[...] In a survey of 1,044 Australians, we found that endorsement of six pre-existing conspiracy theories ranged from 10% to 19%, and that 7% of participants endorsed two that were clearly contradictory together. In addition, 10% of participants endorsed a highly bizarre raccoon army conspiracy theory that we invented, which we interpreted as evidence of insincerity. Endorsement of this conspiracy theory was a strong predictor of endorsing the pre-existing conspiracy theories, including the two contradictory ones together. Moreover, at debriefing, 13% of participants reported that they had responded insincerely in the survey, and this response was also a strong predictor of endorsing the pre-existing conspiracy theories, including the two contradictory ones together. Overall, our results suggest that surveys can produce inflated estimates of the prevalence of belief in conspiracy theories, which raises serious questions about how research on conspiracy theories—and misinformation more broadly—has frequently equated endorsement with belief.