Is there any Internet-age euphemism more telling than “sponsored content”? Publications have been running these sly advertisements—once known as advertorials—since at least the mid-20th century, always labeling them to differentiate them from editorial material. (Slate is one of them.) You’d think that we would have learned to recognize them by now, but a new study indicates that a “significant minority of consumers” may still be in the dark. Ultimately, however, that study doesn’t prove how stupid readers are so much as it demonstrates the elaborate efforts advertisers put into making us trust them.
The study’s authors, University of California, Berkeley, Internet law professor Chris Jay Hoofnagle and lawyer Eduard Meleshinsky, observe that most investigations of paid content devolve into handwringing over its meaning for journalism more generally. Taking a different tack, their work considers these messages from the perspective of consumer perception, asking whether readers could differentiate sponsored content from other materials in a publication. To this end, they confronted 598 survey respondents with a mock blog containing three articles. While two were apparently unobtrusive and inoffensive (the study offers no description of their content), the third was “a realistic advertorial” for what appears to be amphetamines. Though they labeled it as a “sponsored report,” they didn’t specify an author or otherwise indicate its origins.
Surprisingly, that label didn’t keep study participants from misidentifying the source of this dubious post. When Hoofnagle and Meleshinsky asked their study participants who they thought had composed the advertorial, fully 27 percent “thought that journalists and editors wrote the material” while another 29 percent expressed uncertainty. While the authors had drawn their study respondents from a variety of pools—alcoholics, consumers in debt, subscribers to financial publications, and (their control group) general Internet users—they found no significant variance between the groups. Approximately one-quarter of all those they surveyed seemingly confused the paid post with an unbiased one, regardless of other demographic details.
Another, still more telling detail emerged in the process of this inquiry. The advertorial included a deceptive pull-quote about the qualities of the pills it was promoting, accompanied by a photograph of a smiling woman wearing a white Oxford shirt. Half of the respondents saw a version of this image in which the woman poses against a white background, while the other half received one that positions her in front of a wall of blue containers that roughly resemble pill bottles. Those who saw the latter image were disproportionately more likely to identify the woman as a medical professional of some kind, despite the absence of more conventional paraphernalia such as a lab coat. Focusing primarily on the color—and mostly avoiding the pharmaceutical connotations of the containers—Hoofnagle and Meleshinsky suggest that this shows how something as simple as hue can deceive consumers about authority and trustworthiness.
Significantly, Hoofnagle and Meleshinsky repeatedly acknowledge that their results may not be generalizable, and not just because their palette-shifting claims would require more experimentation. As they note, their respondents were drawn from a nonrandom pool of paid participants, none of whom were younger than 34 and more than two-thirds of whom were women, meaning that they weren’t working with a representative sample. Still, they propose that they may have actually underestimated consumer confusion, since the questions they asked may have prompted their test subjects to read more skeptically than they would have otherwise.
It’s possible, however, that this was offset somewhat by the “simulated” quality of the blog they exhibited to participants. Much as we’re more likely to trust a doctor or pharmacist—however we come to identify someone as such—than we are, say, an otherwise unqualified Oxford-clad model, we inevitably find some venues more credible than others. Under ordinary circumstances, we presumably come to different publications with varying levels of confidence and trust. Most of us are likely more willing to accept, say, a New York Times article at face value than one from the New York Post, and those differences presumably affect how we make sense of any ancillary content that they feature. Hoofnagle and Meleshinsky discuss this very fact in their study, though they puzzlingly don’t explore its connection to their own methodology.
And that’s what really matters, because what’s at issue here is trust. The authors cite two presumably intersecting psychological theories that explain how advertorials get to us. First, they work schematically, confronting us in contexts where we’re not normally trained to read skeptically, as we are when we’re confident that we’re looking at an advertisement. We have the general feeling that we’re reading a magazine, so at a basic level we proceed as we would while doing so, even when we know something’s amiss. Second, these promotions work associatively, appropriating the imprimatur of the publication in which they appear. This, of course, is why the lack of context given about the experiment’s fake blog matters, as it left participants with fewer evaluative tools than they might have otherwise.
But this may be why the picture—whatever the reason for the change in trustworthiness—matters so much. In combination with the other elements of the study, it shows how advertisers can rely on subtle details to sidestep, if not outright undermine, our critical instincts. Those small moves matter more today than ever, thanks to the way the Internet has taught us to Google our way to easy answers, making it simple to quickly resolve some broad questions and easier to avoid small ones. In this climate, those little things—context, color, and the like—can add up, inspiring us to believe things that we would otherwise ignore.
Hoofnagle and Meleshinsky offer a series of proactive proposals, including suggesting that the Federal Trade Commission create more explicit anti-advertorial regulations. Ultimately, however, their study suggests that the real devil is in the details. And the details, alas, are often ungovernable.