Marginal values after a shared frailty model

[Navid Asgari]

I am running a shared frailty model (--streg--). As part of post-estimation, when I calculate the marginal values, the margins are 100 times larger than when I run --streg-- without the shared() option (i.e., when I run a frailty model instead of a shared frailty model). Why?

Here is how the results look after running a non-shared frailty model:

And this is how they look after a shared frailty model:

[Stephen Jenkins]

You appear to have shown the same output multiple times. No matter, that one picture suggests that there is something odd about your model specification. Why are all the standard errors so huge? If I were you, I would invest energies in addressing that issue first.