Nonlinear Rank Correlations

Tarsitano, A. (2008) Nonlinear Rank Correlations. Working Paper. Agostino Tarsitano, Italy. (In Press)

[img]
Preview
PDF
noliran.pdf

Download (218kB)
Official URL: http://www.ecostat.unical.it/Tarsitano/Publication...

Abstract

Rank correlation is a fundamental tool to express dependence in case in which the data are arranged in order. There are, by contrast, circumstances where the ordinal association is of a nonlinear type. In this paper we have investigated the effectiveness of thirty measures of rank correlation for assessing agreement between two evaluators in the presence of non costant scale of terms. These measures have been divided into three classes: unweighted rank correlations, weighted rank correlations, correlations of scores. Our findings suggest that none is systematically better than the other in all circumstances. However, a simply weighted version of the Kendall’s $\tau$ provides plausible answers to many special situations where intercategory distances are not the same.

Item Type: Monograph (Working Paper)
Uncontrolled Keywords: ordinal data, nonlinear association
Subjects: 500 Scienze naturali e Matematica > 507 Educazione; Ricerca; Soggetti connessi > 507.2 Ricerca; Metodi statistici
Depositing User: Prof Agostino Tarsitano
Date Deposited: 15 Sep 2008
Last Modified: 20 May 2010 12:01
URI: http://eprints.bice.rm.cnr.it/id/eprint/586

Actions (login required)

View Item View Item