2-1-1 moderated mediation PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
Message/Author
 Yue Zhu posted on Wednesday, May 06, 2015 - 11:48 pm
Dear Professor,

I am trying to test a model 2-1-1 (X-M-Y) with a level 2 moderator (W) on 2-1. Here is my mplus syntax:

TITLE: 2-1-1 moderated mediation
DATA: FILE is mydata.dat;
DEFINE: XW= (X-3.0) * (W-4.0);
VARIABLE: NAMES ARE X M W Y;
USEVARIABLES ARE X M W Y;
CENTERING IS GRANDMEAN (W);
CENTERING IS GRANDMEAN (X);
CENTERING IS GRANDMEAN (M);

CLUSTER= cluster;

BETWEEN = X W;

ANALYSIS: TYPE = TWOLEVEL RANDOM;

MODEL:
%WITHIN%
M Y;
Y on M(b);

% BETWEEN %
X W XW M Y;
M on X(a1)
W XW (a0);
Y on M (b1)
X W XW;

X WITH W XW;
X WITH XW;

MODEL CONSTRAINT:
NEW (ind_h ind_l);
Ind_h = (a0 +a1* 0.85)*b1;
Ind_l = (a0 - a1*0.85)*b1;

NEW(diff);
Diff=ind_h - ind_l;
OUTPUT:
SAMPSTAT;
CINTERVAL;

My question is
1) When I calculate the Ind_h and Ind_l, should I use b or b1?
2) Do I need to employ 2-1-1 model with random slops (example F in Preacher, Zyphur & Zhang (2010))rather than this tradition MSEM?

Thank you so much!!
 Bengt O. Muthen posted on Thursday, May 07, 2015 - 9:45 am
1. You can't give a label after more than one coefficient like you do below

W XW (a0);

2. Your ind=(a0+a1*...)*b1 expressions are wrong when W is the moderator. See our Mediation page for how to write it.

1) You should use b1

2) No.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: