MCAR_meancovTest {MCARtest}

Carry out a test of MCAR using first and second moments.

Description

This is the implementation of Algorithm 1 in Bordino and Berrett (2024).

Usage

MCAR_meancovTest(X, alpha, B)

Arguments

X

The dataset with incomplete data.

alpha

The nominal level of the test.

B

The bootstrap sample B for the bootstrap test.

Value

A Boolean, where TRUE stands for reject MCAR. This is found as outlined in Section 5.2 in Bordino and Berrett (2024).

References

Bordino A, Berrett TB (2024). “Tests of Missing Completely At Random based on sample covariance matrices.” arXiv preprint arXiv:2401.05256.

Examples

library(MASS)
alpha = 0.05
B = 20
m = 500

SigmaS=list() #Random 2x2 correlation matrices (necessarily consistent)
for(j in 1:3){
x=runif(2,min=-1,max=1); y=runif(2,min=-1,max=1)
SigmaS[[j]]=cov2cor(x%*%t(x) + y%*%t(y))
}

X1 = mvrnorm(m, c(0,0), SigmaS[[1]])
X2 = mvrnorm(m, c(0,0), SigmaS[[2]])
X3 = mvrnorm(m, c(0,0), SigmaS[[3]])
columns = c("X1","X2","X3")
X = data.frame(matrix(nrow = 3*m, ncol = 3))
X[1:m, c("X1", "X2")] = X1
X[(m+1):(2*m), c("X2", "X3")] = X2
X[(2*m+1):(3*m), c("X1", "X3")] = X3
X = as.matrix(X)

MCAR_meancovTest(X, alpha, B)