The Dangers of Fake News Spread to Data Visualization

    by Russell Chun
    February 23, 2017
    What might data visualizations add to fake news? Original photo from Library of Congress.

    Fake news stories defined our 2016 election and continue to threaten the social media ecosystem, our lives, and even international diplomacy. Already, the issue has prompted changes in business and sharing practices for Facebook and Google. But despite organizational initiatives to battle the problem, it’s likely to get worse in 2017. Why? For the most part, fake news stories are still relatively crude and unsophisticated, with nothing more than text, a photo grabbed off the internet, and hyper-partisan, attention-grabbing headlines to lure clicks and shares. But fake news creators might begin to include one thing that journalists and researchers already know works to increase engagement: data visualizations.

    Readers Believe Data Visualizations

    The mere inclusion of a simple data visualization, like a chart or map, has shown to significantly increase a story’s believability, whether true or not. Researchers from Cornell University tested readers with two articles on a scientific claim, both identical except in one regard. One included a graph and the other did not. While only 68 percent of readers believed the claim in the article without a graph, nearly all –97 percent– of readers believed the same claim with the graph included. Similar effects were seen when a chemical formula was included, suggesting that the persuasive effects are based on the association of a graph with what the researchers call the “prestige of science.” The researchers conclude that “graphs signal a scientific basis for claims, which grants them greater credibility.” While their conclusions pertain to science stories in particular, it’s just one step to the applicability to news in general to infuse them with an aura of scientific backing.

    "There is a long trail that leads from the raw data to the final visualization, with many opportunities along the way to introduce bogus information, making the accuracy of graphs more difficult to assess."

    Researchers have also demonstrated that graphs can have a measurable impact on attitude changes. In a study from the New York University School of Law, participants were shown data visualizations to bolster arguments on different topics. For those who already held strong beliefs about the selected topics, the graphs had little effect. However, for those who were on the fence about an issue, data visualizations were effective in swaying opinion–a frightening implication for increasing the effectiveness of propaganda and misinformation to reach and to influence the important undecided population.


    Data Visualizations and Social Media

    In addition to the scholarly studies that support the unique power of visual evidence, we know that social media favors the simplicity of the single image. A graph, like a meme, is far more accessible and shareable on Facebook or Twitter than a link to an article. A tweet with an embedded image gets 150 percent more retweets, according to data from Buffer. Sharing an eye-catching data visualization that has an air of credibility (because it’s scientific!) is hard to resist, especially with the low-friction tap of a retweet. A Breitbart story with a bogus map supposedly showing how Trump won the popular vote “in the heartland” was shared widely on Facebook and Twitter. The map itself gives “proof” to those already inclined to believe the headline. A different map purportedly showing how the country would have voted if only millennials participated in the Presidential election went viral and was retweeted hundreds of thousands of times before readers pointed out that the map was based on a survey before the election. The key to spotting the error? The data source, which was cropped out in some versions of the map, pointed to the online poll SurveyMonkey.

    That the data source was easily missed and shared despite its absence hints at how readers are more willing to accept a data visualization without a source than they would a quote without an attribution. But even if a chart clearly cites a credible data source, the methodology, visualization, or framing around the data could be completely off. There is a long trail that leads from the raw data to the final visualization, with many opportunities along the way to introduce bogus information, making the accuracy of graphs more difficult to assess.

    A selection of some of the most shared fake news stories, according to Buzzfeed.


    Fighting Fake News

    As a media arms race emerges between fake news purveyors attempting to deceive and to disseminate doubt, and platforms enabling better ways to detect and debunk, I join with many others to push for better education in media literacy as the best long-term bulwark against misinformation, even more so than any technological measure. While it’s been provocatively suggested that media literacy could actually be the problem and not the solution to fake news, I’m not convinced that the same argument can be applied to visual and data literacy, specifically.

    Having training in numeracy allows us to have a quick “gut check” – the intuitive moment that Malcolm Gladwell describes for decision-making that happens in a “blink,” but relies on early, sustained, and internalized experience. Visual and data literacy should be part of an overall media literacy effort in education. Understanding how graphs and charts are generated from data can dispel much of the mystery that fuels the notion that charts are from experts and hence, more authoritative than other media. Knowing basic chart design helps us not only identify outright lies, but recognize when visual tricks distort the underlying facts.

    Of course, data and visual literacy are valuable beyond recognizing fake news that use graphs and maps. In our increasingly data- and image-saturated environment, those skills become necessities to live, work, and play in an informed manner. We all need to level up our ability to “call bullshit” as two educators from the University of Washington bluntly put it. And we’ll need it more than ever. Recent signals from the Trump administration troublingly suggests that misinformation may become a regular part of the government. The new administration lies and then casually claims “alternative facts” to backup those lies. The president appoints officials with views antithetical to the mission of the agencies they would lead, which should raise red flags about data suppression or distortion. The White House credentialed fake-news purveyor the Gateway Pundit and by doing so validates the site as a legitimate news organization. What all this means is that the reliability of our government-sponsored information and datasets could come into question. Statistics from the government on crime, demographics, employment, and other vital measures have always been the gold-standard, but a Trump administration now forces us to be extra vigilant with federal source data and any visualizations that come from them.

    Russell Chun is an assistant professor of journalism at Hofstra University School of Communication, where he teaches multimedia storytelling and data journalism. Twitter @russellchun.


    Tagged: alternative facts charts data data literacy data visualization fake news graphs maps media literacy social media visual literacy

    Comments are closed.

  • About EducationShift

    EducationShift aims to move journalism education forward with coverage of innovation in the classroom as journalism and communications schools around the globe are coping with massive technological change. The project includes a website, bi-weekly Twitter chats at #EdShift, mixers and workshops, and webinars for educators.
    Amanda Bright: Education Curator
    Mark Glaser: Executive Editor
    Design: Vega Project

    MediaShift received a grant from the Knight Foundation to revamp its EducationShift section to focus on change in journalism education.
  • Who We Are

    MediaShift is the premier destination for insight and analysis at the intersection of media and technology. The MediaShift network includes MediaShift, EducationShift, MetricShift and Idea Lab, as well as workshops and weekend hackathons, email newsletters, a weekly podcast and a series of DigitalEd online trainings.

    About MediaShift »
    Contact us »
    Sponsor MediaShift »
    MediaShift Newsletters »

    Follow us on Social Media