Hi, we have some 24 hour diet data that was collected from a complex sample, with several levels of sampling. As the nutrient data is nowhere near normally distributed we have presented medians (and other percentiles) for the data, and for subgroups. However, when we want to compare subgroups we cannot find a way to compare medians adjusting for the complex sampling. We have been using the svy procedures in STATA for some things, but they do not compare medians. Can anyone help? Peter Peter Herbison Statistician Department of Preventive and Social Medicine Dunedin School of Medicine PO Box 913 Dunedin New Zealand Phone: +64 3 479 7217 Fax: +64 3 479 7298 e-mail: [log in to unmask] %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%