Dear Folks, I was wondering please if splm::spgm accepts factors amongst the regressors or is it that all the independent variables have to be numeric?? splm::spml seems to accept them just fine - but they appear to be a problem with spgm... With many thanks as always, Stuart Reece.