Researchers created fake news. Here's what they found.

Before the term "fake news" became an all-purpose insult for news coverage a person doesn't like, it had a more specific meaning: stories invented from whole cloth, designed to attract social shares and web traffic by flattering the prejudices of their intended audience. Think of untrue claims like the Trump endorsement by Pope Francis or the investigation of the Clinton Foundation for running a pedophile sex ring.

In the immediate aftermath of the election, there was even some speculation that these types of stories were enough to swing the result toward Donald Trump. Sundar Pichai, the chief executive of Google, raised that possibility, and one author of made-up viral news told The Washington Post that "I think Donald Trump is in the White House because of me."

Some new research from two economists throws at least a bit of cold water on the theory that false news was a major influence on the election result. They offer some hard data on how pervasive voters' consumption of fake news really was during the 2016 election cycle. The research also reveals some disturbing truths about the modern media environment and how people make sense of the incoming gush of news.

[How a 23-year-old wrote a fake news masterpiece]

Hunt Allcott of New York University and Matthew Gentzkow of Stanford commissioned a survey in late November hoping to discern just how deeply some of the fake news embedded itself with U.S. voters. The two asked people, among other things, whether they had heard various pieces of news that reflected positively or negatively on one of the candidates — of three varieties.

There was completely true news: Hillary Clinton called some Trump supporters a "basket of deplorables," for example, or Trump refused to say at a debate whether he would concede the election if he lost.

There was fake news, as identified by fact-checking sites like Snopes and PolitiFact — big things like the Pope Francis story and smaller items, like Trump threatening to deport "Hamilton" creator Lin-Manuel Miranda to Puerto Rico.

The third category was most interesting. The researchers created "fake fake" news. That is, they invented some headlines that were the type of thing fake sites produce, but had never actually been published during the campaign. One of these placebo headlines was that "leaked documents reveal that the Clinton campaign planned a scheme to offer to drive Republican voters to the polls but then take them to the wrong place," and its inverse in which it was the Trump campaign scheming to take Democrats to the wrong polling place.

There is some good news in that more people reported having heard, and believed, the true statements than the false statements. Only 15.3 percent of the population recalled seeing the fake news stories, and 7.9 percent recalled seeing them and believing them.

The more interesting result: Those numbers are nearly identical to the proportion who reported seeing (14.1 percent) and believing (8.3 percent) the placebos, the "fake fake" news stories. In other words, as many people recalled seeing and believing fake news that had been published and distributed through social media as recalled seeing fake news that had never existed and was purely an invention of researchers.

That's a strong indication about what is going on with consumers of fake news. It may be less that false information from dubious news sources is shaping their view of the world. Rather, some people (about 8 percent of the adult population, if we take the survey data at face value) are willing to believe anything that sounds plausible and fits their preconceptions about the heroes and villains in politics.

"False remembering is incredibly correlated with people's priors," Gentzkow said. "One way to think about it is we know our memories are very imperfect. And if you ask me if I saw something, I'm partly asking myself how likely it is that I saw it. People are doing inference."

Allcott and Gentzkow do some math around how influential each fake news story would need to have been on people who read it to have meaningfully affected the election results. They conclude that a single news article would need to have been as persuasive as 36 television campaign ads to have been the difference between a Trump victory or loss.

But there's a bigger, and perhaps more worrying, implication of this research. It suggests that the most straightforwardly fraudulent forms of fake news are a small part of what is shaping how people understand the world. People's hunger for information that suits their prejudices is powerful, and in the digital media age, a pile of it emerges to satisfy that demand.