Question: As was stated in, it is difficult to detect nonnormality or heteroscedasticity with only eight data points. Recall, however, the statement of Ruppert et al. (1989) in that suggests that it is likely that we will not have homoscedastic nonlinear regression data that would be fit by the Michaelis-Menten model. What might a practitioner do to try to ensure homoscedasticity when there is a very small sample?