For my SEM-analyses, I applied cluster sampling because of non-independence of the participating adolescents (i.e., in most cases adolescents were part of the same family). Consequently, I tried to ascertain no differences between these clustered adolescents (i.e., the older adolescents versus the younger adolescents). However, is it possible to do a DIFF-test while using cluster sampling? In other words: Can I do a DIFF-test for non-independent groups?
You should not do a multiple group analysis when the groups are not independent. You say you applied cluster sampling. What do you mean by that? If you want to take the non-independence into account, you can use TYPE=COMPLEX; or TYPE=TWOLEVEL;