Hours after the Las Vegas massacre, Travis McKinney’s Facebook feed was hit with a scattershot of conspiracy theories. The police were lying. There were multiple shooters in the hotel, not just one. The sheriff was covering for casino owners to preserve their business.
The political rumors sprouted soon after, like digital
weeds. The killer was anti-Trump, an “antifa” activist, said some; others made
the opposite claim, that he was an alt-right terrorist. The two unsupported
narratives ran into the usual stream of chatter, news and selfies.
“This stuff was coming in from all over my network of
300 to 400” friends and followers, said Mr. McKinney, 52, of Suffolk, Va., and
some posts were from his inner circle.
But he knew there was only one shooter; a handgun
instructor and defense contractor, he had been listening to the police scanner
in Las Vegas with an app. “I jumped online and tried to counter some of this
nonsense,” he said.
In the coming weeks, executives from Facebook and
Twitter will appear before congressional committees to answer questions about
the use of their platforms by Russian hackers and others to spread
misinformation and skew elections. During the 2016 presidential campaign, Facebook
sold more than $100,000 worth of ads to a Kremlin-linked company, and Google
sold more than $4,500 worth to accounts thought to be connected to
the Russian government.
Agents with links to the Russian government set up an
endless array of fake accounts and websites and purchased a slew of
advertisements on Google and Facebook, spreading dubious claims that seemed
intended to sow division all along the political spectrum — “a cultural hack,”
in the words of one expert.
Yet the psychology behind social media platforms — the
dynamics that make them such powerful vectors of misinformation in the first
place — is at least as important, experts say, especially for those who think
they’re immune to being duped. For all the suspicions about social media
companies’ motives and ethics, it is the interaction of the technology with our
common, often subconscious psychological biases that makes so many of us
vulnerable to misinformation, and this has largely escaped notice.
Skepticism of online “news” serves as a decent filter
much of the time, but our innate biases allow it to be bypassed, researchers
have found — especially when presented with the right kind of algorithmically
selected “meme.”
At a time when political misinformation is in ready
supply, and in demand, “Facebook, Google, and Twitter function as a
distribution mechanism, a platform for circulating false information and
helping find receptive audiences,” said Brendan Nyhan, a professor of
government at Dartmouth College (and occasional contributor to The Times’s
Upshot column).
For starters, said Colleen Seifert, a professor of
psychology at the University of Michigan, “People have a benevolent view of
Facebook, for instance, as a curator, but in fact it does have a motive of its
own. What it’s actually doing is keeping your eyes on the site. It’s curating
news and information that will keep you watching.”
That kind of curating acts as a fertile host for
falsehoods by simultaneously engaging two predigital social-science standbys:
the urban myth as “meme,” or viral idea; and individual biases, the automatic,
subconscious presumptions that color belief.
The first process is largely data-driven, experts
said, and built into social media algorithms. The wide circulation of bizarre,
easily debunked rumors — so-called Pizzagate, for example, the canard that
Hillary Clinton was running a child sex ring from a Washington-area pizza
parlor — is not entirely dependent on partisan fever (though that was its
origin).
For one, the common wisdom that these rumors gain
circulation because most people conduct their digital lives in echo chambers or
“information cocoons” is exaggerated, Dr. Nyhan said.
In a forthcoming paper, Dr. Nyhan and colleagues
review the relevant research, including analyses of partisan online news sites
and Nielsen data, and find the opposite. Most people are more omnivorous than
presumed; they are not confined in warm bubbles containing only agreeable
outrage.
But they don’t have to be for fake news to spread
fast, research also suggests. Social media algorithms function at one level
like evolutionary selection: Most lies and false rumors go nowhere, but the
rare ones with appealing urban-myth “mutations” find psychological traction,
then go viral.
There is no precise formula for such digital catnip.
The point, experts said, is that the very absurdity of the Pizzagate lie could
have boosted its early prominence, no matter the politics of those who shared
it.
“My experience is that once this stuff gets going,
people just pass these stories on without even necessarily stopping to read
them,” Mr. McKinney said. “They’re just participating in the conversation
without stopping to look hard” at the source.
Digital social networks are “dangerously effective at
identifying memes that are well adapted to surviving, and these also tend to be
the rumors and conspiracy theories that are hardest to correct,” Dr. Nyhan
said.
One reason is the raw pace of digital information sharing,
he said: “The networks make information run so fast that it outruns
fact-checkers’ ability to check it. Misinformation spreads widely before it can
be downgraded in the algorithms.”
The extent to which Facebook and other platforms
function as “marketers” of misinformation, similar to the way they market shoes
and makeup, is contentious. In 2015, a trio of behavior scientists working at
Facebook inflamed the debate in a paper published
in the prominent journal Science.
The authors analyzed the news feeds of some 10 million
users in the United States who posted their political views, and concluded that
“individuals’ choices played a stronger role in limiting exposure” to contrary
news and commentary than Facebook’s own algorithmic ranking — which gauges how
interesting stories are likely to be to individual users, based on data they
have provided.
Outside critics lashed the study as self-serving,
while other researchers said the analysis was solid and without apparent bias.
The other dynamic that works in favor of proliferating
misinformation is not embedded in the software but in the biological hardware:
the cognitive biases of the human brain.
Purely from a psychological point of view, subtle
individual biases are at least as important as rankings and choice when it
comes to spreading bogus news or Russian hoaxes — like a false report of Muslim
men in Michigan collecting welfare for multiple wives.
Merely understanding what a news report or commentary
is saying requires a temporary suspension of disbelief. Mentally, the reader
must temporarily accept the stated “facts” as possibly true. A cognitive
connection is made automatically: Clinton-sex offender, Trump-Nazi, Muslim
men-welfare.
And refuting those false claims requires a person to
first mentally articulate them, reinforcing a subconscious connection that
lingers far longer than people presume.
Over time, for many people, it is that false initial
connection that stays the strongest, not the retractions or corrections: “Was
Obama a Muslim? I seem to remember that....”
In a recent analysis of the biases that help spread
misinformation, Dr. Seifert and co-authors named this and several other
automatic cognitive connections that can buttress false information.
Another is repetition: Merely seeing a news headline
multiple times in a news feed makes it seem more credible before it is ever
read carefully, even if it’s a fake item being whipped around by friends as a
joke.
And, as salespeople have known forever, people tend to
value the information and judgments offered by good friends over all other
sources. It’s a psychological tendency with significant consequences now that
nearly two-thirds of Americans get at least some of their news from
social media.
“Your social alliances affect how you weight
information,” said Dr. Seifert. “We overweight information from people we
know.”
The casual, social, wisecracking nature of thumbing
through and participating in the digital exchanges allows these biases to
operate all but unchecked, Dr. Seifert said.
Stopping to drill down and determine the true source
of a foul-smelling story can be tricky, even for the motivated skeptic, and
mentally it’s hard work. Ideological leanings and viewing choices are
conscious, downstream factors that come into play only after automatic
cognitive biases have already had their way, abetted by the algorithms and
social nature of digital interactions.
“If I didn’t have direct evidence that all these
theories were wrong” from the scanner, Mr. McKinney said, “I might have taken
them a little more seriously.”