|
|
Serial-multilevel mediation |
|
Message/Author |
|
|
Hi, I'm a new user of Mplus I have data from a diary study with 4 measurements and 3 mediators. I want to assess serial-mediation at level 1. I'm trying to extend model "I" from Preacher's website and wanted to ask if the following syntax is correct (or if I should use a different model): VARIABLE: NAMES ARE ID X M1 M2 M3 Y; USEVARIABLES ARE ID X M1 M2 M3 Y; CLUSTER IS ID; MISSING ARE ALL (99); ANALYSIS: TYPE IS TWOLEVEL RANDOM; MODEL: %WITHIN% ! M1 ON X(aw); M2 ON M2(bw); M3 ON M2(cw); Y ON M3 (dw); Y ON X; Y ON M1; Y ON M2; M3 ON M1; M3 ON X; M2 ON X; %BETWEEN% X M1 M2 M3 Y; M1 ON X(ab); M2 ON M1(bb); M3 ON M2 (cb); Y ON M3 (db); Y ON X; Y ON M1; Y ON M2; M3 ON M1; M3 ON X; MODEL CONSTRAINT: NEW(indb indw); indw=aw*bw*cw*dw; indb=ab*bb*cb*db; OUTPUT: CINTERVAL SAMPSTAT STDYX; |
|
|
You are expressing one of the indirect effects, namely the one that goes via all 3 mediators. There are other ones. Note also the typo where you say: M2 ON M2(bw); It should be M2 on M1 as you have in on Between. |
|
Back to top |
|
|