Greetings fellow R-users, Does anyone know of a way to fit all possible subsets of a global model under glm? I'm looking for something similar to "regsubsets" (from package leaps), but that can handle logistic (binomial) regression. I've searched the archives, books, etc., but have as of yet been unsuccessful. Thanks in advance for any help/advice. P.s., I'm fairly new to R, so my apologies if this problem is a simple one to answer. Todd D. Steury Dept. of Biology Campus Box 1229, 1 Brookings Dr. Washington University St. Louis, MO 63130