The effect of uncertainty communication on public trust depends on belief-evidence consistency

Not everyone perceives the communication of uncertainty about scientific evidence as equally trustworthy. We examine whether these differences in trust can be explained by people's prior beliefs-specifically, the consistency between those beliefs and the evidence presented. We hypothesize that the more consistent the evidence is with people's prior beliefs, the less people will trust both the information and its source when uncertainty is communicated compared with when it is not communicated. Conversely, the less consistent the evidence is with people's prior beliefs, the more people will trust the information and its source when uncertainty is communicated. We tested this crossover interaction in two preregistered online experiments on COVID-19 vaccines (study 1: n = 600) and changes in extreme weather events (study 2: n = 1,001). In both studies, participants rated their prior beliefs on these topics before being randomly assigned to read one of two contrasting evidence conditions, with uncertainty either communicated or not in each condition. The belief-evidence consistency was operationalized by linking people's prior belief scores to the contrasting evidence conditions. We found evidence for our hypothesis for both topics and across people with opposing beliefs. Our results show that people's trust response to the communication of uncertainty depends on how consistent people's beliefs are with the evidence communicated. Furthermore, the effects of uncertainty communication on trust appeared to be independent from its effects on people's subsequent beliefs, with no evidence of polarization in either trust or belief. Implications are discussed.