|
|
Testing Moderation using continuous v... |
|
Message/Author |
|
Lena Jeong posted on Wednesday, August 15, 2018 - 9:25 am
|
|
|
Hi: I am trying to run a SEM model and test moderating effect of a construct with is a continuous variable. I have three constructs V, S, and L in my model and the paths are V->S->L. V influence S and S influence L. I would like to test the moderating effect (M) between V and S, and S and L, simultaneous. Could you please help me how to code this in Mplus? Thank you very much. |
|
|
See the MOD description on page 765 in the UG. In the Model Indirect command, you say using your variable names: L MOD S M(low, high, incr) SM VM V; where low and high give the range of the moderator M, SM is S*M, and VM is V*M. The 2 interactions are created in Define. The outcome L can be latent but the other variables need to be observed. If the others are latent, you can express the effects using Model Constraint in line with our RMA book. |
|
Lena Jeong posted on Wednesday, August 15, 2018 - 10:11 pm
|
|
|
Thank you Prof. Muthen for your response. Yes, All others are latent (V,S,L and M), how can I express the effect using Model constraint? Is there any specific guideline or page I can find in the book? Thank you. |
|
|
See Section 2.6.1 on pages 91 - 93 in the RMA book (3rd printing). |
|
Back to top |
|
|