What should I believe? A conjoint analysis of the influence of message characteristics on belief in, perceived credibility of, and intent to share political posts

Abstract

Research on misinformation and misperceptions often investigates claims that have already reached a critical mass, resulting in little understanding of why certain claims gain widespread belief while others fall into obscurity. Here we consider how various message features factor into why certain claims are more likely to be believed, perceived as credible, and shared with others. Using a conjoint experiment, we randomly assigned participants (N = 1,489) to receive an experimentally manipulated message describing an allegation of political misconduct. Results suggest that partisan cues play a significant role in influencing both belief and perceived credibility. Furthermore, message specificity, language intensity, and whether other users’ comments on the post refute or endorse the post also influenced belief judgment and credibility assessments. We conclude with a discussion of the theoretical and practical importance of these findings for understanding and combating the threat of misinformation.