MuMIn (version 0.13.3)

QAIC: Quasi AIC

Description

Calculates quasi AIC for one or several fitted model objects. This function is provided just as an example of custom rank function for use with model.avg and dredge

Usage

QAIC(object, ..., chat)

Arguments

object

a fitted model object.

...

optionally more fitted model objects.

chat

c - hat

Value

• If just one object is provided, returns a numeric value with the corresponding QAIC; if more than one object are provided, returns a data.frame with rows corresponding to the objects.

encoding

utf-8

Details

rank is specified as a function or a symbol (e.g. a backquoted name) or a character string specifying a function.

Function rank must be able to accept model as a first argument and must always return a scalar.

Examples

budworm <- data.frame(ldose = rep(0:5, 2), numdead = c(1, 4, 9, 13, 18, 20, 0,
	2, 6, 10, 12, 16), sex = factor(rep(c("M", "F"), c(6, 6))))

budworm$SF <- cbind(budworm$numdead, 20 - budworm$numdead)

budworm.qlg <- glm(SF ~ sex*ldose, family = quasibinomial, data = budworm)

dd1 <- dredge(budworm.qlg, rank = "QAIC",
	chat = summary(budworm.qlg)$dispersion)
gm1 <- get.models(dd1, 1:4)

model.avg(gm1)

model.avg(gm1[[1]], gm1[[2]], rank = "QAIC", rank.args = list(chat = 1))