www.cmm.bristol.ac.uk/forum

Yes, the discrepancy in standard errors of the random effects between runmlwin and mixed could be due to differences in how each command handles survey weights at different levels. Different software and commands may apply weights differently to account for clustering, stratification, or unequal ...