﻿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 40 (latent variable version): 1 or more mediators, in parallel if multiple (example uses 1), 3 moderators, 1 moderating both the IV- Mediator path and the direct IV-DV path, 2 moderating the Mediator-DV 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, 3 latent moderators W, V, and Q, each measured by sets of 4 observed variables W1-W4, V1-V4, and Q1-Q4 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 + b2V + b3Q + b4MV + b5MQ + c1'X + c2'W + c3'XW
M = a0 + a1X + a2W + a3XW

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

Y = b0 + b1M + b2V + b3Q + b4MV + b5MQ + c1'X + c2'W + c3'XW
M = a0 + a1X + a2W + a3XW

Hence... substituting in equation for M

Y = b0 + b1(a0 + a1X + a2W + a3XW) + b2V + b3Q + b4(a0 + a1X + a2W + a3XW)V + b5(a0 + a1X + a2W + a3XW)Q + c1'X + c2'W + c3'XW

Hence... multiplying out brackets

Y = b0 + a0b1 + a1b1X + a2b1W + a3b1XW + b2V + b3Q + a0b4V + a1b4XV + a2b4WV + a3b4XWV + a0b5Q + a1b5XQ + a2b5WQ + a3b5XWQ + c1'X + c2'W + c3'XW

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

Y = (b0 + a0b1 + a2b1W + b2V + b3Q + a0b4V + a2b4WV + a0b5Q + a2b5WQ + c2'W) + (a1b1 + a3b1W + a1b4V + a3b4WV + a1b5Q + a3b5WQ + c1' + c3'W)X

Hence...

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

a1b1 + a3b1W + a1b4V + a3b4WV + a1b5Q + a3b5WQ = (a1 + a3W)(b1 + b4V + b5Q)

One direct effect of X on Y, conditional on W:

c1' + c3'W

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, V, and Q, each measured by sets of 4 observed variables W1-W4, V1-V4, and Q1-Q4 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 Q1 Q2 Q3 Q4
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;
Q BY Q1* Q2 Q3 Q4;
Y BY Y1 Y2 Y3 Y4;

W@1;   V@1;   Q@1;

! Create latent interactions
MV | M XWITH V;
MQ | M XWITH Q;
XW | X XWITH W;

! 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 V (b2);
Y ON Q (b3);
Y ON MV (b4);
Y ON MQ (b5);

Y ON X (cdash1);
Y ON W (cdash2);
Y ON XW (cdash3);

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, Q
! for example, of 1 SD below mean, mean, 1 SD above mean

! 3 moderators, 3 values for each, gives 27 combinations
! arbitrary naming convention for conditional indirect and total effects used below:
! HWMVLQ = high value of W, 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 LOW_Q MED_Q HIGH_Q
ILWLVLQ IMWLVLQ IHWLVLQ ILWMVLQ IMWMVLQ IHWMVLQ
ILWHVLQ IMWHVLQ IHWHVLQ
ILWLVMQ IMWLVMQ IHWLVMQ ILWMVMQ IMWMVMQ IHWMVMQ
ILWHVMQ IMWHVMQ IHWHVMQ
ILWLVHQ IMWLVHQ IHWLVHQ ILWMVHQ IMWMVHQ IHWMVHQ
ILWHVHQ IMWHVHQ IHWHVHQ
DIR_LOWW DIR_MEDW DIR_HIW
TLWLVLQ TMWLVLQ THWLVLQ TLWMVLQ TMWMVLQ THWMVLQ
TLWHVLQ TMWHVLQ THWHVLQ
TLWLVMQ TMWLVMQ THWLVMQ TLWMVMQ TMWMVMQ THWMVMQ
TLWHVMQ TMWHVMQ THWHVMQ
TLWLVHQ TMWLVHQ THWLVHQ TLWMVHQ TMWMVHQ THWMVHQ
TLWHVHQ TMWHVHQ THWHVHQ);

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

LOW_Q = -1;   ! -1 SD below mean value of Q
MED_Q = 0;   ! mean value of Q
HIGH_Q = 1;   ! +1 SD above mean value of Q

! Calc conditional indirect effects for each combination of moderator values

ILWLVLQ = a1*b1 + a3*b1*LOW_W + a1*b4*LOW_V + a3*b4*LOW_W*LOW_V +
a1*b5*LOW_Q + a3*b5*LOW_W*LOW_Q;
IMWLVLQ = a1*b1 + a3*b1*MED_W + a1*b4*LOW_V + a3*b4*MED_W*LOW_V +
a1*b5*LOW_Q + a3*b5*MED_W*LOW_Q;
IHWLVLQ = a1*b1 + a3*b1*HIGH_W + a1*b4*LOW_V + a3*b4*HIGH_W*LOW_V +
a1*b5*LOW_Q + a3*b5*HIGH_W*LOW_Q;

