While it’s unlikely to dispel concerns about the impact of fake news on politics and society, a new st udy from researchers at Stanford and New York University suggests that made-up stories circulating on social media most likely did not influence the outcome of the 2016 U.S. presidential election. It’s worth noting however that this conclusion is based on a model incorporating a number of assumptions that are open to dispute.
Using data drawn from the archives of fact-checking sites, audience data, and the results of a new survey of 1,200 U.S. adults, the study determined that fake news stories favoring Donald Trump were shared around 30 million times on Facebook, versus 7.6 million times for fake new stories favoring Hillary Clinton. Further, the average American saw and remembered 0.92 fake news stories favoring Trump, versus 0.23 for Clinton.
Although this reveals a definite slant in favor of Trump in terms of the volume of fake news found on social media, based on the survey results a somewhat smaller proportion of Americans reported seeing and also believing a fake news headline: overall they calculate that the average American saw and believed 0.71 pro-Trump stories, versus 0.18 pro-Clinton stories.
In order to assess the impact of fake news in terms of shifting votes, the researchers relied on previous research for a “voter persuasion model,” plus a separate model used to estimate the persuasiveness of TV political ads. Using these models, their analysis attempted to determine whether fake news could have influenced voters enough to shift the national voting results by 0.51% in favor of Trump, giving him victory in the Electoral College.
On this basis, the researchers calculate that for fake news to have shifted the outcome of the election, “a single fake news story would need to have convinced about 0.7 percent of Clinton voters and non-voters who saw it to shift their votes to Trump.” That means that each fake news story would have to be 36 times as persuasive as the average TV political ad. The researchers note: “Our data suggest that social media were not the most important source of election news, and even the most widely circulated fake news stories were seen by only a small fraction of Americans.”
However, there are obvious questions about the assumptions made in the study, including a number acknowledged by the researchers. One of the most important is the use of TV political ads as a benchmark of persuasiveness – an assumption that is open to doubt following the 2016 election, in which Clinton outspent Trump by a considerable margin but was apparently bested by Trump’s social media “ground game.” Even if the political TV ad benchmark is valid, it doesn’t seem impossible that a fake news story could be 36 times as persuasive as a TV ad, since ads are viewed as inherently partisan and untrustworthy whereas news stories are presented as true.On an editorial note, the debate over the impact of fake news on political outcomes would seem to miss the much larger and more important issue – namely, the widespread failure of the American education system to impart critical thinking skills to the public, which must be the case if so many people are indeed ready to accept transparently false news stories at face value.