comment
1. Extending the approaches of this study in the future, we can investigate the higher-order structure be treated as multidimensional and unidimensional approach. In the other words, we can simulate the true model as higher-order model (within or between), but we use the multidimensional model (within or between) and unidimensional model to analysis.
2. Another question I want to extend to investigate is that how bias does the higher-order score obtained by plausible value method for SEM have after the first order scores are analyzed as "multidimensional model", rather than as unidimensional model. It means the score in one dimension have been adjusted by the correlation with other dimension at the first order, but we still treat the higher order score to affect those first order score. I am curious the effect of this strange design.