Andrich, Marais and Sappl (2023) adapted Rasch’s concept of a meta-metre (Wohlk Olsen, 2003; Rao, 1958), which governs the rate of growth in the physiological variable of weight of all individuals of a population, to characterise growth on some intelligence and attainment tests. Within the meta-metre, a single, linear rate of growth of each individual can be estimated. Given the estimate for each individual, comparisons among the mean rates of growth of subpopulations, such as by demographic groups, can be conducted with standard analysis of variance procedures. In such procedures, the variance among individuals absorbs the error variance of the estimate of each individual. Andrich et al. (2023) did not provide estimates of the error variance of the estimates of individual rates of growth. This paper is concerned with directions for estimating this variance which, because of regression effects of both measurement error and growth, is affected by the dependence between successive times of measurement.