On Feb 23, 2011, at 21:38 , Ben Bolker wrote: > > Potentially, but I am personally much more interested in enabling > drop1(), which seems to be a much more legitimate tool for testing terms > in models than step(), which is so easy to abuse ...
Yes, although repeated use of drop1() easily leads to backwards elimination.... However, I have a different point: To make drop1() a better generic, I suspect that something needs to be done about the test = c("none", "Chisq") bit. It would be nice if the list of possible tests could vary according to model type, e.g. by doing all tests via anova(model1,model2,...). I'm not quite up to figuring out how complicated that would be. -- Peter Dalgaard Center for Statistics, Copenhagen Business School Solbjerg Plads 3, 2000 Frederiksberg, Denmark Phone: (+45)38153501 Email: pd....@cbs.dk Priv: pda...@gmail.com ______________________________________________ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel