LGCM convergence problem PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
Message/Author
 Joyce posted on Friday, December 09, 2016 - 9:49 am
Hello,

I am a graduate student trying to run a Latent Growth Curve Model. I have 5 different time points for a scale, but the participants are different for each time point.

I keep running into the error message: no convergence. number of iterations exceeded.

I have increased the iterations, and have tried different time scores as well.

A recommendation I have seen in reading the manual is to use preliminary parameter estimates as starting values. I am having a hard time figuring out what that means and what time scores to try.

VARIABLE:
NAMES ARE ms_t1 ms_t2 ms_t3 ms_t4 ms_t5 gender;
USEVARIABLES ARE ms_t1 ms_t2 ms_t3 ms_t4 ms_t5;
MISSING ARE all (-99);

ANALYSIS:
COVERAGE=0;
ITERATIONS=1,000,000,000,000,000;

MODEL:
I S | ms_t1@0 ms_t2@1 ms_t3@2 ms_t4@3 ms_t5@4;

OUTPUT: SAMPSTAT STANDARDIZED;

Any advice on what syntax I can use to estimate model parts separately to obtain appropriate starting values?

Thank you,
Joyce
 Bengt O. Muthen posted on Friday, December 09, 2016 - 11:37 am
Growth modeling is for when the same people are measured at several time points. With different people you cannot study growth. Instead you have 5 different groups, so independent observations across groups - hence zero correlations across time and zero coverage across time.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: