I am trying to figure out how to test for mediation in this more complex model. The fact that I am new to mediation does not make it any easier. However, I would very much like to know, how you would approach a model like the one attached?

All the variables are measured on a 7-point likert scale.

From what I can see, literature generally suggests that SEM is more fitted for testing this somewhat more complex model. However, as I am more familiar with regression I would like to know if this is even doable?

I have currently had success in running the macro for multiple mediation test provided by Preacher and Hayes (2008). However, this only allows med to test for mediation in the first part of the model that is X → M1, M2, M3, M4 → Y1. I do not think that it is possible to test for all the direct and indirect effects in one model by using regression. This is why I am considering splitting up the model and running two additional mediation tests. One where the mediators becomes IV and Y1 becomes a mediator, that is X1, X2, X3, X4 → M → Y2 and then a last one that tests the full mediation between X → Y1 → Y2.

I know it is a long shot since, I have not been able to locate any literature that support splitting up a model in this way. This is why I am afraid that I will lose some effects by doing so.

I hope some of you can enlighten me and push me in the right direction.

Thank you in advance :)