Some Statistical Methods for Comparison of Growth Curves

Abstract

Tests for comparison of average growth curves for groups of individuals treated differently have been developed. Some of these tests are exact and useful when the sample size is small. An attempt is made to transform the time axis in such a way that average rate of growth is uniform, to the extent possible, with respect to the new time metameter for each group. The differences between groups in the rates of growth so defined are subject to a test. The adequacy of a common transformation for all the groups is also examined. Tests appropriate for large samples have also been derived. Other topics discussed are the factor analytic models for individual growth curves and discrimination of growth curves when observations are taken continuously.