pm.wilcox.test: Wilcoxon test for Partially Matched Two Sample Data

In robustrank: Robust Rank-Based Tests

Description

Performs rank-based two sample test for partially matched two sample data by combining information from matched and unmatched data

Usage

pm.wilcox.test(Xpaired, Ypaired, Xextra = NULL, Yextra = NULL,
 alternative = c("two.sided", "less", "greater"),
 method = c("SR-MW", "MW-MW", "all"), mode = c("test",
 "var", "power.study"), useC = FALSE, correct = NULL,
 verbose=FALSE)

Arguments

Xpaired
Ypaired
Xextra
Yextra
alternative
method	String. SR-MW is recommended, all is for development only.
mode	String. Do not change it to var, for developpment only.
useC	Boolean. Do not set it to TRUE, for development only.
verbose
correct	Continuity correction.

Details

If Xpaired and Ypaired have NAs, the corresponding unpaired data in Ypaired and Xpaired will be combined with Yextra and Xextra.

Value

An htest object.

Examples

set.seed(1)
z=rnorm(20, sd=0.5) # induces correlation between X and Y
X=rnorm(20)+z
Y=rnorm(20,mean=0.8)+z
X[1:10]=NA
boxplot(X,Y,names=c("X","Y"))

pm.wilcox.test(X,Y)
# for comparison
wilcox.test(X,Y,paired=TRUE)
wilcox.test(X,Y,paired=FALSE)# often a conservative test due to the correlation

# no paired data
Y1=Y
Y1[11:20]=NA
pm.wilcox.test(X,Y1)
# should match the following
wilcox.test(X,Y1,paired=FALSE)

# only 1 pair of matched data
Y1=Y
Y1[12:20]=NA
pm.wilcox.test(X,Y1)

robustrank documentation built on July 15, 2020, 5:08 p.m.