1-1-1 MESM with one slope fixed
Message/Author
 Niki Su posted on Thursday, February 01, 2018 - 11:42 pm
I want to run a 1-1-1 MSEM with one slope fixed and I don't know how to write the script to estimate the indirect effect. Here is my script:

model:

%within%
m on x(aw);
sb|y on m;
sc|y on x;

%between%

m on x(ab);
y on m(bb);
y on x;
[sb](bw);
 Niki Su posted on Friday, February 02, 2018 - 2:01 am
One more question, in my experiment. x is a within-subject variable with 2 levels. Thus x has no between-group variation. Can I estimate the indirecr effect through a model in which x is't claimed as "Within are"?
 Niki Su posted on Friday, February 02, 2018 - 5:38 am
Following the last question, since x only has within-group variation, the mediation effect must occur strictly at the within-group level, and there must be no Between indirect effect though both m and y have between-group and within-group variation. Then the script of the 1-1-1 MSEM can be the following:

within=x;
analysis:type is twolevel random;

model:

%within%
m on x(aw); !The slope is fixed because the correlation of random slope and intercept is 1.

sb|y on m;
sc|y on x;

%between%
!m on x(ab); The path is commented out because x only has within-group variation.

y on m(bb);

!y on x; The path is commented out because x only has within-group variation.

[sb](bw);

model constraint:
new(indw);
indw=aw*bw;

Please tell me whether I am right. Thank you very much.
 Bengt O. Muthen posted on Friday, February 02, 2018 - 4:47 pm
This is correct. There is no added covariance term for the indirect effect because m on x does not have a random slope.