Mplus Discussion >> Problem in APIM mediation

Topics
Last Day
Last 3 Days
Last Week
Tree View

Edit Profile


Problem in APIM mediation

Mplus Discussion > Structural Equation Modeling >

Message/Author

lluvia Park posted on Thursday, July 21, 2016 - 2:36 pm

Hello,

I'm trying to test mediation in (distinguishable) dyad data

So I'm using APIM mediation model like following (M and F are for male and female data):

model: ZRE_M on NNI_M Sat_M;
ZRE_F on NNI_F Sat_F;
Sat_M on ZRE_M ZRE_F NNI_M NNI_F;
Sat_F on ZRE_M ZRE_F NNI_M NNI_F;
Sat_F with Sat_M;
NNI_M with NNI_F;
ZRE_M with ZRE_F;
Model indirect: Sat_M ind NNI_M;
Sat_M ind NNI_F;
Sat_F ind NNI_M;
Sat_F ind NNI_F;
output: res;
cinterval(bcbootstrap);

and when I run it, I get

"ALL INDIRECT EFFECTS CANNOT BE COMPUTED BECAUSE THE ABSOLUTE VALUE OF AN
EIGENVALUE OF THE REGRESSION MATRIX IS
GREATER THAN OR EQUAL TO 1."

I'm not sure what the problem is.
Is there a way to fix it??

Thank you

Bengt O. Muthen posted on Thursday, July 21, 2016 - 5:25 pm

You have a reciprocal interaction model like

y1 on y2 x1;
y2 on y1 x2;

The SEM literature describes how indirect effects need special care for such models and that these eigenvalues need to be less than 1. The model needs to be changed in some way to avoid this problem.

Bengt O. Muthen posted on Friday, July 22, 2016 - 8:56 am

lluvia Park posted on Saturday, July 23, 2016 - 4:25 am

Thank you!
I'll see how I can change the model