Congenital biliary atresia (CBA) is a disorder of early infancy that needs early and accurate diagnosis because the curability may be increased if surgical therapy (hepatic porto-enterostomy) is performed early in the disease course of this condition. But immediate differentiation of CBA from neonatal hepatitis (NH) is very difficult, and even special tests for differential diagnosis are of limited value because of some hazards, including the length of time required, technical difficulties, and so on. The data on 14 patients with CBA and 20 with NH admitted to Tenri Hospital were analyzed by two statistical methods (discriminant analysis and probit analysis) and then scores combining several simple variables were calculated. We tried to differentiate these two conditions by the scoring systems mentioned above, and checked the results of these methods against cases in other hospitals (CBA 6 cases and NH 15 cases). Our scoring systems presented here showed higher discriminant power than any single variable test, including the 131I-BSP excretion test, and can be used where there is no time for differential diagnosis.