Plotting a count (negative binomial) ... PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
 Chris Greenwood posted on Monday, December 17, 2018 - 12:35 am

I have run a negative binomial growth model with intercept, slope, quadratic, and cubic growth factors and am estimating the variance of the intercept and slope.

I am trying to plot the expected values in excel (although I understand I can plot within Mplus). I believe that I need to exponentiate the mean growth factors, however I can't replicate the values that Mplus plots.

Although, I can replicate the values, when the variance of all the growth parameters is fixed at 0.

Any help would be greater appreciated.

 Bengt O. Muthen posted on Monday, December 17, 2018 - 4:45 pm
The variances of the growth factor means that you need to use numerical integration over the growth factors to get the estimated values - you cannot just plug in the growth factor means.
 Chris Greenwood posted on Monday, December 17, 2018 - 5:17 pm
Thank you very much for the quick response.

Following this, is numerical integration not used in growth models which specify the data as continuous?

I ask this because I am able to replicate the expected values in a continuous model with estimated growth parameter variances.

Thanks again for clearing this up.
 Bengt O. Muthen posted on Tuesday, December 18, 2018 - 8:10 am
That's right, numerical integration is not needed for continuous outcomes. Numerical integration is needed when you have a combination of continuous latent variables (factors, growth factors, random effects) and observed non-continuous variables (binary, ordinal, nominal, count, censored).
 Chris Greenwood posted on Tuesday, December 18, 2018 - 2:33 pm
Thanks again for clarifying this.
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