At the risk of seeming churlish, given that it appears to remain very popular, I will mention that the non-parametric statistics book by the late Sidney Siegel first published in 1956 received an absolutely scathing review from I.R. Savage in JASA in 1957, detailing its many mistakes and misconceptions. Unfortunately, I have no idea whether the revision by the late N. John Castellan did much to rectify matters. But I have a notion that its popularity arises as much from its layout, allowing you to find almost instantly the non-parametric test you supposedly need, and from its worked examples, as from any statistical or scientific merit. Nick [log in to unmask] %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%