Is it appropiate to run pls sem individually on each independent variable to see the impact on 2 Dvs and 2 serial mediat

[Ayla Wajahat]

I am studying the impact of teaching styles on student satisfaction and learning. The study has six teaching styles, two serial mediators, and 2 DVs.

[jmbecker]

Usually, I would include both DVs into one model, if there are no other data problems that would complicate such an analysis. The model can build on the larger nomological network to estimate the composite scores and relations, which should make it more reliable.
However, if it is just the same predictors and mediators, the differences between separate and combined model are usually small if your sample sizes are large enough and if there is no relation between the DVs.

[Ayla Wajahat]

My model has six different datasets based on each teaching style. The data is in dyad forms. The data on teaching style and one mediator were collected from teachers, and data on DVs and one mediator were taken from students. For each style, I have run the separate pls-sem. Is it appropriate?