In agricultural research, different ordinal scales of measurement are often used to estimate disease severity. However, the characteristics of the distribution of diseased leaves in the population and the number of classes in a disease scale often affect the accuracy of the resulting estimates of mean disease severity. The purposes of this study are to compare various interval-scale estimates to nearest percent estimates and to further investigate the effects of the number of classes in certain disease scales. A simulation method was employed to execute the study. The criterion for comparison was the mean squared error for each of the different scales used for estimation. The results of this study indicate that, when preparing numeric category scales for rating disease severity, scales with grades of ?7 are preferable. Moreover, linear category scales with sensitivity to low disease severity are preferable to nonlinear category scales for assessing disease severity. For practical application, to take one example, pear scab is one of the most notorious diseases around the world. However, so far no objective criteria for assessing pear scab severity have been identified in the literature. Therefore, real data from the field was used to verify the simulation results. We believe that the results of our study will be helpful in improving the accuracy of disease severity estimates in plant epidemiology and related areas of research.