Hello, this question is conceptual as much as anything else: when I run svy: mean var1 var2 var3, the weighted means I get are slightly different than if I ran them all separately: svy: mean var1, then svy: mean var2, and so on. Can someone explain to me what's going on under the hood that causes this discrepancy? It seems as if the number of observations is quite different based on which method I do it, which is obviously also part of this dilemma...