Linear Mixed Models with same variance for contrasts.

This function provide a small modification of the lme4 function link{lmer} in order to estimated mixed models with the same variance for orthonormal contrasts. Check formula parameter for details.

gANOVA(
  formula,
  data = NULL,
  REML = TRUE,
  control = lmerControl(),
  start = NULL,
  verbose = 0L,
  subset,
  weights,
  na.action,
  offset,
  contrasts = NULL,
  devFunOnly = FALSE,
  ...
)

Arguments

formula	a lme4 formula. If you provide a formula of type (1|id:g) where id is the grouping variable and g is a factor then the covariance structure will estimate the same variance for all orthonormal (contr.poly) contrasts in g. WARNINGS The identifier of the grouping variable id must be written as the first terms to the right of the "|" because all other terms will be "reduced" by an orthonormal contrasts. See Details for the reduced notation.
data	a data frame. See lmer for more details.
REML	a logical that indicate which criterion to optimize. See lmer for more details.
control	Some parameters. See lmerControl or lmer for more details.
start	starting values for the paramters. See lmer for more details.
verbose	See lmer for more details.
subset	an expression to selecte a subset of the data. See lmer for more details.
weights	an optional vector of weights. See lmer for more details.
na.action	a function that handle NA's. See lmer for more details.
offset	specify a priori component in the predictor. See lmer for more details.
contrasts	a list of contrasts. See lmer for more details.
devFunOnly	a logical set by default to FALSE. See lmer for more details.
...	addition arguments. See lmer for more details.

Details

summary and anova method are copied from the lmerTest package in order to have p-values. Given 2 factors, f and g, and a grouping variable id, the formula (1|id) + (1|id:f)+ (1|id:g)+ (1|id:f:g) can be reduced to (1|id|f*g). However the grouping variable (herer id) should be written as one variable.

See also

lmer.

Other estimation: refit.lmerModgANOVA()