# mdTest: Mahalanobis Distance Test for GMPMs In gmpm: Generalized Multilevel Permutation Models

## Description

Performs multiparameter tests on GMPM objects (or matrices) using Mahalanobis distances calculated on the permutation matrix.

## Usage

 `1` ```mdTest(x, y) ```

## Arguments

 `x` Either a `GMPM` object or a permutation matrix `y` Either a list of vectors, or a single vector. The vector (or vectors) should index the columns of the matrix on which to perform the test. They can either be numeric indices or the names of the matrix columns. If `y` is a list, then a separate test will be performed for each vector in the list.

## Value

A `Mdtest` object

S4 classes `Mdtest`, `GMPM`, and function `gmpm`
 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20``` ```# create a within-subjects design # we are interested in main effect of A (3-level variable) # this requires a multivariate test df1 <- data.frame(SubjID=rep(1:12,each=6), A=rep(c("Aa","Ab","Ac"),each=2), B=c("Ba","Bb")) df1\$Y <- rep(c(0,1,-1),each=2) + rnorm(72) # parametric analysis summary(aov(Y ~ A*B + Error(SubjID), df1)) # randomization analysis; should increase maxruns to something like 999 df1.gmpm <- gmpm(Y ~ A*B | SubjID, gaussian, df1, c("A","B"), gmpmControl=list(maxruns=99)) # retrieve the permutation matrix pmx <- getPermMx(df1.gmpm)[["within"]] # perform test using permutation matrix mdTest(pmx, c("A1","A2")) ```