[RsR] Kolmogrov-Smirov test and Wilcoxon test

[Wei-han Liu]

Hi R users:

Initially both Kolmogrov-Smirov dominance
test and Wilcoxon rank sum and signed rank test are designed for independent
samples but now there are algorithm developed for related samples.
�
I applied both Kolmogrov-Smirov dominance
test and Wilcoxon rank sum and signed rank test on two related financial series
in order to find the ordering of the data involved. I have four series to
compare, so I need to conduct the same procedure to make 6 comparisons (take 2
out of 4, 4C2) among the four data series.
�
However, the testing outcomes out of
these comparisons by the two tests are sometimes somewhat different or even
conflicting. For example, K-S test summarizes A1<A2<A3<A4, but
Wilcozon test gives A1<A2?A3<A4 or K-S test summarizes A1<A2<A3<A4,
but Wilcozon test gives A1>A2>A3>A4 in other comparison round. What
can I interpret the difference? Any implication behind these?
�
Best regards,
�
Weihan
	[[alternative HTML version deleted]]