By LEVI BOXELL, MATTHEW GENTZKOW and JESSE M. SHAPIRO DEC. 6, 2017,
New York Times [Original article contains link and charts.]
image from article
Is social media responsible for our democracy’s current crisis? An increasing amount
of political information (and misinformation) gets disseminated online, and many
Americans do not trust the media, do not trust Congress and do not trust the
president. By many measures, voters are as polarized now as they have ever been in
recent memory.
Many observers — even, before he left office, President Barack Obama — have
tagged social media as a key driver of this crisis. The digital world offers no shortage
of potential villains: targeted Russian ads; shadowy purveyors of fake news; political
consultants like Cambridge Analytica wielding big data and cutting edge psychology;
and formerly fringe media players like Breitbart leaping into the mainstream.
But we risk giving too much weight to the newest and most frightening media
technologies. If any media platform is to blame, it is not the web. It is more likely
television, which is a more important source of political information. Growing
polarization may also result from structural economic changes, like rising inequality,
that have occurred in recent decades.
A few facts can help keep the role of social media in perspective. The share of
Americans who use social media as their primary source of political news and
information is rising fast but remains relatively small. Recent work by Mr. Gentzkow
and Hunt Allcott finds that only 14 percent of American adults reported that social
media was their most important source of news for the 2016 election. On the other
hand, 57 percent of American adults said that TV (cable, network or local) was their
most important source.
It’s also important which demographic groups use social media. In a paper
recently published in the Proceedings of the National Academy of Sciences, we show
that polarization has been growing as fast or faster among elderly Americans — those
least likely to use social media — as among those aged 18 to 39. This applies across a
broader set of demographic groups: Polarization increased as fast or faster among
the Americans least likely to use the internet compared with those who are most
likely.
In the 2016 election, President Trump was most popular among demographic groups
least likely to use social media. According to our calculations, Mr. Trump gained
support relative to Mitt Romney among non-internet-using voters, but actually lost
support among internet-using voters. An analysis by the media researchers Keith
Hampton and Eszter Hargittai likewise finds that Hillary Clinton’s supporters were
more likely to use Twitter and Reddit than Mr. Trump’s supporters.
And polarization was climbing steadily long before the rise of social media. In
our study, we find it has been growing since the 1980s — long before the internet, let
alone Facebook or Twitter, became popular choices for media consumption. We see
no clear increase in this trend in the period when digital sources were introduced.
Social media are likely to grow in importance over time, and none of these facts
rule out its having important effects today. We should be concerned about the effects
it has on its users, even if these users do not account for the bulk of overall
polarization. Young voters who use social media may share polarized views with
older voters who do not, and inflammatory content on social media can be picked up
and amplified by more mainstream outlets. Nevertheless, we believe these and other
data suggest social media are unlikely to be a main cause of rising polarization in
America. We think it is important not to lose sight of other factors that may play a
more important role.
Which brings us back to television — and in particular, the rise of partisan cable
like Fox News and MSNBC. A recent analysis using large-scale data identifies cable
television news as a major contributor to polarization. This narrative arguably fits
with the timing of the rise in interparty animus, and it is consistent with the rise in
polarization among groups — such as the elderly — with limited internet use but
high rates of television viewing.
We would also look to the behavior of politicians and the parties. Data from
congressional roll-call voting show that the increase in polarization in the Senate and
House started well before it can be detected for voters, and though elected leaders
follow their constituents much of the time, they can also lead them. Today the
parties seem to speak different languages, with Republicans talking about “illegal
aliens” and “the death tax” while Democrats talk about “undocumented workers”
and “the estate tax.” An analysis of the Congressional Record that Gentzkow and
Shapiro conducted with the statistician Matt Taddy finds that this linguistic rift
opened up right around the time of the Contract With America, when the Republican
leadership adopted a successful strategy of using wordcraft to frame the issues of the
day.
Where we suspect the most important causes lie is in the deeper structural
changes that have caused the experiences of those in the red and blue parts of the
country to diverge. A voter’s party identification is increasingly related to his or her
position in the income distribution, with the top quintile containing
disproportionately more Republicans. At the national level, income inequality and
polarization in Congress track each other closely. Furthermore, congressional
districts that were adversely affected by the rise of Chinese imports have been shown
to elect less centrist representatives from both parties.
The social-media-as-villain narrative is gripping because it plays vividly to our
fears. But that is not the only reason it holds such sway. It also lets us collectively off
the hook. Why are half of Americans thinking and acting in ways the other half
cannot comprehend? Why did one of those halves choose Donald Trump to be their
president? Easy, the social media narrative would say: They were brainwashed. They
were duped by the bad guys — fake news or Russian robots or big-data-driven
algorithmically targeted psi-ops propaganda.
At some fundamental level, they don’t really mean what they are saying, and if
only they weren’t so gullible or so vulnerable they would see things our way. To find
solutions we don’t need to look at our own behavior, or values, or consumption
patterns — we just need to beat the bad guys at the gate.
As tempting as that story is, it is at best incomplete. Like many inflection points
in history, this one was probably not caused by any single change, but by the fact that
many important changes happened to converge at the same time. The factors that
likely matter the most are those that have caused the real experiences of Americans
to diverge.
Levi Boxell is a Ph.D. student in economics at Stanford, where Matthew Gentzkow is an
economist. Jesse M. Shapiro is an economist at Brown.
No comments:
Post a Comment