A multiphasic approach for describing serial height data of Fels children:A hexaphasic-logistic-additive growth model

In this paper, we reported the results obtained from fitting a new growth m odel to serial height data of 80 Fels children. The model assumed that huma n height growth curves are due to the combined effects of six macroscopic l ogistic growth processes, each reaching the same asymptotic height value. I t was named the Walker and Walker-Hexaphasic-Logistic-Additive (WWHLA) grow th model. An advantage to using this model is that it allowed us to easily fit entire growth curves with 14 biologically interpretable parameters. We tested the model using a computerized nonlinear least squares technique. Th e results showed that the new model worked extremely well. The fits resulte d in high R, R-2, and adjusted R2 values, large F values, relatively low re sidual mean squares, Durbin-Watson statistics that were very close to 2, an d relatively small standard error estimates for the model parameters. In ad dition, the normality and constant variance test passed for more than 95 pe rcent of the children, and the graphs of the residuals essentially showed n o model bias. The new model identified the six growth components or process es in both male and female growth curves. The processes were named accordin g to when they reached their peak height velocity: neonatal, infantile, ear ly-childhood, middle-childhood, late-childhood, and pubertal. Preliminary r esults suggest that the WWHLA model appears to be the best that is currentl y available at this time for describing the human growth curve.