AbstractBiopsy specimens are taken from the heart at regular intervals during the weeks following cardiac transplantation. Three specimens at most have been taken at each session, and classified by the pathologist on a histological four‐point scale of rejection. Since the rejection reaction in the myocardium is uniform neither in intensity nor distribution, there is often disagreement between the findings from a single session; even after three negative specimens there is some doubt as to whether rejection has taken place. In an attempt to quantify this residual uncertainty, we have adapted a procedure of Dawid and Skene1for ‘estimating error rates’ when the true state is not known. The model is shown to provide a good explanation of the observed data, and provides guidance on the number of negative specimens required to give a defined level of certainty that rejection is not taking