My colleagues and I currently are conducting a multi-group EFA with loadings and thresholds constrained across groups. We have unbalanced samples, so we adjusted the country specific survey weights so that the effective sample size per group is equal (say 2500/group). The challenge we’re having is that when we extract factor scores and weights and save them, the weights that are returned are adjusted—seemingly to the observed (group) sample size.
The questions we have are twofold: Are the estimations with the weights adjusted before or postestimation? What is the process by which the adjustment is made, and can Mplus be made to use the weights as provided without being adjusted?
Many thanks and apologies if this has been asked before,
The weights are adjusted before the estimation. The weights within each group are scaled so that the sum of the weights equals the sample size of the group. It is not possible to change that at this point.