AnovaMix function

Mixed model least squares analysis of variance (mixed ANOVA).

Uses output from lm() in mixlm package to compute ANOVA table, variance components and errors. latin1

AnovaMix(object, SStype)

Arguments

• object: object fitted by lm (mixlm package) containing at least one random effect.
• SStype: type of sums-of-squares (I/II/III) for Analysis of Variance.

Details

AnovaMix can either be invoked directly or through the Anova() function (with type III error).

Returns

• lm: linear model fitted by lm in package mixlm.

• anova: ANOVA table.

• err.terms: list of denominator information for F tests.

• denom.df: numeric of denominator degrees of freedom for F tests.

• restricted: logical indicating if ANOVA used restricted modelling.

• exp.mean.sq: character containing expected mean squares.

• var.comps: numeric containing variance components.

• random.effects: character containing the random effects.

• ind.randoms: numeric with indices of random effects in the model.

• formula.text: character containing all effects of the model.

Author(s)

Kristian Hovde Liland

Note

Only balanced models are fully supported.

See Also

print.AnovaMix, Anova, lm

Examples

mydata <- data.frame(y = rnorm(12),
                                                          x = factor(c(rep(2,4),rep(1,4),rep(0,4))),
                                                          z = factor(rep(c(1,0),6)))
mixlm <- lm(y~x*r(z),
                        data = mydata)
Anova(mixlm,type="III")