|
|
 |
| 3-level Monte Carlo analysis of RI-CLPM |
 
|
| Message/Author |
|
|
|
|
Hello, I am using the Monte Carlo function to test random slopes of the cross-lagged coefficients in a Random Intercept Cross-Lagged Panel model. Unfortunately, the whole model doesn't fit into the message here, but the most important parts are the following (I deleted the definitions of the random intercepts and the latent within-level deviation variables):
%WITHIN%
cy2-cy5 pon cy1-cy4*.5;
cx2-cx5 pon cx1-cx4*.5;
cy2-cy5 pon cx1-cx4*.3;
cx2-cx5 pon cy1-cy4*.3;
cx1-cx5 pwith cy1-cy5*.2;
RI_xw*.90; RI_yw*.90;
cx1*1; cy1*1;
cx2-cx5*.66; cy2-cy5*.66;
%BETWEEN%
RI_xb*.10; RI_yb*.10;
[RI_xb*1 RI_yb*1];
The model works up to here, but then I change to Type = twolevel random; and:
r | cy2-cy5 pon cx1-cx4;
plus on the between level:
r*.10; [r*.3);
The idea was to add a small between-level variance to this cross-lagged coefficient.However, I get the following error message:
THE POPULATION COVARIANCE MATRIX THAT YOU GAVE AS INPUT IS NOT POSITIVE DEFINITE AS IT SHOULD BE.
Can you see what I've done wrong?
Best,
Fredrik Falkenström |
|
|
|
|
| We need to see the full output - please send to Support. |
|
|
| Back to top |
|
|
