Examining Individual Differences in Misinformation Engagement among Social Media Users in a Simulated Twitter Task

Background Psychological traits and social identities have been linked to misinformation sharing, but most studies measure passive or low-effort engagement. Less is known about high-effort engagement that requires users to write a response. We examined whether cognitive reflection, motivated reasoning variables (political orientation, religiosity, trust in medical scientists) and personality traits are differentially associated with liking, retweeting, replying to, and quoting COVID-19 misinformation and public-health-unaligned tweets.Methods US MTurk participants (N = 858; September 2021) engaged with real COVID-19 tweets in a simulated Twitter feed. Replies and quotes were manually coded as endorsive or combative. We fit OLS regressions for each engagement type (N = 851) and estimated relative importance using Shapley (LMG) decomposition of R².ResultsEndorsive replies and quotes were rare (7.3% and 4.3% of participants for misinformation). Political conservatism and religiosity were associated with endorsive misinformation replies, and motivated reasoning variables were the largest contributors in all reply and quote models, though these models explained little variance (adjusted R² ≤ .03). Cognitive reflection was negatively associated only with liking, and grandiose narcissism was the largest contributor to retweeting.ConclusionClassical reasoning, personality and motivated reasoning explanations of misinformation propagation were each supported, depending on the type of engagement. Simulation tasks that measure multiple engagement behaviors can help clarify inconsistent findings in the misinformation literature.