Effects of observed correlation structure among tumor types on experimentwise false positive rate in animal carcinogenicity studies.

Abstract

Probabilities P1, P2, ..., Pn of n events E1, E2, ..., En can impose constraints on the probability of a further event E. deFinetti's fundamental theorem of probability characterizes the interval of coherent probability of E. Lad, Dickey, and Rahman (5) extended deFinetti's theorem in a few directions. In this work we will show some applications of these results to the calculation of experimentwise false positive error rate in multiple hypotheses testing procedure.