Logistic regression analysis was applied to data from a population of donkeys to obtain statistical models estimating the risk of equine sarcoid. Two models were constructed and compared. The group model was derived using the prevalence of animals with sarcoids in groups, classified according to the explanatory variables gender, age at first exposure (considered as age at entry to the population), duration of exposure, and gender by duration of exposure as an interaction. The subject model was derived from individual animal data and consisted of the factor gender with the covariates age at first exposure and duration of exposure. Age at first exposure was represented in the model by linear and quadratic terms, and as an interaction with duration of exposure. Both models provided a good approximation to the observed data. The analyses identified young male donkeys with short duration of exposure as being at highest risk. We concluded that risk assessment using the group model could be used effectively in the clinical management of sarcoids in the population of donkeys. Frequent examination of high risk groups might allow early diagnosis, therapeutic intervention and improved animal welfare.