A model design of path analysis
Message/Author
 cchien posted on Friday, July 23, 2010 - 9:20 am
Hi, I have a question about the path analysis with Mplus. In example 3.11 of Mplus user's guide, there are two variables y1 and y2 fitted by x1, x2, and x3, and then x2, y1, and y2 are defined as independent variables to fit the last variable y3. I am wondering, if there is no variable y3, is this path analysis still available? Meanwhile, the analysis is just stopped by y1 and y2, which are defined as the terminal outcome variables. I tried this program by withdrawing the last line "y3 on y1 y2 x2" in model statement, and still obtained an output. However, I am wondering whether this structure is reasonable to do path analysis. There is no obvious information about whether this model design is available or unavailable while searching on the Internet, so hope I can get some suggestions from here. Any comment is appreciated. Thank you.
 Linda K. Muthen posted on Friday, July 23, 2010 - 11:36 am
When you have a set of observed variables, the paths that are estimated or not estimated are determined by theory. Having a path model without y3 is possible.
 Annie Desrosiers posted on Thursday, January 12, 2012 - 9:29 am
Hi, It is possible to have 3 dependent variables with 1 independent variable in a Path model ?

y1 y2 y3 on x1;

Thank you
 Linda K. Muthen posted on Thursday, January 12, 2012 - 9:40 am
Yes.
 Annie Desrosiers posted on Thursday, January 12, 2012 - 9:47 am
But I always have a error message :

THE MODEL ESTIMATION TERMINATED NORMALLY

THE STANDARD ERRORS OF THE MODEL PARAMETER ESTIMATES MAY NOT BE
TRUSTWORTHY FOR SOME PARAMETERS DUE TO A NON-POSITIVE DEFINITE
FIRST-ORDER DERIVATIVE PRODUCT MATRIX. THIS MAY BE DUE TO THE STARTING
VALUES BUT MAY ALSO BE AN INDICATION OF MODEL NONIDENTIFICATION. THE
CONDITION NUMBER IS -0.919D-17. PROBLEM INVOLVING PARAMETER 9.

Is it because my third dependent variable is the interaction of the other two ??

Thank you again
 Linda K. Muthen posted on Thursday, January 12, 2012 - 10:49 am