Message/Author 

Ali posted on Tuesday, July 01, 2014  2:54 am



Dear Muthens I have specified the following montecalo SEM model : model: f1 by y1@1 y2y8*.7; f2 by y9@1 y10y14*.7; f3 by y15@1 y16y21*.7; f4 by f1@1 f2f3*.7; f5 by y22y24*.7; f6 by y25@1 y26y27*.7; f6 on f5*.5; f4 on f5*.3; f4 on f6*.7; y1y27*.3; f1f3*1; f4*.5; f6*1; f4*1; f5*.5; But I received the following error: *** FATAL ERROR THE POPULATION COVARIANCE MATRIX THAT YOU GAVE AS INPUT IS NOT POSITIVE DEFINITE AS IT SHOULD BE. I'm not sure where I went wrong. I would appreciate your help 


The error message refers to your Model Population command not your Model command. Often it implies that you have a zero variance, perhaps one not mentioned. If that doesn't help, send to Support. 

Ali posted on Tuesday, July 01, 2014  9:32 pm



The Population command is exactly the same as the Model command. I guess I have included everything.Isn't it related to the proportion of the coefficients in the model? I mean should each factor loading squared+the respective residual variance for each obesrved variable add up to 1? 


They don't have to add up to 1. To be able to judge this better, you may want to send the output to support@statmodel.com with your license number. 

Back to top 