You would know the difference between a “real” news story and a story written for or by an advertiser, right? Especially when a story is labeled “advertisement” right there at the top of your screen. Even if that label was something murkier like “BrandVoice” and pushed down to the bottom of the story, most of us would recognize sponsored content as advertising, right?
A few years ago, noting that $3.2 billion had been spent on native advertising in 2014, two University of Georgia researchers set out to take a closer look into what happens in our minds when we see native ads. Bart Wojdynski and Nate Evans defined native advertising as “any paid advertising that takes the specific form and appearance of editorial content from the publisher itself.”
In a fairly straightforward experiment, Wojdynski and Evans started with a news story about advances in lithium ion auto batteries presented under the banner of a recognizable news organization with a headline, byline and all the conventional appearances of an online news story. But the story definitely had a slant. It prominently featured a particular brand of battery in a positive light and quoted only one source—an executive at the company. Then the researchers designed 12 different versions that varied based on whether the content was labeled “Advertisement,” “Sponsored Content,” “BrandVoice” or “Presented by…” and whether the label showed up at the top, middle or bottom of the story.
They sent the 12 different versions to 242 U.S. residents ranging in age from 19 to 73 and representing diverse demographic groups. Each person was randomly assigned only one version of the story. The idea was to see how different labels and different placement of disclosures would activate critical thinking about the story and the company. For example, would the story be seen as more credible or less credible if the label was presented at the top compared to the middle or bottom of the story? The most remarkable finding from the study: Across all conditions, only 17 of the 242 participants even recognized the article as advertising.
Maybe it was just a fluke. Maybe the people taking the survey just weren’t taking it seriously. (Participants were recruited via Amazon’s Mechanical Turk, which is often criticized for the validity of the data it yields.)
Wojdynski and Evans then ran a second study. This time they recruited undergraduates who looked at native advertising while being monitored with eye-tracking software to see if their gaze would track more toward disclosures at the top, middle or bottom of stories. Of the 60 students, only 11 recognized the material as advertising.
Results from a more recent Stanford study of students across six different universities (including Stanford itself), show equally disheartening results in students’ abilities to apply critical thinking in evaluating claims on social media.
Inspired by the growing body of research in native advertising research that followed those early UGA studies, I started working with University of Florida College of Journalism and Communications Assistant Professor of Adverising Eunice Kim and graduate students Yoo Jin Chung and Brad Horn on some new projects. Two theoretical concepts we are looking at now are sidedness (one-sided or two-sided stories) and expectancy violation (expectations can be violated in negative or positive ways). For example, if someone were to read a very one-sided story on CNN about Samsung’s commitment to safety following its phones blowing up, that type of native advertising (in theory) may violate reader expectations and lead to negative attitudes toward CNN and lower perceptions of CNN’s credibility. It’s a researchable question whether or not Samsung’s brand would be helped or harmed. On the other hand, a surprisingly two-sided story as branded content hosted by Samsung might violate expectations in a good way, leading to more positive credibility or attitudes toward Samsung and its products.
We started by developing several conditions and running a pilot study with 97 undergraduates here in our college. By random assignment, 19 of them were presented with a story deliberately written to be one-sided in favor of Samsung—something that we thought would violate expectations on a news site (in this case we put the story under a CNN banner). When asked whether the one-sided CNN story met their expectations, 13 of 19 of our students said yes, and 10 of those 13 identified the story source as “journalists” rather than “Samsung” or “journalists sponsored by Samsung.” Again, this was just a pretest, and we have a long way to go improving our study design with a larger sample before we can draw any conclusions about sidedness and expectancy violations. But we also have a long way to go in educating generations to come in digital media literacy.
Those of us working in media and communication programs in colleges and universities need to up our game when it comes to digital literacy. We can’t expect K-12 educators to bear all the responsibility, and we shouldn’t be lulled into assumptions that digital literacy is mostly a matter of remedial education by the time students reach college. As teacher-scholars and scholar-teachers, we need to apply what we know from theory and research to our curricula, and we need use our social science research skills to seek evidence of success in teaching digital literacy.
We don’t have to start from scratch. Major research-education partnerships such as the Stanford History Education Group that ran the study mentioned above, Stony Brook’s Center for News Literacy, and The Consortium for Media Literacy’s Medialit.org offer excellent resources.
For example, we could start with Medialit.org’s “Five Questions for Media Literacy” to outline intersections of media theory, research, and teaching:
- Who created this message?
- What creative techniques are used to attract my attention?
- How might different people understand this message differently from me?
- What lifestyles, values and points of view are represented in—or omitted from—this message?
- Why is this message being sent?
Even the very first question forces us to reckon with the fact that “fake news” is actually a problematic term, because satire from Comedy Central is different than branded content on Samsung’s web page, which is different than native advertising on CNN, which is different than a deliberately misleading viral Facebook story propagated by kids in Macedonia, etc. Just developing a clear taxonomy of “fake news” is a big step.
We have a shopping list of theoretical concepts that could then be applied: persuasion knowledge model, confirmation bias, source attribution, layered sources, heuristic cues, gatekeeping, source proximity, framing, third-person effect, agenda building, critical-cultural approaches, uses and gratifications, personalization—the list goes on and on.
Our mission, after all, is not just to do the research but also to educate informed media producers and consumers for the future of our democracy.
Tom Kelleher, Ph.D., is professor and chair of the Department of Advertising in the College of Journalism and Communications at the University of Florida. Kelleher joined the UF faculty in 2014. He had been a member of the faculty at the University of Hawaii for 12 years and served as chair of the School of Communications there from 2010 to 2013. He also served on the faculty at the School of Journalism and Communication at the University of North Carolina at Chapel Hill from 2004 to 2006. Kelleher has designed and taught 22 different courses at three flagship state universities (Florida, North Carolina, and Hawaii) and has published in numerous journals including Journal of Communication, Journal of Computer-Mediated Communication, Journal of Public Relations Research, and Journal of Mass Media Ethics. In January 2017, Oxford University Press published his second book, Public Relations, an introductory textbook that covers fundamental student learning outcomes with organic integration of practical ethics and digital, social, and mobile technologies.