Examining Engagement Behavior on Social Media Related to Fake News

Abstract

Considering political events in the aftermath of the 2016 presidential election fake news is a hotly debated topic. This topic is made more complicated by the growing prevalence of social media as a source of news. This research examines what factors contribute to the engagement behaviors with fake news on social media. A review of the literature uncovered a gap in theoretical models regarding the engagement with real and fake news on social media. The Truth Bias and Elaboration Likelihood Model (ELM) were used as foundations to build a new model to help explain social media behaviors. Truth Bias was relevant due to the social nature of engagement, in particular relational development. The ELM expands the truth bias model with the explanation of central and peripheral routes of messages. Critical consumption media literacy helped to fill in the gaps between the two models above in a social media context. An experiment was designed to investigate the relationships described therein. The results of this experiment show support for the model as well as distinct differences between engagement behavior with real and fake news headlines. While the original truth bias factor of relational development displayed the expected relationships with the truth bias construct, truth bias itself displayed the opposite of the anticipated relationship with behavior with fake news. Overall political alignment did not perform as expected. Political alignment with the user profile was significantly related to relational development for only real headlines, while neither form of political alignment (user profile or outlet) were significant in the fake headlines group. Generalized communicative suspicion was a significant moderator in both the models, although to different degrees. Additionally, issue involvement and critical consumption media literacy were found to be moderators in the real headlines group but were direct effects on behavior in the fake headlines group. Interestingly, several control variables were significant for fake news headlines, including the perceived realism, whether the article had been seen before, the gender, age, and the respondent’s political party. Of these control variables, only age was found significant for the real headlines group. The results imply that truth bias plays a role in engagement behaviors and that a different set of factors contribute to the engagement with fake news headlines than with real news headlines.