|
|
| Message/Author |
|
|
|
|
Dear Professor Muthen
Here are two questions for moderating syntax for SEM. All are continuous variables. The moderation model is that organizational trust (OT) moderates the relationship between breach (PCBR) and violation (PCVI).
1. Should I use "Analysis: Estimator = MLM;" in my model? Or use "ANALYSIS: TYPE = RANDOM; ALGORITHM = INTEGRATION;"?
2. The interaction term always becomes a separate latent variable in my output.
Following <mplus> on page 71, I write syntax as follows:
TITLE:
SEM three indicator with moderation model
DATA:
FILE =/Users/xi/Documents/voice&silence/new reverse listwise deletion.csv;
VARIABLE:
NAMES = CASE_t1 CASE_t2 sex age otenure ft
ot1 ot2 ot3 ot4 ot5 ot6 ot7
pcb1 pcb2 pcb3 pcb4 pcb5
pcv1 pcv2 pcv3 pcv4;
Usevariables = ot1 ot4 ot7
pcb3 pcb4 pcb5
pcv2 pcv3 pcv4;
ANALYSIS:TYPE = RANDOM;
ALGORITHM = INTEGRATION;
MODEL:
OT BY ot1 ot4 ot7;
PCBR BY pcb3 pcb4 pcb5;
PCVI BY pcv2 pcv3 pcv4;
PCVI ON PCBR;
BROT | PCBR XWITH OT;
PCVI ON BROT;
OUTPUT:
Unlike figure in page 71, the output shown in my example is that the interaction term becomes a separate latent variable.Thanks so much. |
|
|
|
|
You use XWITH so you need Type=Random and we recommend MLR (which is the default).
I don't know which document you refer to when you say page 71. I think you need to send your output to Support to show what you mean by the interaction term becoming a separate latent variable. |
|
|
|
|
| Thank you so much for your reply. The page 71 is for <mplus>. I do not know if I use MLR, how do I modify my syntax? I will send my output to Support. ThanksŁĄ |
|
|
| Back to top |
|
|
