Key Takeaways from the Project
Background
Little is known about how different linguistic forms of headlines affect the spread of misinformation. This study examined the effects of various linguistic forms of true and false headlines on users’ click and share intentions. Our hypotheses were that interrogative headlines with questions that generate knowledge gaps between the headlines and the full articles would be marked by participants as more clickable, whereas assertive headlines with explicit and concrete information would be marked by participants as more shareable.
Method
536 participants were recruited via Prolific. They were presented with 36 COVID-19-related (18 true and 18 false) headlines and asked to indicate their intent to click and share. We manipulated the headlines into one of nine speech acts, which we classified into three categories for analysis: assertive, interrogative, and a combination of both. This manipulation enabled us to build headlines with explicit information and a lack of gaps, as well as headlines with implicit inferences and gap-inducing properties.
Results
We employed Binomial Logistic Regression to examine the binary choice between the intent to click/share or not. The results showed that the combined forms and the interrogative forms were significant unique predictors of both click and share intentions. People were less likely to click “yes” and share “yes” for the combined and interrogative headlines than for the assertive headlines. Whether the headline was true or false was not a significant predictor of both click and share intentions.
Discussion
While our hypotheses were only partially supported by the findings, they suggest that linguistic forms are significant predictors of clicking and sharing intentions.
Video
Click below to view the discussion on Youtube.