ILWMVLQ = a1*b1 + a3*b1*LOW_W + a1*b4*MED_V + a3*b4*LOW_W*MED_V +
a1*b5*LOW_Q + a3*b5*LOW_W*LOW_Q;
IMWMVLQ = a1*b1 + a3*b1*MED_W + a1*b4*MED_V + a3*b4*MED_W*MED_V +
a1*b5*LOW_Q + a3*b5*MED_W*LOW_Q;
IHWMVLQ = a1*b1 + a3*b1*HIGH_W + a1*b4*MED_V + a3*b4*HIGH_W*MED_V +
a1*b5*LOW_Q + a3*b5*HIGH_W*LOW_Q;

ILWHVLQ = a1*b1 + a3*b1*LOW_W + a1*b4*HIGH_V + a3*b4*LOW_W*HIGH_V +
a1*b5*LOW_Q + a3*b5*LOW_W*LOW_Q;
IMWHVLQ = a1*b1 + a3*b1*MED_W + a1*b4*HIGH_V + a3*b4*MED_W*HIGH_V +
a1*b5*LOW_Q + a3*b5*MED_W*LOW_Q;
IHWHVLQ = a1*b1 + a3*b1*HIGH_W + a1*b4*HIGH_V + a3*b4*HIGH_W*HIGH_V +
a1*b5*LOW_Q + a3*b5*HIGH_W*LOW_Q;

ILWLVMQ = a1*b1 + a3*b1*LOW_W + a1*b4*LOW_V + a3*b4*LOW_W*LOW_V +
a1*b5*MED_Q + a3*b5*LOW_W*MED_Q;
IMWLVMQ = a1*b1 + a3*b1*MED_W + a1*b4*LOW_V + a3*b4*MED_W*LOW_V +
a1*b5*MED_Q + a3*b5*MED_W*MED_Q;
IHWLVMQ = a1*b1 + a3*b1*HIGH_W + a1*b4*LOW_V + a3*b4*HIGH_W*LOW_V +
a1*b5*MED_Q + a3*b5*HIGH_W*MED_Q;

ILWMVMQ = a1*b1 + a3*b1*LOW_W + a1*b4*MED_V + a3*b4*LOW_W*MED_V +
a1*b5*MED_Q + a3*b5*LOW_W*MED_Q;
IMWMVMQ = a1*b1 + a3*b1*MED_W + a1*b4*MED_V + a3*b4*MED_W*MED_V +
a1*b5*MED_Q + a3*b5*MED_W*MED_Q;
IHWMVMQ = a1*b1 + a3*b1*HIGH_W + a1*b4*MED_V + a3*b4*HIGH_W*MED_V +
a1*b5*MED_Q + a3*b5*HIGH_W*MED_Q;

ILWHVMQ = a1*b1 + a3*b1*LOW_W + a1*b4*HIGH_V + a3*b4*LOW_W*HIGH_V +
a1*b5*MED_Q + a3*b5*LOW_W*MED_Q;
IMWHVMQ = a1*b1 + a3*b1*MED_W + a1*b4*HIGH_V + a3*b4*MED_W*HIGH_V +
a1*b5*MED_Q + a3*b5*MED_W*MED_Q;
IHWHVMQ = a1*b1 + a3*b1*HIGH_W + a1*b4*HIGH_V + a3*b4*HIGH_W*HIGH_V +
a1*b5*MED_Q + a3*b5*HIGH_W*MED_Q;

ILWLVHQ = a1*b1 + a3*b1*LOW_W + a1*b4*LOW_V + a3*b4*LOW_W*LOW_V +
a1*b5*HIGH_Q + a3*b5*LOW_W*HIGH_Q;
IMWLVHQ = a1*b1 + a3*b1*MED_W + a1*b4*LOW_V + a3*b4*MED_W*LOW_V +
a1*b5*HIGH_Q + a3*b5*MED_W*HIGH_Q;
IHWLVHQ = a1*b1 + a3*b1*HIGH_W + a1*b4*LOW_V + a3*b4*HIGH_W*LOW_V +
a1*b5*HIGH_Q + a3*b5*HIGH_W*HIGH_Q;

ILWMVHQ = a1*b1 + a3*b1*LOW_W + a1*b4*MED_V + a3*b4*LOW_W*MED_V +
a1*b5*HIGH_Q + a3*b5*LOW_W*HIGH_Q;
IMWMVHQ = a1*b1 + a3*b1*MED_W + a1*b4*MED_V + a3*b4*MED_W*MED_V +
a1*b5*HIGH_Q + a3*b5*MED_W*HIGH_Q;
IHWMVHQ = a1*b1 + a3*b1*HIGH_W + a1*b4*MED_V + a3*b4*HIGH_W*MED_V +
a1*b5*HIGH_Q + a3*b5*HIGH_W*HIGH_Q;

