These data are the weight losses of subjects randomly assigned to one of three
diets, and these additional covariates sex, initial age, and body mass index.
Is there any evidence from these data that the mean weight loss differs for
the different diets, after accounting for the effect of the covariates? How big
are the difference?