Prevalence-Induced Concept Change (PICC) is the tendency for people to change their concepts in response to shifts in the prevalence of corresponding information in the environment (Levari et al., 2018; Devine et al., 2023). For instance, when people see fewer overweight bodies, they shift their concept of an overweight body and start judging some average bodies as overweight (Devine et al., 2022). Some people, particularly older adults, are not susceptible to the PICC effect (Devine et al., 2023). Other studies show that older adults are generally less likely to change their beliefs than younger adults and show overconfidence in their decisions (Prims & Moore, 2023). We suggested that high confidence in the original concepts (before the prevalence shift) may determine the susceptibility to the PICC effect. In a preregistered study (N = 81), we asked participants to categorise fictional characters (Bouba and Kiki) and rate the confidence in their judgements. The prevalence of the target stimuli (Bouba) was stable (50%) across 16 blocks of trials in one group and dropped to 10% after 4 blocks in the other group. Our results show that higher confidence before the prevalence shift is associated with a lower probability of false alarm errors after the prevalence shift. The same association is found between the false alarm errors and the decision boundary parameter, which can also be interpreted as a final confidence in the decision (Lee et al., 2023). Thus, we found the first evidence that confidence may serve as a mediator of the PICC effect.