're on board, and you prefer the corner case when you're just getting started and Twitter is making recommendations.

That is just the corner case.

Scenario: My friend Bob "likes" and "retweets" five things. Four are links to Zero Hedge and the other is a link to some Anderson Cooper something or other. My news feed only shows me the Anderson Cooper thing.

Strictly speaking, you'd be right about the algorithm, but if that kind of thing happens all the time then it leaves room for the criticism that the algorithm disproportionately favors 25% of content.

I, personally, don't know how the algorithm makes its decisions, and I despise Twitter, so I have no skin in the game here. But if you were trying to suggest that the matter is simple and transparent, I don't think you've managed to convince me.

You "follow" Bob if you want to see all his stuff, including retweets.

There is no Twitter "friend" concept or control.

You might see and "like" other tweets when you expand a conversation, or after a search, but those likes do not cause those new people to ever appear in your timeline.

The reason I am a little bit animated here is that it really is remarkably transparent that Twitter gives you the *reason* every item appears in your timeline.

I just reviewed my "home" view and can confirm that every entry is directly from a "follow" or is annotated with how it got there.

Apparently 14 people I follow each follow George Conway, so an occasional tweet from him pops in.

With Trump as president, why would his followers go to Facebook in search of fake news? They wouldn't, for they get all the fake news they want by following Trump on Twitter. It's the supply side explanation.

Only we old people go to Facebook.

Truth Tornado/Tweet from President Donald J. Trump:

"Democrats are becoming the party of late-term abortion (me: infanticide), high taxes, open borders, and crime!" 8:30AM, 31 January 2019 AD

Howard Schultz 2020.

Becoming? What were they before? I love President Donald J. Trump more than my own children, but this is a rare gaffe.

True!

If I edited his tweets, it would read. "The Democrat Party is the party of ignorance and evil."

It's a shame your talents are wasted here. People need to know about the Democrat traitors.

Although I don't believe this claim, specifically, a similar one strikes me as being more plausible: If the objective of fake news was to promote the election of Donald Trump, then a decline in fake news after the election of Donald Trump makes perfect sense.

However, this implies that there is also reduced expenditure on a fake-news reelection campaign.

Rayward, my apologies. Your point speaks directly to engagement, which is the topic of the article. My point didn't even address engagement, and thus may not be relevant even if true! I retract. :)

I'd like to start with a definition of "fake." If by that we mean fully contrived stories with no basis in fact. Well okay.

Because of course even the major media have been offering mediated, curated, cherry-picked topics, selective emphasis, unasked questions, etc. - pretty much since the beginning of time.

What's changed is that this approach to storytelling has been democratized, monetized, and weaponized.

This is a good point. The Murray Gell-Mann Amnesia effect is a good example of this. Newspapers and TV used to be the only source of much day to day data until recently. So like one you have only one clock, it is always correct. Now we have thousands of clocks so we debate a lot more which is correct, of course some of them are totally obviously wrong - but it doesn't mean that the official clocks were totally correct.

The Gell-Mann effect. That's great. I had not heard it.

I often use teachable moments from or local newspaper. Some story will upset a friend/neighbor because the reporter got it so wrong, they end up wondering if that reporter was actually even at the same event as them.

I just read Crichton's article on the Gell-Mann effect and discovered that his claim about Antarctic sea ice is totally, stupidly wrong. (Pay attention to the difference between land ice and sea ice, the latter does not contribute to sea level rise.) So why would I believe what he says about the Gell-Mann effect? LOL.

Remember the Maine!

The world is round!

Incidences of fake news are not declining among the lying, liberal media [redundant].

Can anyone here identify the date on which they ceased pretending to be journalists?

When did President Donald J. Trump declare his candidacy? There's your answer.

By the way I love that 'redundant' joke, and I appreciate you using it all the time. It never gets old!

#MAGA2020

This story was indeed after Trump declared his candidacy:

https://www.reuters.com/article/us-russia-lavrov-norway-refugees-idUSKCN0V41DN

From the second paragraph: "Norway’s right-wing government toughened policies for asylum seekers in December and wants to send back those who crossed the Arctic border from Russia, saying Russia is safe for them. "

Since Norway's government in a small fraction of their decisions did something a red wine sipping soft cheese eating liberal is against, the Norwegian government is suddenly "Right Wing". This is ridiculous. Norway hasn't had a (de facto) right wing government since 1940-1945, and even that one was socialist.

