﻿figure it out - a statistical consultancy from the Institute of Work Psychology, University of Sheffield

## Mplus code for mediation, moderation, and moderated mediation models

Model 70 (latent variable version): 1 or more mediators, in parallel if multiple (example uses 1), 2 moderators, both moderating the Mediator- DV path with all 2-way and 3-way interactions, 1 of which also moderates the IV-Mediator path

Example Variables: 1 latent predictor X measured by 4 observed variables X1-X4, 1 latent mediator M measured by 4 observed variables M1-M4, 2 latent moderators W and V, each measured by sets of 4 observed variables W1-W4 and V1-V4 respectively, 1 latent outcome Y measured by 4 observed variables Y1-Y4

Preliminary notes:

The code below assumes that

• The latent IV (factor X) is measured by continuous observed variables X1-X4.
• Any latent moderator(s) (factors W, V, Q, Z) are measured by continuous observed variables W1-W4, Z1-Z4, V1-V4, Q1-Q4 respectively.
• Any latent mediator(s) (factor M, or factors M1, M2, etc.) are measured by continuous observed variables M1-M4 or M1_1-M1-4, M2_1-M2_4 respectively.
• The latent outcome Y is measured by continuous observed variables Y1-Y4.

Model Diagram (factor indicator variables omitted for space/clarity reasons):

Statistical Diagram (factor indicator variables omitted for space/clarity reasons):

Model Equation(s):

Y = b0 + b1M + b2W + b3V + b4MW + b5MV + b6WV + b7MWV + c'X
M = a0 + a1X + a2W + a3XW

Algebra to calculate indirect and/or conditional effects by writing model as Y = a + bX:

Y = b0 + b1M + b2W + b3V + b4MW + b5MV + b6WV + b7MWV + c'X
M = a0 + a1X + a2W + a3XW

Hence... substituting in equation for M

Y = b0 + b1(a0 + a1X + a2W + a3XW) + b2W + b3V + b4(a0 + a1X + a2W + a3XW)W + b5(a0 + a1X + a2W + a3XW)V + b6WV + b7(a0 + a1X + a2W + a3XW)WV + c'X

Hence... multiplying out brackets

Y = b0 + a0b1 + a1b1X + a2b1W + a3b1XW + b2W + b3V + a0b4W + a1b4XW + a2b4WW + a3b4XWW + a0b5V + a1b5XV + a2b5WV + a3b5XWV + b6WV + a0b7WV + a1b7XWV + a2b7WWV + a3b7XWWV + c'X

Hence... grouping terms into form Y = a + bX

Y = (b0 + a0b1 + a2b1W + b2W + b3V + a0b4W + a2b4WW + a0b5V + a2b5WV + b6WV + a0b7WV + a2b7WWV) + (a1b1 + a3b1W + a1b4W + a3b4WW + a1b5V + a3b5WV + a1b7WV + a3b7WWV + c')X

Hence...

One indirect effect(s) of X on Y, conditional on W, V:

a1b1 + a3b1W + a1b4W + a3b4WW + a1b5V + a3b5WV + a1b7WV + a3b7WWV = (a1 + a3W)(b1 + b4W + b5V + b7WV)

One direct effect of X on Y:

c'

Mplus code for the model:

! Latent predictor variable X measured by X1-X4
! Latent mediator M measured by 4 observed variables M1-M4
! Latent moderators W and V, each measured by sets of 4 observed variables W1-W4 and V1-V4 respectively
! Latent outcome variable Y measured by Y1-Y4

USEVARIABLES = X1 X2 X3 X4 M1 M2 M3 M4
W1 W2 W3 W4 V1 V2 V3 V4
Y1 Y2 Y3 Y4;

ANALYSIS:
TYPE = GENERAL RANDOM;
ESTIMATOR = ML;
ALGORITHM = INTEGRATION;

! In model statement first state measurement model
! Then create any latent interactions required
! Then state structural model naming each path and intercept using parentheses

MODEL:

! Measurement model
! This makes these factors standardised
X BY X1 X2 X3 X4;
M BY M1 M2 M3 M4;
W BY W1* W2 W3 W4;
V BY V1* V2 V3 V4;
Y BY Y1 Y2 Y3 Y4;

W@1;   V@1;

! Create latent interactions
MW | M XWITH W;
MV | M XWITH V;
XW | X XWITH W;
WV | W XWITH V;
MWV | M XWITH WV;

! Fit structural model and name parameters
! Note that intercepts of M, Y are fixed = 0 since they are latent vars
! so no code to state and name them as parameters
Y ON M (b1);
Y ON W (b2);
Y ON V (b3);
Y ON MW (b4);
Y ON MV (b5);
Y ON WV (b6);
Y ON MWV (b7);

Y ON X(cdash);

M ON X (a1);
M ON W (a2);
M ON XW (a3);

! Use model constraint subcommand to test conditional indirect effects
! You need to pick low, medium and high moderator values for W, V
! for example, of 1 SD below mean, mean, 1 SD above mean

! 2 moderators, 3 values for each, gives 9 combinations
! arbitrary naming convention for conditional indirect and total effects used below:
! MEV_LOQ = medium value of V and low value of Q, etc.

