Abstract
Two models for growth in length and weight early childhood are compared for goodness of fit and other qualities. One model is linear in its parameters, while the second is nonlinear. Both models are fitted to the data for length and weight of each child in a large sample which has fairly complete and reliable records of measurements. The results are used to study issues such as model reliability, efficiency and precision, overall and age-specific goodness of fit, and adequacy of assumptions.