lrtest: Likelihood ratio test

In epiDisplay: Epidemiological Data Display Package

Likelihood ratio test

Description

Likelihood ratio test for objects of class 'glm'

Usage

lrtest (model1, model2)

Arguments

model1, model2

Two models of class "glm" having the same set of records and the same type ('family' and 'link')

Details

Likelihood ratio test checks the difference between -2*logLikelihood of the two models against the change in degrees of freedom using a chi-squared test. It is best applied to a model from 'glm' to test the effect of a factor with more than two levels. The records used in the dataset for both models MUST be the same. The function can also be used with "clogit", which does not have real logLikelihood.

Author(s)

Virasakdi Chongsuvivatwong [e-mail]

See Also

'glm', 'logLik', 'deviance'

Examples

model0 <- glm(case ~ induced + spontaneous, family=binomial, data=infert)
model1 <- glm(case ~ induced, family=binomial, data=infert)
lrtest (model0, model1)
lrtest (model1, model0) # same result
lrtest (model1, model0) -> a
a

epiDisplay documentation built on May 18, 2022, 5:11 p.m.