Hi Gang,
We're looking at the output of a Tree Growth model using 1dSEM. Our understanding is that for the parsimonious fit index a value closer to 1 is optimal and for AIC the lower the value the better. However, at what point would you throw out a model? In other words, what values would be considered acceptable versus not acceptable? Is there some point at which the model is considered "good" or "poor"?
The output of the Tree Grow model we used is:
++ with new cost = 0.248954, chisq = 72.0948, ntheta = 11
++ parsimonious fit index = 0.312205 Akaike's Information Criterion = 94.0948
Thanks very much for all your help.
best,
Steve