[R] Likelihood ratio based confidence intervals for logistic regression
I'm applying logistic regression to a moderate sized data set for which I believe Wald based confidence intervals on B coefficients are too conservative. Some of the literature recommends using confidence intervals based on the likelihood ratio in such cases, but I'm having difficulty locating a package that can do these. Any help would be immensely appreciated. Best, Jeff Hanna -- View this message in context: http://r.789695.n4.nabble.com/Likelihood-ratio-based-confidence-intervals-for-logistic-regression-tp2077303p2077303.html Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Likelihood ratio based confidence intervals for logistic regression
jh556 wrote: I'm applying logistic regression to a moderate sized data set for which I believe Wald based confidence intervals on B coefficients are too conservative. Some of the literature recommends using confidence intervals based on the likelihood ratio in such cases, but I'm having difficulty locating a package that can do these. Any help would be immensely appreciated. Are you looking for profile-likelihood based CIs? Package MASS has those, via the confint function (if my memory is correct...) fm1 - glm(..., family = binomial) confint(fm1) __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Likelihood ratio based confidence intervals for logistic regression
Some quick googling suggests that they are the same thing. Thanks for the help! -- View this message in context: http://r.789695.n4.nabble.com/Likelihood-ratio-based-confidence-intervals-for-logistic-regression-tp2077303p2077354.html Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Likelihood ratio based confidence intervals for logistic regression
On 2010-04-30 12:42, jh556 wrote: Some quick googling suggests that they are the same thing. Thanks for the help! And note that profile likelihood CIs are produced by default on glm objects, i.e. R uses MASS's confint.glm for glm objects. confint.default(your model) let's you compare with Wald CIs. -- Peter Ehlers University of Calgary __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.