Path Analysis PreviousNext
Mplus Discussion > Structural Equation Modeling >
 Fei Xing posted on Tuesday, February 27, 2018 - 4:30 pm
MODEL I tested:
e, m, n ON s;
score, m, r, p ON n;
e, r, p ON m;
score, e ON p;
score ON r;
score WITH e@0;
I do not understand why every time the program will automatically consider the correlation between e and score as a parameter. They are endogenous variable.
In addition, I got weird results. Why did I get a negative solution for the parameter "score ON r" while all others are positive?
STDYX Standardization

Estimate S.E. Est./S.E. P-Value

N 0.249 0.069 3.607 0.000
P 0.212 0.058 3.656 0.000
R -0.130 0.058 -2.249 0.025

N 0.360 0.057 6.282 0.000
M 0.356 0.060 5.965 0.000

N 0.287 0.059 4.845 0.000
M 0.333 0.056 5.977 0.000
 Bengt O. Muthen posted on Wednesday, February 28, 2018 - 3:55 pm
Endogenous variables at the end of the chain are given residual covariances as the default because that is often needed. Note that these are not correlations between the variables but correlations between the residuals of the variables.

I have no comment on your negative finding - you may want to post this interpretation question on SEMNET.

Note that you don't have to insert commas in your ON statements.
 Fei Xing posted on Friday, March 02, 2018 - 5:05 pm
Thank you very much for your response, Dr. Muthen. If there is a high and significant residual covariances between endogeneous variables. What does that mean?
 Bengt O. Muthen posted on Saturday, March 03, 2018 - 6:40 am
It means that there are left-out covariates for both outcomes which because they are not included ends up in the residuals. Because the covariates are correlated and are part of both residuals, the residuals are correlated.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message