Kendall Rank Correlation Coefficient

Abstract

The Kendall (1955) rank correlation coefficient evaluates the de- gree of similarity between two sets of ranks given to a same set of objects. This coefficient depends upon the number of inversions of pairs of objects which would be needed to transformone rank order into the other. In order to do so, each rank order is repre- sented by the set of all pairs of objects (e.g., [a,b] and [b,a] are the two pairs representing the objects a and b), and a value of 1 or 0 is assigned to this pair when its order corresponds or does not correspond to the way these two objects were ordered. This cod- ing schema provides a set of binary values which are then used to compute a Pearson correlation coefficient.