Dear all,

I would be grateful for your ideas on the question what one might consider as an appropriate sample size for level 2 units when conducting a twolevel SEM.

For example, when modeling a twolevel CFA with 5 indicators within and 5 indicators between groups (level 2), does this mean that the standard sample size recommendations for single-level factor analysis (e.g. Tabachnik and Fidell recommend 300 cases per factor) hold for twolevel analysis as well? Or are "things different" for twolevel CFA, e.g. because the level 2 units should have a higher reliability because they represent means of level-1 indictors?

Many thanks for your answers!