Seems Legit: An Investigation of the Assessing and Sharing of Unverifiable Messages on Online Social Networks

Unverifiable messages abound on the Internet. Why do people share messages they cannot verify? This study develops an in-depth understanding of how messages containing unverifiable product information differ and why users share such messages over online social networks (OSNs). We develop a classification that identifies different types of unverifiable messages that OSN users encounter prior to the release of a new product and conduct two studies to investigate the resharing of true (information leak) and false (rumor) messages originating from unofficial channels. We contend that such differences (true vs. false) are likely to result in differentiating message characteristics. Employing a dual-processing theoretical lens, we further hypothesize that because these messages are unverifiable, recipients will take a holistic approach and rely on both content (plausibility) and noncontent (vividness, sender credibility) message characteristics when assessing the message. Specifically, when faced with an unverifiable message, the presence of content characteristics amplifies the effect of noncontent characteristics, suggesting that plausibility enhances the value of vividness and sender credibility. These characteristics jointly help recipients assess a message's diagnosticity and novelty, which are the primary psychological factors driving the reshare decision. We employ a multimethod approach with Study 1 leveraging secondary data collected from Twitter to assess objective behavior and Study 2 employing a controlled experiment to assess psychological processes. Together, the studies offer compelling evidence in support of our model, indicating that leaks and rumors exhibit different message characteristics; that recipients employ a synergistic processing strategy when assessing unverifiable messages; and that unverifiable messages are reshared when they are perceived to be helpful or novel. The findings from this research have implications for both research and practice.