[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
Has anyone used rpoisson? If you have, I'd really appreciate some advice about the best way to compare models. Fitstat and lrtest commands do not work after this command.
Apologies for repeating this request, but I didn't get any replies on this when I posted it to the list first time round.
Imperial College London
[mailto:firstname.lastname@example.org]On Behalf Of Michael Hills
Sent: 08 April 2004 10:53
Subject: Re: st: ORs for non-rare outcomes
You could try glm with family binomial and a log link instead of logit.
This will give you risk ratios.
* For searches and help try: