“…Analyses of data on the frequency of claims in Pennsylvania, New York, New Jersey, and Florida find heterogeneity of risk within a given physician specialty (Rolph, 1981(Rolph, , 1991Nye and Hofflander, 1988;Venezian, Nye, and Hofflander, 1989;Ellis, Gallup, and McGuire;1990). 2 Sloan and Hassan (1990) show that residual heterogeneity within risk classification groups is reduced but not eliminated when the a priori risk groups are based on more information than just physician specialty. Their study provides the best support to date for the presence of residual heterogeneity in claims.…”