Calculate the autocorrelation for the residuals from the logistic growth model fit to the U.S. population data in Section 17.4.2. Recalling the discussion of autocorrelated errors in linear regression (in Chapter 16), does autocorrelation appear to be a serious problem here?

Exercise 17.9 Using nonlinear least squares, refit the logistic growth model to the U.S. population data (given in Table 17.2) assuming multiplicative rather than additive errors:

Yi ¼ β1 1 þ expðβ2 þ β3XiÞ

εi Which form of the model appears more adequate for these data?