I can't think of an example where R does not work better than SAS except for a few cases of mixed effects regression models and for processing enormous datasets when the R user does not want to learn about the latest R tools for large datasets. I quit using SAS in 1991 (in favor of S-Plus and transitioned to R around 2000) and have never looked back. Lately what has really made R powerful is its ability to interface with other languages and especially the way it works in a reproducible analysis/dynamic report document context. Frank -- Frank E Harrell Jr Professor and Chairman School of Medicine Department of Biostatistics Vanderbilt University