On modeling the correlates of conspiracy thinking

Abstract

Abstract While a robust literature on the psychology of conspiracy theories has identified dozens of characteristics correlated with conspiracy theory beliefs, much less attention has been paid to understanding the generalized predisposition towards interpreting events and circumstances as the product of supposed conspiracies. Using a unique national survey of 2015 U.S. adults from October 2020, we investigate the relationship between this predisposition—conspiracy thinking—and 34 different psychological, political, and social correlates. Using conditional inference tree modeling—a machine learning-based approach designed to facilitate prediction using a flexible modeling methodology—we identify the characteristics that are most useful for orienting individuals along the conspiracy thinking continuum, including (but not limited to): anomie, Manicheanism, support for political violence, a tendency to share false information online, populism, narcissism, and psychopathy. Altogether, psychological characteristics are much more useful in predicting conspiracy thinking than are political and social characteristics, though even our robust set of correlates only partially accounts for variance in conspiracy thinking.