Both Reuters and Spiegel qualify as fake news.

Facebook has put a lot of time, effort, and money into reducing the amount of fake news its users see. Seems like it's working.

Sometimes the answer is so obvious that it's hard for people to see it.

I am wondering if stories by the lying German Pulitzer price winner was counted as fake news

http://www.spiegel.de/international/the-relotius-case-answers-to-the-most-important-questions-a-1244653.html

The two authors of this Medium article tried to tell Spiegel that they had published faker news to no avail:

https://medium.com/@micheleanderson/der-spiegel-journalist-messed-with-the-wrong-small-town-d92f3e0e01a7

Only when a journalist told them that a journalist was lying did The Spiegel believe it.

Der Spiegel itself tells another story (but they would, wouldn't they?) - 'That misery came in the form of an email, one which, as chance will have it, arrived some 17 hours before the glamor of the awards ceremony, at 3:05 a.m. The message came from a woman named Jan, short for Janet, who was doing media work for a vigilante group in Arizona conducting patrols along the border to Mexico. She asked Relotius -- who two weeks earlier had written an article ostensibly about this vigilante group in the darkly dazzling DER SPIEGEL report "Jaeger's Border" -- what exactly he was up to. How, she wanted to know, could Relotius have written about her group without even bothering to stop by for an interview? She found it very strange, she wrote, that a journalist would write stories without gathering facts locally.

The story "Jaeger's Border" would prove to be Relotius' undoing. It was one fabricated story too many, because this time, he had a co-author, who sounded the alarm while also collecting facts to counter his fiction. That co-author, Juan Moreno, has been traveling the world as a reporter for DER SPIEGEL since 2007. In the dispute with and surrounding Relotius, Moreno risked his own job, at times even desperately seeking to re-report his colleague's claims at his own expense. Moreno would go through three or four weeks of hell because his colleagues and senior editors in Hamburg didn't initially believe that Relotius could be nothing more than a liar.

In late November and into early December, some at DER SPIEGEL even believed that Moreno was the real phony and that Relotius was the victim of slander. Relotius skillfully parried all allegations and all of Moreno's well-researched evidence, constantly coming up with new ways of sowing doubt, plausibly refuting accusations and twisting the truth in his favor. Until, ultimately, his tricks stopped working. Until he could no longer sleep at night for fear that he might get caught. Relotius caved in last week when a superior, Özlem Gezer, deputy head of the "Gesellschaft" section where he worked, confronted him and told him outright that she no longer believed him. On Thursday, he sat down with his section head and the editor-in-chief and came clean -- or at least his version of clean.' http://www.spiegel.de/international/zeitgeist/claas-relotius-reporter-forgery-scandal-a-1244755.html

Oddly though, Der Spiegel comes out looking poorly from its own reporting about its own failures.

One of Der Spiegel's "star" reporters was an outright fraud. Of course they come off poorly. You don't get a gold medal for confessing after you were caught lying.

The key point is, as RatInPutinsMaze points out, how many of the remaining authors are cheats that haven't been caught yet.

canadian snowrollers or
gender positive postmodern canuck conquistador/as with itchy red rashes who didn't get their measles vaccine
you make the call
https://www.oregonlive.com/news/2019/02/vancouver-area-measles-outbreak-grows-to-43-cases.html

I have yet to see an unfake news source, how did they do the split?

Seems like an example of looking for the horses that have already fled the barn. Those existing sites may have lost their engagement but this study says nothing about new sites or methods for sharing fake news.

https://www.washingtonpost.com/lifestyle/2019/02/04/this-man-with-no-hands-can-shave-grill-drive-car-he-helps-other-amputees-learn-how/?utm_term=.6d5fef526fd4&wpisrc=nl_rainbow&wpmm=1

This seems ambiguously worded:
"In comparison, interactions with other news, business, or culture sites have followed similar trends on both platforms. "
So, if all sorts of stories are following the same trend, doesn't that just show a movement from FB to Twitter in general, with no conclusions possible regarding 'fake news'? Or did they forget the word "not" in the sentence (not that that ever happens)?

Comments for this post are closed