Just the facts? How providing the truth can exacerbate ‘fake news’

With concerns growing around the spread of misinformation, particularly about the proliferation of “fake news” from media outlets, many people assume that the act of fact-checking will reduce the incentive for media outlets to publish biased information. Not so fast, cautions David Godes, the Clyde F. and Ruth E. Williams Professor in Business at Johns Hopkins Carey Business School.

“We find, on the contrary, that revealing the truth, such as through the process of fact-checking, may not remedy the problem and, in some cases, may actually exacerbate biased reporting,” says Godes. His recent research paper, “Will the Truth Free Us from Misinformation?” offers support for this conclusion in two domains: American football score predictions and the fact-checking of U.S. news outlets.

Signaling slanted news

While most scholarship into media bias has focused on dealing with consumer response, says Godes, “there’s been very little focus on the supply side of media misinformation— media outlets themselves.”

He argues that biased media outlets cater to a subset of consumers.

“These are people who have a set of beliefs they have already decided upon, which are part of their identity,” he says. “They are relatively less interested in finding out the truth and more interested in finding news outlets consistent with their worldview. They want to build their inventory of stories in order to maintain their group status and build their social currency.”

The act of fact-checking, Godes says, “makes it easier for consumers to sort media outlets to find those that are shading information in their direction.

Fact-checking may thus create an increased incentive for biased media outlets to go even further in slanting information, he argues in his paper, since the “outlet’s profit is directly proportional to the precision with which they can signal to consumers that they are a particular type of provider of slanted news.”

Going to the max in implausibility

To see whether actual media practices were consistent with his theory, he looks first to ESPN.com’s weekly “NFL Nation” score predictions, comparing those made by local sports columnists for regular season games taking place between 2015 and 2018. These predictions offer a rich database and are targeted at customers, he notes, “who are likely to prefer slanted news” toward the team they support. For each game to be played during an upcoming weekend, two writers—one each from the two teams—offer their score predictions.

In his analysis, Godes found that notwithstanding the presence of vast amounts of commonly known information—such as injury reports, home field advantage, or weather forecasts—sports writers on average forecast a score differential that is biased by roughly 2.5 points in favor of “their” team. This is true even though the “truth” will be revealed within days and that readers will, over time, identify biased reporting.

“These results are very much consistent with the view that the constant presence of the truth is not likely to free us from the intentional communication of misinformation,” he concludes in his paper.

Godes next looks at the effect that the 2013 establishment of PunditFact, a nonprofit fact-checking media website, had on false reporting. In his paper, he analyzes more than 1,000 PunditFact checks between 2013 and 2018. He assumes network broadcast news outlets will act like “honest” outlets, which “will communicate to consumers the truth,” and cable news outlets will be more “strategic” in that they “may be motivated to slant the news to fit the preferences of a customer segment.”

After completing his analysis, Godes writes, “We find evidence, consistent with our theory, that cable news outlets present more claims later determined to be false after PunditFact than before, and relative to traditional, legacy broadcast outlets.”

Explaining the motivation that might guide biased news outlets in increasing their level of misinformation after fact-checking, Godes says, “Think of a consumer trying to figure out which outlet is going to give them the slant that they’re looking for. One source of information they use is the stories they read. A fact check though provides yet another signal—possibly a more precise signal—of the outlet’s approach.”

Valuing truth

If fact-checking is counter-productive in the fight against the supply of “fake news,” what other strategy might be more effective? Godes notes that his model doesn’t offer a “silver bullet.” But he says his analysis suggests the best policy option may be “to invest in increasing the proportion of the news-consuming public that places a high value on the truth.” Doing so, he notes, would “reduce supply-side incentives for the production of misinformation.”

Such investment could take the form of public service announcements, he suggests, or creating a school curriculum that educates children from an early age about the importance of living in a society where the truth prevails—a society that values having a free media with honest reporting.

“Emphasizing this point in schools, and through our leaders, educators, and whoever has influence—that’s the strongest policy recommendation I can make,” says Godes.