simulateQuasif.fnc: Simulate data for quasi-F analysis and compare models

In languageR: Analyzing Linguistic Data: A Practical Introduction to Statistics

Description

This function creates a user-specified number of simulated datasets, and compares mixed-effects models with quasi-F and F1 and F2 analyses. It should be run with the version of R and the version of languageR used by Baayen, Davidson & Bates (2008, JML), as mcmcsamp no longer supports models with random correlation parameters.

Usage

simulateQuasif.fnc(dat, with = TRUE, nruns = 100, nsub = NA, nitem = NA, ...)

Arguments

dat	Data frame with a data set with as variables Subject, Item and SOA, as in the quasif data set.
with	Logical, if TRUE, an effect of SOA is built into the simulation.
nruns	Integer for the number of simulation runs.
nsub	Integer denoting the number of subjects.
nitem	Integer denoting the number of items.
...	other parameters to be passed through to plotting functions.

Details

Model parameters are estimated from the input data set.

For each completed simulation run, a dot is added to the R console.

Value

A list with components

alpha05	Description of 'comp1'
alpha01	proportion of runs in which predictors are significant at the 05 significance level.
res	Data frame with simulation results.
with	Logical, TRUE if SOA effect is built into the simulations.

Author(s)

R. H. Baayen

See Also

See also subjects.quasif.fnc.

Examples

## Not run: 
data(quasif)
library(lme4)

quasif.sim = simulateQuasif.fnc(quasif, nruns = 1000, with = TRUE) 
quasif.sim$alpha05

## End(Not run)

languageR documentation built on May 2, 2019, 10:02 a.m.