For a variety of reasons, predictive power is not always a useful criterion of model validity. One alternative is a closed-form analysis of error propogation, but often the exact functional form is too unwieldy to use. A similar analysis is performed here, using a Taylor series approximation to...