clusterlm: Integration of the quasi-F statistics into the 'clusterlm'...

In jaromilfrossard/permucoQuasiF: Permutation test for quasi-F statistic

Description

Provides cluster-mass tests based on permutations for quasi-F statistics

Usage

clusterlm(formula, data = NULL, np = 5000, method = NULL,
  test = "fisher", threshold = NULL, aggr_FUN = NULL,
  multcomp = "clustermass", effect = NULL, ...)

Arguments

formula	A formula object. The formula for the quasi-F statistic should be written using the notation + Error(id/(A*C)) + Error(id/(B*C)) to specify random effect associate to subjects and items, where A is a within id factor, B is a within item factor and C is within id and item factors.
data	similar to the permuco package.
np	The number of permutations. Default value is 5000.
method	similar to the permuco package. The method is set to "terBraak" and "terBraak_logp" are available for the quasi-F statistic.
test	similar to the permuco package. the argument does not influence the quasi-F statistic.
threshold	similar to the permuco package.
aggr_FUN	similar to the permuco package.
multcomp	similar to the permuco package.
effect	a numeric specifying the effect to test. The value correspond to the "assign" attribute of the model.matrix argument of the fixed effect. The default is NULL tests all effects.
...	Futher arguments, see details.

Details

Similar to the permuco package.

Value

A list containing : a table of the clusters, or a multcomp object for the other multiple comparison procedures. Use the plot.clusterlm method to have a quick overview of the results.

Author(s)

[e-mail]

Examples

 ## Not run: 

# packages

library(permuco)
devtools::install_github("jaromilfrossard/permucoQuasiF")
library(permucoQuasiF)

# data

data("signal")
data("design")

model <- clusterlm( signal ~ A*B*C + Error(id/(B*C)) + Error(item/(A*C)), data=design, np = 5)
plot(model)
model
## End(Not run)

jaromilfrossard/permucoQuasiF documentation built on May 5, 2019, 12:30 p.m.