The base rate neglect effect is a stable phenomenon which can be observed in classification learning and in other contexts of human judgment and decision making. An experiment by M. A. Gluck and G. H. Bower (1988) and their explanation of the base rate neglect effect with the delta rule explanation is described. Their results can also be accounted for by two variants of Bayes' model, the base rate equalizing hypothesis and the single symptom interpretation hypothesis. The current study aimed at comparing the three models regarding their explanatory power of the base rate neglect effect. 60 subjects saw a combination of six symptoms and were asked to predict the correct disease. At the end of the experiment, they estimated the probabilities of each disease in the presence of certain symptoms. The pattern of results are best accounted for by the single symptom interpretation hypothesis and not by the two other models.