Why are some countries more prone to frequent anti-government protests than others under similar conditions? Using Bayesian networks to explore causal misconceptions, we propose a model of subjective belief formation that gives a prominent role to history in explaining how protest cultures are formed and persist. We propose that past anti-regime actions may influence the inferences citizens make regarding the effect of protests on economic performance, even in the absence of a direct causal relationship. When regime strength is related to both economic performance and protests, it induces a correlation between the two variables, where the strength of the correlation is driven by the empirical frequency of protesting in the polity’s history. When citizens believe in a strongman narrative that postulates that regime strength produces good economic outcomes, citizens reason that fewer instances of past protests suggest a stronger correlation, rendering protests less likely in the current period.