Unless you have detailed simulations to back up the performance of this method I would avoid it. It violates several statistical principles. Frank Hari wrote> Hello R geeks, > > Waiting for an reply. > > Thanks, > Hari----- Frank Harrell Department of Biostatistics, Vanderbilt University -- View this message in context: http://r.789695.n4.nabble.com/stepwise-discriminant-analysis-using-wilks-lambda-tp4668817p4669235.html Sent from the R help mailing list archive at Nabble.com.