Unlike previous work showing that people underestimate others’ support for climate action, this study finds systematic pluralistic ignorance in both directions: frequent attitudes and actual behaviours are underestimated but rare ones are overestimated. This can be explained by general cognitive processes.