calcSatterth: F-test based on the Satterthwaite's approximation for...

In alku86/lmerTest: Tests in Linear Mixed Effects Models

Description

Produces a list with the values for an approximate F-test based on the Satterthwaite's approximation.

Usage

calcSatterth(model, L)

Arguments

model	linear mixed effects model (lmer object).
L	hypothesis contrast matrix or a vector
...	other potential arguments.

Details

F test for the null hypothesis H_0: L β = 0, where β is a vector of the same length as fixef(model)

Value

A list with the results from the F test

denom	numeric. Denominator degrees of freedom, calculated with the Satterthwaite's approximation
Fstat	numeric. F statistic
pvalue	numeric. p-value of the corresponding F test
ndf	numeric. Numerator degrees of freedom

Author(s)

Alexandra Kuznetsova, Per Bruun Brockhoff, Rune Haubo Bojesen Christensen

References

Schaalje G.B., McBride J.B., Fellingham G.W. 2002 Adequacy of approximations to distributions of test Statistics in complex mixed linear models

See Also

anova

Examples

## import lme4 package and lmerTest package
library(lmerTest)

## specify lmer model for the sleepstudy data from the lme4 package
m <- lmer(Reaction ~ Days + (1 + Days|Subject), sleepstudy)


L <- cbind(0,1) ## specify contrast vector
calcSatterth(m, L) ## calculate F test

## specify model for the ham data
m.ham <- lmer(Informed.liking ~ Product + (1|Consumer), data = ham)

## specify contrast vector for testing product effect
L <- matrix(0, ncol = 4, nrow = 3)
L[1,2] <- L[2,3] <- L[3,4] <- 1
calcSatterth(m.ham, L)

## by using anova function we get the same result
anova(m.ham)

alku86/lmerTest documentation built on May 10, 2019, 9:22 a.m.