-1
$\begingroup$

Currently, I am working with with Kendall tau distance between two rankings. I get a tau value of 0.849, Z (statistical significance of 3.9) and a Kendall score of 56. What does this Kendall score tell us about? Here under are my results and I would appreciate if anyone could enlighten me with these values of S, D and varS - if they are worthy of adding to a research evaluation. Or is only the kendall tau and Z are enough to tell the story?

kendal$s
    [1] 56
    attr(,"Csingle")
    [1] TRUE
kendal$D
[1] 66.00001
attr(,"Csingle")
[1] TRUE
>
kendal$varS
[1] 212.6667
attr(,"Csingle")
[1] TRUE
$\endgroup$
1
$\begingroup$

It looks like this "Kendall score" is only an intermediate step in the computation of τ (the Kendall correlation) and not of any substantive interest itself. There's no need to report it. Focus on the τ.

$\endgroup$
1
$\begingroup$

Sort the data so that one variable is in order (call it $A$) then take each value of $B$ except the rightmost and count how many values of $B$ to the right of it are larger (call it $p$) and smaller (call it $q$). The add up all the $p$ and call it $P$ and similarly to form $Q$. Then $S = P - Q$. Whether that has any meaning in your scientific question is up to you. It is a way of defining the way in which Kendall decided to measure correlation. $S$ also plays a fundamental role in many other so-called non-parametric tests.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.