README

varTestnlme implements the likelihood ratio test (LRT) for testing the presence of random effects in linear, generalized linear and nonlinear mixed-effects model. The test can be used to answer questions of the type:

It is possible to compare two models with different random effects, provided that the random structures of the two models are nested.

The package works on models that were fitted using nlme, lme4, or saemix packages.

Reference

Baey C, Cournède P-H, Kuhn E, 2019. Asymptotic distribution of likelihood ratio test statistics for variance components in nonlinear mixed effects models. 135:107–122 (2019), https://doi.org/10.1016/j.csda.2019.01.014

Installation

install.packages("varTestnlme")

# install.packages("devtools")
#devtools::install_github("baeyc/varTestnlme")

devtools::load_all()

Example

library(nlme)
data("Orthodont")

# using nlme, with correlated slope and intercept
m1 <- lme(distance ~ 1 + Sex + age + age*Sex, random = pdSymm(Subject ~ 1 + age), data = Orthodont, method = "ML")
m0 <- lme(distance ~ 1 + Sex + age + age*Sex, random = ~ 1 | Subject, data = Orthodont, method = "ML")

vt <- varTest(m1,m0)
#> Variance components testing in mixed-effects models
#> (models fitted using the nlme package)
#> Testing that the variance of age is null
#> model under H1: distance ~ 1 + Sex + age + age * Sex  (fixed effects),  pdSymm(Subject ~ 1 + age)  (random effects)
#> model under H0: distance ~ 1 + Sex + age + age * Sex  (fixed effects),  ~1 | Subject  (random effects)
#> 
#> Likelihood ratio test statistics: 
#>  LRT =  0.83311 
#> 
#> Limiting distribution:
#> mixture of 2 chi-bar-square distributions with degrees of freedom 1, 2
#> lower-bound for p-value: 0.51035  upper bound for p-value: 0.51035