3-way interaction (one latent and two... PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Lemi Baruh posted on Thursday, April 19, 2012 - 5:15 am
Hi,

I have data from a 2 by 2 experiment: Info_Amo (high vs. low) and Pro_gend (female vs. male) are conditions.

Also,
-Two ordinal, Q1A and Q1D as the dependent variable
-One nominal, Q1F_2ka as a mediator.
-Three ordinal variables Q5A, Q5B, Q5C to measure “social comparison”

Is there any way for me to test the three way interaction between social comparison (Latent) and Info_Amo (observed) and Pro_gend (observed)?

I have attempted the following but it said only two-way interactions are available.


Data:
File is Tess045_MPLUS.csv;
Variable:
Names Are
Q1A Q1D Q5A Q5B Q5C Q1F_2ka Info_Amo Pro_Gend Geninfo;
Missing is
Q1A (-1) Q1D (-1) Q5A (-1) Q5B (-1) Q5C (-1) Q1F_2ka (-1);
USEVARIABLES ARE Q1A Q1D Q5A Q5B Q5C Q1F_2ka Info_Amo Pro_Gend Geninfo;
CATEGORICAL = Q1F_2ka;
DEFINE:
GENINFO = Info_Amo*Pro_Gend;;
ANALYSIS: TYPE = RANDOM;
ALGORITHM = INTEGRATION;
MODEL:
F1 BY Q1A Q1D;
F2 BY Q5A Q5B Q5C;
VOINFO | F2 XWITH Info_Amo;
GENVOY | F2 XWITH Pro_Gend;
VOGEINFO | F2 XWITH INFO_Amo XWITH Pro_gend;
Q1F_2ka ON Info_Amo Pro_Gend GENINFO;
F1 ON Q1F_2ka Info_Amo Pro_Gend F2 GENINFO GENVOY VOINFO VOGEINFO;
OUTPUT: TECH1 TECH8;


Thank you very much.
 Linda K. Muthen posted on Thursday, April 19, 2012 - 12:12 pm
There are two things you can do. You can create the interaction between the two observed variables in DEFINE and then use that in XWITH with the latent variable. Or you can use XWITH as follows:

int2 | latent XWITH observed1;
int3 | int2 XWITH observed2;
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: