Not sure how to interpret standard errors produced by model.tables(object, type="effects", se=T) For example, simplest possible case is single stratum anova with two equally replicated treatments (n observations each). The effects displayed are plus and minus half the difference between the two treatment means, and the SE is given as sqrt(RMS/n), where RMS = residual mean square. But the usual calculation would give the SE as this divided by sqrt(2). Can't see this question raised previously in the archives, so perhaps I am missing something basic here. I.White Ashworth Laboratories, West Mains Road Edinburgh EH9 3JT E-mail: i.m.s.white at ed.ac.uk -- The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.