|
|
Multiple mediation analysis |
|
Message/Author |
|
|
Hi, I don't have a lot of MPlus experience. I am trying to modify an example program from the manual to run a multiple mediation analysis and could use some assistance. I have time 1 measurements for an independent variable, two mediator variables, and two outcomes. I have time 2 measurements of the two mediators. I have time 3 measurements of the two outcomes. I would like to test the total indirect effect of both time 2 mediators on time 3 outcomes, as well as each of their specific indirect effects after accounting for the other mediator, and I also would like to control for time 1 levels of both mediators and both outcomes. Below is my input file. My questions are: 1. What modifications to the model paragraph would control for time 1 measurement of the mediators and the outcomes? 2. Does the rest of the syntax look correct for the intended analysis? Thanks, Patrick Palmieri data: file is C:\mplus\mediation.dat; format is free; type is individual; nobservations are 625; variable: names are iv t1med1 t1med2 t1dv1 t1dv2 t2med1 t2med2 t3dv1 t3dv2; usevariables iv t1med1 t1med2 t1dv1 t1dv2 t2med1 t2med2 t3dv1 t3dv2; missing are all (-9); analysis: type = general missing H1; estimator = ML; iterations = 50000; bootstrap = 5000; model: t2med1 on iv; t2med2 on iv; t3dv1 on t2med1 t2med2 iv; t3dv2 on t2med1 t2med2 iv; t2med1 with t2med2; t3dv1 with t3dv2; model indirect: t3dv1 IND t2med1 iv; t3dv1 IND t2med2 iv; t3dv2 IND t2med1 iv; t3dv2 IND t2med2 iv; output: sampstat standardized cinterval(bcbootstrap); |
|
|
I think you want to regress t1dv1 t1dv2 on t1med1 t1med2. I don't see that in the MODEL command. That would be: t1dv1 t1dv2 ON t1med1 t1med2; Otherwise, I can't see any problems. The best way to see if you have set the input up correctly is to run the analysis and see if you get what you want. |
|
Back to top |
|
|