Iím interested in whether there are different subtypes of married couples in terms of variables such as self-reported marital satisfaction, marital discord, etc. (e.g., some couples are in agreement, others arenít) . Iíve tried LCA using the husband and wife self-reports as separate indicators (Perhaps allow residual covariances, as needed?). I realize that individuals are nested within dyads and therefore a multilevel LCA may be more appropriate (Iíve read Henry & Muthťn, 2010, and Chapter 10 in the Mplus Userís Guide). However, given that n = 2 for each cluster MLCA seems problematic. I would appreciate any advice you could give me with this!
You do not have to use multilevel modeling in your cases. The clustering due to couple is taken into account because you have a multivariate model with different variables for both members of the couples.