xtlogit number of iteration

[felipe cabezon]

Dear Statalist,

I am running a xtlogit regression and the iterations are endless, even though after iteration 5 the log likelihood does not change and says (backed up).
I would like to reduce the number of iterations but the option iterate(#) is not working. I was wondering if there is any way of hurry up the estimation.
Thanks!

Felipe

[Clyde Schechter]

When the log likelihood is not changing and it says "(backed up)" your model is failing to converge. You can't hurry that up: it's just a failed estimation. However, I don't know why the -iterate()- option would fail to work. It should terminate the estimation after the specified number of iterations. In fact, that is the simplest way to find out why your model fails to converge: inspecting the output from terminating while the iteration is "spinning its wheels" will typically reveal the source of the problem. I suggest you post your command(s) and output here for further advice.

[Carlo Lazzaro]

Felipe:
another approach would be to start it all over again, add one predictor at time and see when log likelihood begins to gasp.