Comparing two growth models PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
 CC.J. T. posted on Friday, June 08, 2018 - 1:28 am
I tested latent basis model and linear growth model in my data. Both showed a good model fit.
latent basis model, chi-square = 2.841 (p =.41), RMSEA=.000, CFI = 1.00, TLI = 1.00, SRMR = .038

linear growth model, chi-square = 6.094 (p =.297), RMSEA=.032, CFI = .979, TLI = .975, SRMR = .066.

I would like to know how to decide which one is better, using chi-square difference test? or BIC ? or BIC and AIC? Thank you very much.
 Bengt O. Muthen posted on Friday, June 08, 2018 - 1:48 pm
If the two models are nested, you can use a chi-square difference test. You can check if they are nested by using our new NESTED feature in Mplus Version 8.1.
 CC.J. T. posted on Monday, June 11, 2018 - 1:29 am
Thank you very much. I have not got Mplus 8 yet. But I am curious if they are nested models statistically. From your answer, it seems that sometimes they are but sometimes they are not. So how to decide ? Thanks
 Bengt O. Muthen posted on Monday, June 11, 2018 - 5:36 pm
It can be difficult to decide. No general rules are available. The new NESTED procedure is the way to go. You can read our paper on this - as well as the bentler-satorra paper it refers to - to understand nesting better.
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