The objective of this work was to estimate minimum sample size (n) of animals to detect significant differences (D) among monthly replicate evaluations of scrotal circumference from animals of the Nellore breed, from 10 to 30 months of age, considered as 21 evaluation conditions or repeated measurements (treatments). The value of n that allows to detect the D among means when each subject is observed for t > 2 treatments, was obtained by means of the Statistical Analysis System - SAS developed program, the t-variate normal distribution (t=21), as a vector of zero mean and S covariate matrix, Hotelling T² statistics and F-distribution with noncentrality parameter of d²D. Different levels of type I error (a), power of the test (1- b), minimum correlation among repeated measurements (r min) and D were considered. The n estimates ranged from 4 to 299, being mostly affected by changes in the D.
F distribution with noncentrality parameter δ2Δ; type I error (α); power of the test (1-β); repeated measurements; simulation
