Scientific Publications by FDA Staff

Biometrical J 1996;38(7):767-77

On Choice of Subjects and Replicates in Estimating Among and Within Subject Variation

Wong WK, Lachenbruch PA, Clements PJ

We consider the problem of choosing the number of replicates and number of subjects in a components of variance problem which optimizes various criteria. The case study here involves patients suffering from systemic sclerosis (Scleroderma), a form of rheumatic disease that is potentially disabling. Under the physical constraints imposed on the study, we find that using 2 or 3 replicates with as many patients as possible is the optimal strategy for several criteria.