MODEL CONSTRAINT:
NEW(LOW_W MED_W HIGH_W LOW_V MED_V HIGH_V
ILOW_LOV IMEW_LOV IHIW_LOV ILOW_MEV IMEW_MEV IHIW_MEV
ILOW_HIV IMEW_HIV IHIW_HIV
TLOW_LOV TMEW_LOV THIW_LOV TLOW_MEV TMEW_MEV THIW_MEV
TLOW_HIV TMEW_HIV THIW_HIV);

LOW_W = -1;   ! -1 SD below mean value of W
MED_W = 0;   ! mean value of W
HIGH_W = 1;   ! +1 SD above mean value of W

LOW_V = -1;   ! -1 SD below mean value of V
MED_V = 0;   ! mean value of V
HIGH_V = 1;   ! +1 SD above mean value of V

! Calc conditional indirect effects for each combination of moderator values

ILOW_LOV = a1*b1 + a3*b1*LOW_W + a1*b4*LOW_W + a3*b4*LOW_W*LOW_W +
a1*b5*LOW_V + a3*b5*LOW_W*LOW_V + a1*b7*LOW_W*LOW_V +
a3*b7*LOW_W*LOW_W*LOW_V;
IMEW_LOV = a1*b1 + a3*b1*MED_W + a1*b4*MED_W + a3*b4*MED_W*MED_W +
a1*b5*LOW_V + a3*b5*MED_W*LOW_V + a1*b7*MED_W*LOW_V +
a3*b7*MED_W*MED_W*LOW_V;
IHIW_LOV = a1*b1 + a3*b1*HIGH_W + a1*b4*HIGH_W + a3*b4*HIGH_W*HIGH_W +
a1*b5*LOW_V + a3*b5*HIGH_W*LOW_V + a1*b7*HIGH_W*LOW_V +
a3*b7*HIGH_W*HIGH_W*LOW_V;

ILOW_MEV = a1*b1 + a3*b1*LOW_W + a1*b4*LOW_W + a3*b4*LOW_W*LOW_W +
a1*b5*MED_V + a3*b5*LOW_W*MED_V + a1*b7*LOW_W*MED_V +
a3*b7*LOW_W*LOW_W*MED_V;
IMEW_MEV = a1*b1 + a3*b1*MED_W + a1*b4*MED_W + a3*b4*MED_W*MED_W +
a1*b5*MED_V + a3*b5*MED_W*MED_V + a1*b7*MED_W*MED_V +
a3*b7*MED_W*MED_W*MED_V;
IHIW_MEV = a1*b1 + a3*b1*HIGH_W + a1*b4*HIGH_W + a3*b4*HIGH_W*HIGH_W +
a1*b5*MED_V + a3*b5*HIGH_W*MED_V + a1*b7*HIGH_W*MED_V +
a3*b7*HIGH_W*HIGH_W*MED_V;

ILOW_HIV = a1*b1 + a3*b1*LOW_W + a1*b4*LOW_W + a3*b4*LOW_W*LOW_W +
a1*b5*HIGH_V + a3*b5*LOW_W*HIGH_V + a1*b7*LOW_W*HIGH_V +
a3*b7*LOW_W*LOW_W*HIGH_V;
IMEW_HIV = a1*b1 + a3*b1*MED_W + a1*b4*MED_W + a3*b4*MED_W*MED_W +
a1*b5*HIGH_V + a3*b5*MED_W*HIGH_V + a1*b7*MED_W*HIGH_V +
a3*b7*MED_W*MED_W*HIGH_V;
IHIW_HIV = a1*b1 + a3*b1*HIGH_W + a1*b4*HIGH_W + a3*b4*HIGH_W*HIGH_W +
a1*b5*HIGH_V + a3*b5*HIGH_W*HIGH_V + a1*b7*HIGH_W*HIGH_V +
a3*b7*HIGH_W*HIGH_W*HIGH_V;

! Calc conditional total effects for each combination of moderator values

TLOW_LOV = ILOW_LOV + cdash;
TMEW_LOV = IMEW_LOV + cdash;
THIW_LOV = IHIW_LOV + cdash;

TLOW_MEV = ILOW_MEV + cdash;
TMEW_MEV = IMEW_MEV + cdash;
THIW_MEV = IHIW_MEV + cdash;

TLOW_HIV = ILOW_HIV + cdash;
TMEW_HIV = IMEW_HIV + cdash;
THIW_HIV = IHIW_HIV + cdash;

! Use loop plot to plot conditional indirect effect of X on Y for each combination of low, med, high moderator values
! Could be edited to show conditional direct or conditional total effects instead
! NOTE - values from -3 to 3 in LOOP() statement since
! X is factor with mean set at default of 0

PLOT(PLOW_LOV PMEW_LOV PHIW_LOV PLOW_MEV PMEW_MEV PHIW_MEV
PLOW_HIV PMEW_HIV PHIW_HIV);

LOOP(XVAL,-3,3,0.1);

PLOW_LOV = ILOW_LOV*XVAL;
PMEW_LOV = IMEW_LOV*XVAL;
PHIW_LOV = IHIW_LOV*XVAL;

PLOW_MEV = ILOW_MEV*XVAL;
PMEW_MEV = IMEW_MEV*XVAL;
PHIW_MEV = IHIW_MEV*XVAL;

PLOW_HIV = ILOW_HIV*XVAL;
PMEW_HIV = IMEW_HIV*XVAL;
PHIW_HIV = IHIW_HIV*XVAL;

PLOT:
TYPE = plot2;

OUTPUT:
CINT;