Three-way cross-level interaction ana... PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
Message/Author
 Wim Beyers posted on Tuesday, May 28, 2019 - 12:45 am
(2) If so, would the following syntax allow me to transform the score of the latent factor a SD up and down to test the simple effects? It seems wrong to me, as I don't think the slopes I create to test the simple effects take in consideration the main effect of W and Z on Y, yet I don't know how to include it in the syntax.

MODEL CONSTRAINT: NEW (LO_W HI_W LO_Z HI_Z LOW_LOZ LOW_HIZ HIW_LOZ HIW_HIZ);
LO_W = +1; !I assume that the SD of W is 1, because it was standardized with W@1;
HI_W = -1;
LO_Z = +.6325; !the SD of Z is .6325
HI_Z = -.6325;

LOW_LOZ = b1 + b2*LO_W + b3*LO_Z + b23*LO_W*LO_Z;
LOW_HIZ = b1 + b2*LO_W + b3*HI_Z + b23*LO_W*HI_Z;
HIW_LOZ = b1 + b2*HI_W + b3*LO_Z + b23*HI_W*LO_Z;
HIW_HIZ = b1 + b2*HI_W + b3*HI_Z + b23*HI_W*HI_Z;

Thank you!
 Bengt O. Muthen posted on Wednesday, May 29, 2019 - 4:17 pm
You have

s1 = [s1] + b2*W + b3*cZ + b23*W*cZ + residual

Typically you look for a moderator - say W -and consider the moderator function

(b2 + b23*cZ)*W

in which case the simple slope is

(b2 + b23*cZ)

We request that postings are limited to 1 window. For longer postings and input/output, please send to Support along with your license number.
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: