permTest_Contrast: Perform a permutation test on a model object

In RichieHonor/PermutateR: Performs a Permutation Test

Perform a permutation test on a model object

Description

This function performs a permutation test on a model object. The permutation test is based on the test statistics derived from the function "summary", which are typically contrasts. One must specify the variable(s) to be randomized, the test parameter(s), and the amount of replication to be performed. This function can perform extract multiple test parameters from the same fitted model thus allowing permutation tests of multiple parameters to be performed at once, efficiently. This function works on lm, glm, lmer, glmer, glmmTMB, and Lavaan classes.

Usage

permTest_Contrast(
  Model_Object,
  Test_Parameter,
  Randomize_Variables,
  Test_Statistic,
  Replication,
  UseAllAvailableCores = TRUE,
  OutputData = FALSE,
  Data_Supplement = NULL
)

Arguments

Model_Object	A statistical model object.
Test_Parameter	A character string indicating the parameter that the permutation test is to be performed on. For example, 'var1:var2' or "var1". This function also has functionality to extract multiple test statistics from a single permutation test. To perform permutation tests on multiple parameters, input a vector of parameters. For example c("var1:var2","var1","var3").
Randomize_Variables	A character vector indicating the variables that are to be randomized in the permutation test
Test_Statistic	A Character string indicating the desired test statistic to conduct the permutation test.This can be an f value, chi-squared value, p-value...etc, depending on what test statistics are available for your model object. These are different for lm, glm, glmmTMB, etc.
Replication	The number of simulations in the permutation test.
OutputData	Should the simulated test statistics be outputted ?
Data_Supplement	The data frame that the model is built on. This is necessary to supply for models of the class "lavaan" as there is not functionality to extract the data frame from the model object.

Value

A list of two items. The first is a data frame of results of the permutation test. The second is a histogram displaying the sampling distribution of the simulated test statistics, with a red line displaying the test statistic of the original (non-randomized) data. If OutputData=T, then a vector of simulated test statistics will also be included in the output list. If multiple test parameters are supplied, then this output will be in a list for each test parameter.

RichieHonor/PermutateR documentation built on July 2, 2023, 2:07 p.m.