Re: [R] model selection with step()

Karen, Look at the help for the drop1() function. ?drop1 There you will see, "The hierarchy is respected when considering terms to be added or dropped: all main effects contained in a second-order interaction must remain, and so on." So, for fit2, the step() function will only consider dropp

[R] model selection with step()

I am using the step() function to select a model using backward elimination, with AIC as the selection criterion. The full regression model contains three predictors, plus all the second order terms and two-way interactions. The full model is fit via lm() using two different model formulae. One