The use of item response theory models for analyzing the results of judgmental standard setting studies (the Angoff technique) for establishing minimum pass levels is discussed. A comparison of three methods indicates the traditional approach may not be best. A procedure based on generalizability theory is suggested. (GDC)