윌콕슨의 순위합 검정(Wilcoxon rank sum Test)

install.packages("exactRankTests")

library(exactRankTests)

# H0 : mu_A equal mu_B

# H1 : mu_A not equal mu_B

# 1. Wilcoxon Rank Sum Test : no tie
diet.A <- c(79.3, 76.1, 63.4, 58.9, 68.0)
diet.B <- c(77.0, 76.6, 60.2, 59.8, 64.8)

wilcox.test(diet.A, diet.B, alternative = "two.sided")

# 결과물
Wilcoxon rank sum test

data: diet.A and diet.B
W = 13, p-value = 1
alternative hypothesis: true location shift is not equal to 0

# 결론 : 유의확률이 1.000이므로 유의수준 0.05에서 귀무가설을 기각할 수 없다.

# 2. Exact Wilcoxon rank sum test
blood.pressure.man <- c(114, 96, 80, 102, 94, 94, 98, 92, 94, 100, 108, 110, 90, 90, 82, 106)

blood.pressure.woman <- c(108, 98, 88, 86, 100, 98, 104, 102, 94)

exactRankTests::wilcox.exact(blood.pressure.man, blood.pressure.woman, alternative = "two.sided")

# 결과물
Exact Wilcoxon rank sum test

data: blood.pressure.man and blood.pressure.woman
W = 67, p-value = 0.7916
alternative hypothesis: true mu is not equal to 0

# 결론 : 유의확률이 0.7916이므로 유의수준 0.05에서 귀무가설을 기각할 수 없다.

[출처] R을 이용한 비모수 통계학, 임동훈 지음, 자유아카데미, p32~41