0
$\begingroup$

In R, wilcoxsign_test in library coin calculates both a "teststatistic" (output as "Z") and a "linearstatistic" (I think this is either Siegel's (1956) T or Glantz' (2005) W).

What is "Z", which should I report, and why?

Example:

x <- 1:10
y <- 11:20
library(coin)
wilcoxsign_test(y ~ x, distribution="exact")
#   Exact Wilcoxon-Signed-Rank Test
# 
# data:  y by x (neg, pos) 
#    stratified by block
# Z = 3.1623, p-value = 0.001953
# alternative hypothesis: true mu is not equal to 0
wilcoxsign_test(y ~ x, distribution="exact")@statistic@linearstatistic
# [1] 55
wilcoxsign_test(y ~ x, distribution="exact")@statistic@teststatistic
#      neg 
# 3.162278
$\endgroup$
  • 2
    $\begingroup$ coin vignettes and JSS article provide an answer to your questions. $\endgroup$ – Tim Jan 12 '15 at 10:16
  • $\begingroup$ @Tim From the information you gave me, is it correct to say that the symbol for "linearstatistic" is "T"? $\endgroup$ – user14650 Jan 13 '15 at 12:04
  • 1
    $\begingroup$ Yes, and you can extract those using statistic(object, "linear") or statistic(object, "standardized"). $\endgroup$ – Tim Jan 13 '15 at 12:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy