A modified logistic model applied to human populations

The use of the logistic curve for forecasting human populations [in the United Kingdom] considered by Leach is re-examined. A modification of the use of the logistic curve is suggested which changes the emphasis from fitting a logistic trend to providing a forecast logistic trend. The assumption used by Leach that the variance of the additive disturbance term is constant is replaced by a more realistic supposition that the variance of the proportional disturbance is constant. The forecasting performance of the modified logistic model is shown to be superior to that of the model used by Leach. A reply by Leach (pp. 496-7) and a response by the author (p. 498) are included. (EXCERPT)