Mediation in SEM with two independant variables

nenu

Hi !

I watn to run my first SEM, and I have a model with latent variables and a mediation.

This is how I done for one independent variable (E_1)

M =~ M1 +M2 +M3 +M4+M5+M6

E_1 =~ QE1+QE2+QE3+QE4+QE5+QE6+QE7+QE8

T_2 =~ QT1+QT2+QT3+QT4+QT5

# regression

M ~ alpha * E_1

VD~ beta*M + direct* E_1

indirect :=alpha*beta

total :=indirect + direct

proportion := indirect/total

But how can I do for the add the other one (T_2)?

Thanks for you help!

EJ

Hmm this is really a question of how you specify models in lavaan. So I would look online for lavaan documentation first. We are in the process of hiring team members with SEM expertise (our previous expert accepted a job as an assistant professor) and then we can address your question -- but it seems to me it is more of a lavaan question than a JASP question, so the lavaan community should be able to help you out more efficiently. Could you post the answer here once you have figured it out?

Cheers,

E.J.

nenu

Hi EJ

Thanks! Indeed I got an answer, and it was quite simple! 

Here is what I was told  :

------------

Hi,

You would just add T_2 into both regressions as well. So in M ~ alpha * E_1 + indirect_something * T2 and

VD~ beta*M + direct* E_1 + direct_something * t2.

And then add those as a separate indirect effect and to your total.

-------

I also have another question, is it possible to automatically make the diagram path on JASP? I did not find an answer... 

Thanks

EJ

We don't produce the path diagrams. That certainly would be very nice to add.

Cheers,

E.J.