“If it bleeds, it leads.” In news reporting, this prevalent adage describes how negative stories captivate audiences. Consistent with this idea, prior studies have shown that negative information captures attention and is more likely to be read and shared. However, negative messaging can harm mental health, distort memory, and discourage action to address a problem. In a series of behavioral experiments, we investigated the diverse effects of negative framing on behavior and tested an alternate strategy for increasing engagement—evoking positive affect by describing action and progress toward future goals. 

In two experiments, we adapted environmental news headlines to feature different aspects of each story, emphasizing Crisis or Opportunity. Both Crisis and Opportunity framing (and negative and positive affect, respectively) motivated reading and sharing, relative to the unaltered headlines. Crucially, consistent with theoretical predictions, we identified a trade-off: Crisis framing had the strongest effects on immediate engagement (increasing sharing and real charitable donations), but impaired memory for peripheral details; in contrast, Opportunity framing enhanced memory for news content.

In a third study, we investigated real-world engagement with news about climate change by using a large language model to computationally classify content in >25,000 news headlines on social media. Positive and negative affective framing strategies were both associated with increased likes, reposts, and replies. Overall, we demonstrate that the affective framing of news modulates reading, sharing, donations, and memory in laboratory and real-world settings.

In ongoing work, we are conducting an fMRI study with an adapted version of this paradigm. We aim to test whether Crisis and Opportunity framing engage neuromodulatory systems associated with imperative motivation (noradrenergic system) and interrogative motivation (dopaminergic system), respectively. Using representational similarity analyses, we will also explore how the brain represents multiple dimensions of naturalistic information, including semantic content, affect, and valuation.

Project Contributors:
Allie Sinclair, University of Pennsylvania / Rice University
Emily Falk, University of Pennsylvania
Danielle Cosme, University of Pennsylvania
José Carreras-Tartak, University of Pennsylvania