An examination of partisan disinformation sharing behavior on social media

The study attempts to develop an ordinal logistic regression model to identify the predictors of partisan disinformation sharing behavior using the survey data collected from 500 U.S. adults. Since the 2016 U.S. presidential election, disinformation has become a quotidian leitmotiv in American politics. Previous research has concentrated on examining individual attitudes towards disinformation susceptibility, but relatively few studies have investigated how disinformation is transmitted from a behavioral perspective. Instead of asking "why people believe it," this study focuses on "how people share it." The following question is addressed in this study: How partisanship, political literacy, educational attainment, information credibility, social media engagement, and one's demographic characteristics contribute to partisan disinformation sharing behavior? To answer the research question, I designed an online survey experiment, in which 500 respondents were recruited using Amazon's Mechanical Turk crowd sourcing platform. The survey was fields in April 2020. In the survey experiment, research participants were randomly assigned to four groups and were asked to report how they would share a set of fabricated social media postings. To rule out confounding factors, respondents were presented with fabricated social media messages unrelated to real-world events and political actors. I find that partisans, including Democrats and Republicans, are more likely to share politically congenial disinformation. In other words, Democrats are more likely to share political disinformation favorable towards Democrats and unfavorable towards Republicans; Republicans are more likely to share political disinformation favorable towards Republicans and unfavorable towards Democrats. I also find that political knowledge is negatively associated with disinformation sharing behavior: The more politically literate an individual is, the less likely the individual shares disinformation on social media. In addition, males are more likely to share partisan disinformation as opposed to females. Age and social media engagement effectively predict one's disinformation sharing behavior, and they are generally positively associated with disinformation sharing behavior. Surprisingly, information credibility is not a robust predictor of partisan disinformation sharing behavior.