The growth curves of young bulls tested at performance-test stations were spread out into genetic, animal?s permanent environment, and residual components. Heritability of body weight was on a moderate level and quite steady. The tested statistical models showed only small differences in describing the variability and population genetic parameters. The number of parameters in the random regression function is more important than the type of function. For traits such as live weight, in which the different stages are highly correlated, it is suitable to use Linear Spline function for random effects. For describing the growths of bulls, random regression functions with more than three parameters could be recommended. Editing the database plays a role in estimating genetic parameters.