Predicting adult stature: a comparison of methodologies

Three methods of predicting a child's adult height are in common use: Bayley and Pinneau (1952), Tanner et al. (1975) and Roche et al. (1975). The relative accuracies of these methods have been assessed using growth records (22 male, 24 female) from the Child Research Council, Denver, Colorado. A series of six ages from five years to mid-adolescence was examined. Testing all methods and ages within each sex with a multivariate design yielded significant differences in the methods' accuracies, but inspection of the data by age disclosed that the major discrimination occurred during adolescence, not childhood. Results also indicated that the method of skeletal age assessment (either Greulich--Pyle or TW2) is more critical to accurate height prediction than is the choice of prediction method per se. This is because of inter-population differences in the rate and pattern of progress towards maturity and thus indicates the need to compare the children under examination to the most appropriate population standards available.