Calibration is an important issue in research investigating the appropriateness of probability judgements. Traditionally, calibration has been studied using normative standards, which suggested a general bias (overconfidence). Instead of the normative approach, which considers the rightness of probability judgements, the support theory focuses on how judgements are made. The authors' aim was to study the effect of knowledge and feedback on calibration. They selected groups differing in the amount of their knowledge and in the frequency of the feedback relating to this knowledge. They found that only the amount, and not the feedback, influenced calibration. The results are interpreted in accordance of the support theory: intensity (the stereotypical image about the group's knowledge) is used as a basis for calibration, while the weight of the information (feedback) is neglected.