Considering the Effects of Embedded Conflict and Sensationalization: Results from a Message Effects Experiment on Conflicting Health Information

There is evidence that exposure to conflicting health information—are eggs really good for you, or not?—can spill into the rejection of unrelated (and non-controversial) health recommendations. But the boundaries of these effects are unclear. Grounded in Nagler’s model of spillover effects, this experiment assesses whether spillover is amplified by the use of sensationalized language or by the level of conflict already associated with a topic. U.S. adults (N = 700) were randomly assigned to read four short health news stories or to a no-message control group. Among those exposed to stories, two factors were crossed in a 3 × 2 factorial design: story-level conflict (no conflict, conflicting findings described plainly, or the same conflicting findings described with sensationalized language) and topic-embedded conflict (topics rated in a separate study as higher or lower conflict). Stories with conflicting findings increased confusion, backlash, and negative affect and marginally reduced intentions to follow novel, but not established, health recommendations. Adding sensationalized language marginally increased backlash but did not affect other outcomes. Stories on higher-conflict topics decreased—rather than increased—backlash. Findings underscore the importance of distinguishing among different operationalizations of conflicting information.