Hello everyone,

I am looking for some help using Mplus trying to do a longitudinal mediation. My study involved taking pre-post measurements of several variables as part of an intervention that had two groups; a control group and a treatment group.

I need the mediator and outcome variables to be change scores and I am uncertain how to tell Mplus to interpret the mediator and outcome variables as change scores in the Define category . My model has the correct syntax to create the change scores , I just don’t know how to tell Mplus to use the change score variable as the mediator and outcome variable.

I have missing data for the post measurements as several people dropped off during the study. However I am doing an intent-to-treat analysis. Do I need to add anything on the syntax to account for this? I am already specifying that values with -999 are missing and I am also using the ML estimator.

Do I need to rename it somehow?

Any help is much appreciated.

DATA: FILE IS Organized Full Sheet.dat;

VARIABLE:

NAMES ARE

ActiveC !ActiveC is a binary categorical independent variable

CIBS CIBSA

CRBS CRBSA

GSI GSIA

PST PSTA

PSDI PSDIA

N NA

ND NDA

Met MetA

NGSE NGSEA !New General Self Efficacy Scale; complete in NGSA and

DIET DIETA !about 50% on DIETA.

SES SEA

NEval NEvalA

GloH GloHA

Mas MasA

Enj EnjA

Sys SysA

Dia DiaA

BPM BPMA

BFM BFMA

Wght WghtA; !Weight variable; complete in Wght and about 47% in WghtA.

usevariables NGSE NGSEA Wght WghtA X M Y;

missing = all ;

DEFINE:

X = ActiveC;

M = NGSA – NGS;

Y = WghtA – Wght;

ANALYSIS:

TYPE = GENERAL;

ESTIMATOR = ML;

BOOTSTRAP = 10000;

MODEL:

NGSEA ON NGSE@1;

NGSEA@0;

;

delta by NGSEA@1;

delta*;

;

delta with NGSE;

Wghta ON Wght@1;

Wghta@0;

;

delta by Wghta@1;

delta*;

;

delta with Wght;

Y ON M ;

Y ON X ;

M ON X ;

MODEL CONSTRAINT:

NEW ;

a1b1 = a1*b1;

TOTAL = a1*b1 + cdash;

OUTPUT:

STAND CINT ;