Hello, Statalist community!
I feel a little foolish posting this question but I'm hoping that someone will be kind enough to share some insights.
I am reviewing several different models with varying structural differences. I am modeling panel data from 55 neighborhoods during 1990-2014 I am trying to figure out which model has the best fit. 1. XT (GEE, POISSON, NBREG) each reviewing the different correlation structures of independent, exchangeable, identity, unstructured, or ar1)
2. ME (POISSON, NBREG) ) each reviewing the different correlation structures of independent, exchangeable, identity, unstructured, or ar1)
3. GLM (Gamma distribution and identity link)
Currently, I run each model then use the test and lrtest commands, as well as compare the AIC BIC. As well as using the test and lrtest commands. I am using Stata 15.1.
Is there anything else I should be doing?
Sincerely,
Brennan
Related Posts with Goodness of fit for panel data
Difference interaction with country variable and multilevelHey, I'm about to examine how a two-way interaction differs depending on the country (I examine thre…
Using reg or xtreg?Hi I'm new to STATA. Just completed a survey capturing about 3000+ responses from different city/pro…
New post-estimation command -suregr- available on SSC: calculates robust, or cluster robust variance post -sureg- estimation.Thanks to Kit Baum, I have made available on SSC a new post estimation command -suregr- (mnemonic fo…
Problem with renaming varsHi, I would like to rename a set of vars if they end with odd numbers and are between 1 and 42. The …
xtabond2 with p-value of AR(1) test > 0.1Dear All, In a typical application (xtabond2 to a dynamic panel data model), one often needs to chec…
Subscribe to:
Post Comments (Atom)
0 Response to Goodness of fit for panel data
Post a Comment