Log-Likelihood of an lmList Object

If pool=FALSE, the (restricted) log-likelihoods of the lm components of object are summed together. Else, the (restricted) log-likelihood of the lm fit with different coefficients for each level of the grouping factor associated with the partitioning of the object components is obtained.

# S3 method for class 'lmList'
logLik(object, REML, pool, ...)

Arguments

object

an object inheriting from class "lmList", representing a list of lm objects with a common model.

REML

an optional logical value. If TRUE the restricted log-likelihood is returned, else, if FALSE, the log-likelihood is returned. Defaults to FALSE.

pool

an optional logical value indicating whether all lm components of object may be assumed to have the same error variance. Default is attr(object, "pool").

...

some methods for this generic require additional arguments. None are used in this method.

Value

either the sum of the (restricted) log-likelihoods of each lm component in object, or the (restricted) log-likelihood for the lm fit with separate coefficients for each component of object.

Author

José Pinheiro and Douglas Bates bates@stat.wisc.edu

See also

lmList, logLik.lme,

Examples

fm1 <- lmList(distance ~ age | Subject, Orthodont)
logLik(fm1)   # returns NA when it should not
#> 'log Lik.' -144.9816 (df=55)