ILWHVHQ = a1*b1 + a3*b1*LOW_W + a1*b4*HIGH_V + a3*b4*LOW_W*HIGH_V +
a1*b5*HIGH_Q + a3*b5*LOW_W*HIGH_Q;
IMWHVHQ = a1*b1 + a3*b1*MED_W + a1*b4*HIGH_V + a3*b4*MED_W*HIGH_V +
a1*b5*HIGH_Q + a3*b5*MED_W*HIGH_Q;
IHWHVHQ = a1*b1 + a3*b1*HIGH_W + a1*b4*HIGH_V + a3*b4*HIGH_W*HIGH_V +
a1*b5*HIGH_Q + a3*b5*HIGH_W*HIGH_Q;

! Calc conditional direct effects for each combination of moderator values

DIR_LOWW = cdash1 + cdash3*LOW_W;
DIR_MEDW = cdash1 + cdash3*MED_W;
DIR_HIW = cdash1 + cdash3*HIGH_W;

! Calc conditional total effects for each combination of moderator values

TLWLVLQ = ILWLVLQ + DIR_LOWW;
TMWLVLQ = IMWLVLQ + DIR_MEDW;
THWLVLQ = IHWLVLQ + DIR_HIW;

TLWMVLQ = ILWMVLQ + DIR_LOWW;
TMWMVLQ = IMWMVLQ + DIR_MEDW;
THWMVLQ = IHWMVLQ + DIR_HIW;

TLWHVLQ = ILWHVLQ + DIR_LOWW;
TMWHVLQ = IMWHVLQ + DIR_MEDW;
THWHVLQ = IHWHVLQ + DIR_HIW;

TLWLVMQ = ILWLVMQ + DIR_LOWW;
TMWLVMQ = IMWLVMQ + DIR_MEDW;
THWLVMQ = IHWLVMQ + DIR_HIW;

TLWMVMQ = ILWMVMQ + DIR_LOWW;
TMWMVMQ = IMWMVMQ + DIR_MEDW;
THWMVMQ = IHWMVMQ + DIR_HIW;

TLWHVMQ = ILWHVMQ + DIR_LOWW;
TMWHVMQ = IMWHVMQ + DIR_MEDW;
THWHVMQ = IHWHVMQ + DIR_HIW;

TLWLVHQ = ILWLVHQ + DIR_LOWW;
TMWLVHQ = IMWLVHQ + DIR_MEDW;
THWLVHQ = IHWLVHQ + DIR_HIW;

TLWMVHQ = ILWMVHQ + DIR_LOWW;
TMWMVHQ = IMWMVHQ + DIR_MEDW;
THWMVHQ = IHWMVHQ + DIR_HIW;

TLWHVHQ = ILWHVHQ + DIR_LOWW;
TMWHVHQ = IMWHVHQ + DIR_MEDW;
THWHVHQ = IHWHVHQ + DIR_HIW;

! 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(PLWLVLQ PMWLVLQ PHWLVLQ PLWMVLQ PMWMVLQ PHWMVLQ
PLWHVLQ PMWHVLQ PHWHVLQ
PLWLVMQ PMWLVMQ PHWLVMQ PLWMVMQ PMWMVMQ PHWMVMQ
PLWHVMQ PMWHVMQ PHWHVMQ
PLWLVHQ PMWLVHQ PHWLVHQ PLWMVHQ PMWMVHQ PHWMVHQ
PLWHVHQ PMWHVHQ PHWHVHQ);

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

PLWLVLQ = ILWLVLQ*XVAL;
PMWLVLQ = IMWLVLQ*XVAL;
PHWLVLQ = IHWLVLQ*XVAL;

PLWMVLQ = ILWMVLQ*XVAL;
PMWMVLQ = IMWMVLQ*XVAL;
PHWMVLQ = IHWMVLQ*XVAL;

PLWHVLQ = ILWHVLQ*XVAL;
PMWHVLQ = IMWHVLQ*XVAL;
PHWHVLQ = IHWHVLQ*XVAL;

PLWLVMQ = ILWLVMQ*XVAL;
PMWLVMQ = IMWLVMQ*XVAL;
PHWLVMQ = IHWLVMQ*XVAL;

PLWMVMQ = ILWMVMQ*XVAL;
PMWMVMQ = IMWMVMQ*XVAL;
PHWMVMQ = IHWMVMQ*XVAL;

PLWHVMQ = ILWHVMQ*XVAL;
PMWHVMQ = IMWHVMQ*XVAL;
PHWHVMQ = IHWHVMQ*XVAL;

PLWLVHQ = ILWLVHQ*XVAL;
PMWLVHQ = IMWLVHQ*XVAL;
PHWLVHQ = IHWLVHQ*XVAL;

PLWMVHQ = ILWMVHQ*XVAL;
PMWMVHQ = IMWMVHQ*XVAL;
PHWMVHQ = IHWMVHQ*XVAL;

PLWHVHQ = ILWHVHQ*XVAL;
PMWHVHQ = IMWHVHQ*XVAL;
PHWHVHQ = IHWHVHQ*XVAL;

PLOT:
TYPE = plot2;

OUTPUT:
CINT;