[R] How can I extract the AIC score from a mixed model object produced using lmer?

David Hewitt dhewitt at vims.edu
Wed Dec 19 16:42:40 CET 2007

David Barron-3 wrote:
> You can calculate the AIC as follows:
> (fm1 <- lmer(Reaction ~ Days + (Days|Subject), sleepstudy))
> aic1 <- AIC(logLik(fm1))

Is AIC() [extractAIC()] "valid" for models with random effects? I noticed
that the help page for extractAIC() does not list models with random
effects. I think this boils down to the difference between the likelihoods
for models with and without random effects, and I don't know. Just

> On 12/18/07, Peter H Singleton <psingleton at fs.fed.us> wrote:
>> I am running a series of candidate mixed models using lmer (package lme4)
>> and I'd like to be able to compile a list of the AIC scores for those
>> models so that I can quickly summarize and rank the models by AIC. When I
>> do logistic regression, I can easily generate this kind of list by
>> creating
>> the model objects using glm, and doing:
>> > md <- c("md1.lr", "md2.lr", "md3.lr")
>> > aic <- c(md1.lr$aic, md2.lr$aic, md3.lr$aic)
>> > aic2 <- cbind(md, aic)
>> but when I try to extract the AIC score from the model object produced by
>> lmer I get:
>> > md1.lme$aic
>> Warning message:
>> In md1.lme$aic : $ operator not defined for this S4 class, returning NULL
>> So... How do I query the AIC value out of a mixed model object created by
>> lmer?
R-help at r-project.org mailing list
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.

David Hewitt
Virginia Institute of Marine Science
View this message in context: http://www.nabble.com/How-can-I-extract-the-AIC-score-from-a-mixed-model-object-produced-using-lmer--tp14406832p14419438.html
Sent from the R help mailing list archive at Nabble.com.

More information about the R-help mailing list