predictmeans (version 1.0.9)

permlmer: Permutation Test of random or fixed effects for lmer model.

Description

This function provides permutation tests for the terms in a linear mixed model of lmer.

Usage

permlmer(lmer0, lmer1, nperm = 999, ncore=3, plot=FALSE, seed)

Value

Permutation p-value.

Arguments

lmer0

lmer model under H0, note that lmer0 model must nest within lmer1 model.

lmer1

lmer model under H1, note that lmer0 model must nest within lmer1 model.

nperm

Number of permutation, the default value is 999.

ncore

Number of core for parallel computing, the default value is 3.

plot

Plot permutation distribution or not, the default value is FALSE.

seed

Specify a random number generator seed, for reproducible results.

Author

Dongwen Luo, Siva Ganesh and John Koolaard

References

Oliver E. Lee and Thomas M. Braun (2012), Permutation Tests for Random Effects in Linear Mixed Models. Biometrics, Journal 68(2).

Examples

# library(predictmeans)
## Test random effects
# fm1 <- lmer(Reaction ~ Days + (Days | Subject), sleepstudy)
# fm2 <- lmer(Reaction ~ Days + (Days || Subject), sleepstudy)
# fm3 <- update(fm1, . ~ . - (Days | Subject) + (1 | Subject))
# anova(fm1, fm2, fm3)
# permlmer(fm3, fm2)
# permlmer(fm2, fm1)

## Test fixed effects
# Oats$nitro <- factor(Oats$nitro)
# fm0 <- lmer(yield ~ nitro+Variety+(1|Block/Variety), data=Oats)
# fm <- lmer(yield ~ nitro*Variety+(1|Block/Variety), data=Oats)
# permlmer(fm0, fm)