![Code and results[![][1]](https://i.stack.imgur.com/5f83c.png)
]2
I don't understand why these results differ (see both Pictures attached).. Any suggestions? Thanks in advance!
Asked
Active
Viewed 23 times
0
-
Which part of the output are you referring to, specifically? – Erik Ruzek Mar 28 '20 at 02:17
-
the coefficient and p-value of the interaction term especially. Thanks for your help again :-) – annwy Mar 28 '20 at 15:23
-
They are not different enough to worry about, in my opinion. Can you report the AIC and BIC of the xtlogit fit in Stata? It may not provide it, and if so, just refit the model with `melogit` and after the model is estimated, run `estat ic`. See also https://stats.stackexchange.com/questions/443142/advice-for-r-packages-for-glmm-and-adaptive-gauss-hermite-quadrature/443183#443183 – Erik Ruzek Mar 28 '20 at 16:14