milan lee posted on Friday, December 06, 2013 - 8:22 am
Hello, I have run a 2-class GMM with 1) int and linear slope first and then with 2) quadratic slope in addition. The means of the quadratic growth factor in each class are sig. different from zero. I am not sure if I should compare the model fit between these two models and pick the better fitting one. If so, should I use BIC/AIC or a chi-square difference test to compare the two models? When I looked at the plots of these two models, both linear and quadratic trajectories were supported. Thank you very much!
When you add the quadratic growth factor, if its variance is fixed at zero, you can just look at the z-test of the mean If you want to compare the model with intercept and slope growth factor means and variances to the model with intercept, slope, and quadratic growth factor means and variances, use BIC.