Abstract
In cases where only linear relationships are suspected, Pearson’s correlation is generally used to measure the strength of the association between variables. It is well-known, however, that when a non-linear or non-linearizable connection exists, the use of Pearson’s coefficient on original values can be deceiving. On the other hand, rank correlations should perform satisfactorily because of their properties and versatility. There are many coefficients of rank correlation, from simple ones to complicated definitions invoking one or more special transformations. Each of these methods is sensitive to a different feature of dependence between variables The purpose of this article is to find a coefficient, if one exists, that tends to be different from zero at least in a meaningful way more often than others when the relationship between two rankings is of a non linear type. In this regard, we analyze the behavior of a few well-known rank correlation coefficients by focusing on some frequently encountered non-linear patterns. We conclude that a reasonably robust answer to the special needs arising from non-linear relationships could be given by a variant of the Fisher–Yates coefficient, which has a more marked tendency to reject the hypothesis of independence between pairs of rankings connected by several forms of non-linear interaction.
Similar content being viewed by others
References
Amerise IL, Tarsitano A (2016) A new symmetrical test of bivariate independence (under review)
Amerise IL, Marozzi M, Tarsitano A (2016) pvrank: Rank correlations. R package version 1.1.1. http://CRAN.R-project.org/package=pvrank
Barton DE, Mallows CL (1965) Some aspects of the random sequence. Ann Math Stat 36:236–260
Bhat DN, Nayar SK (1997) Ordinal measures for visual correspondence. In: Proceedings of the IEEE computer society conference on computer vision and pattern recognition, San Francisco (CA, USA), 18–20 June 1996, pp 351–357
Blomqvist N (1950) On a measure of dependence between two random variables. Ann Math Stat 21:593–600
Borroni CG (2013) A new rank correlation measure. Stat Pap 54:255–270
Borroni CG, Zenga M (2007) A test of concordance based on Gini’s mean difference. Stat Methods Appl 16:289–308
Cameron PJ, Wu T (2010) The complexity of the weight problem for permutation and matrix groups. Electron Notes Discret Math 28:109–116
Campano F, Salvatore D (1988) Economic development, income inequality and Kuznets’ U-shaped hypothesis. J Policy Model 10:265–280
Conover WJ, Iman RL (1981) Rank transformations as a bridge between parametric and nonparametric statistics. Am Stat 35:124–129
Critchlow DE (1986) A unified approach to constructing nonparametric rank tests. Technical Report, Department of Statistics, Purdue University, Indiana, 86-15. Available at http://www.stat.purdue.edu/docs/research/tech-reports/1986/tr86-15.pdf
Dallal GE, Hartigan JA (1980) Note on a test of monotone association insensitive to outliers. J Am Stat Assoc 75:722–725
Diaconis P, Graham RL (1977) Spearmans footrule as a measure of disarray. J R Stat Soci B 39:262–268
Fieller EC, Pearson ES (1961) Tests for rank correlation coefficients: II. Biometrika 48:29–40
Filliben JJ (1975) The probability plot correlation coefficient test for normality. Technometrics 17:111–117
Fisher R, Yates F (1938) Statistical tables for biological, agricultural and medical research, 5th edn. Oliver and Boyd, Edinburgh
Genest C, Plante J-F (2003) On Blest’s measure of rank correlation. Can J Stat 31:35–52
Gideon RA, Hollister A (1987) A rank correlation coefficient resistant to outliers. J Am Stat Assoc 82:656–666
Gini C (1914) Di una misura delle relazioni tra le graduatorie di due caratteri. Tipografia Cecchini, Roma
Gordon AD (1979a) A measure of the agreement between rankings. Biometrika 66:7–15
Gordon AD (1979b) Another measure of the agreement between rankings. Biometrika 66:327–332
Hájek J, Šidak Z, Sen PK (1999) Theory of rank tests, 2nd edn. Academic Press, San Diego
Hamming RW (1950) Error detecting and error correcting codes. Bell Syst Tech J 29:147–160
Hoeffding W (1948) A non-parametric test of independence. Ann Math Stat 19:546–557
Hotelling H, Pabst MR (1936) Rank correlation and tests of significance involving no assumption of normality. Ann Math Stat 7:29–43
Kemeny JG (1959) Mathematics without numbers. Daedalus 88:577–591
Kendall MG (1938) A new measure of rank correlation. Biometrika 30:91–93
King TS, Chinchilli VM (2001) Robust estimators of the concordance correlation coefficient. J Biopharm Stat 11:83–105
Knuth DE (1973) The art of computer programming, 2nd edn. Vol 1: Fundamental algorithms. Addison-Wesley Publishing Company, Reading
Lieberson S (1964) Limitations in the application of non-parametric coefficients of correlation. Am Sociol Rev 29:744–746
Mango A (2006) A distance function for ranked variables: a proposal for a new rank correlation coefficient. Metodološki Zvezki 3:9–19
Millard SP (2013) EnvStats: an R package for environmental statistics. Springer, New York
Moore GH, Wallis AW (1943) Time series significance test based on sign differences. J Am Stat Assoc 38:153–165
Moran PAP (1950) A curvilinear ranking test. J R Stat Soc B 12:292–295
Niewiadomska-Bugaj M, Kowalczyk T (2006) A new test of association and other tests based on the Gini mean difference. Metron 64:399–409
Peters CC (1946) A new descriptive statistic: the parabolic correlation coefficient. Psychometrika 11:57–69
Rényi A (1959) On measures of dependence. Acta Math Hung 10:441–451
Salama IA, Quade D (1997) The asymptotic normality of a rank correlation statistic based on rises. Stat Probab Lett 32:201–205
Salama IA, Quade D (2001) The symmetric footrule. Commun Stat Part A Theory Methods 30:1099–1109
Salvemini T (1951) Sui vari indici di concentrazione. Statistica 2:133–154
Sarndäl CE (1974) A comparative study of association measures. Psychometrika 39:165–187
Scarsini M (1984) On measures of dependence. Stochastica 8:201–218
Schweizer B, Wolff EF (1981) On nonparametric measures of dependence for random variables. Ann Stat 9:879–885
Shaper AG, Wannamethee G, Walker M (1988) Alcohol and mortality in British men: explaining the U-shaped curve. Lancet 2:1267–1273
Spearman C (1904) The proof and measurement of association between two things. Am J Psichol 15:72–101
Spearman C (1906) “Footrule” for measuring correlation. Br J Psychol 2:89–108
Sugano O, Watadani S (1993) An association analysis based on a statistic orthogonal to linear rank statistics. Behaviormetrica 29:17–33
Tarsitano A, Amerise LI (2015) On a new measure of rank-order association. J Stat Econom Methods 4:83–105
Tarsitano A, Lombardo R (2013) A coefficient of correlation based on ratios of ranks and anti-ranks. Jahrnbucher fr Nationalkonomie und Statistik 233:206–224
Theil H (1950) A rank-invariant method of linear and polynomial regression analysis. Part 3. Proceedings of Koninklijke Nederlandse Akademie van Wetenenschappen 53:1397–1412
Ulam SM (1961) Monte Carlo calculations in problems of mathematical physics. In: Beckenbach F (ed) Modern mathematics for the engineer: second series. McGraw-Hill, New York
Wanniski J (1978) Taxes, revenues, and the ‘Laffer curve’. Public Interest 50:3–16
Whitfield JW (1950) Uses of the ranking method in psychology. J R Stat Soc Ser B 12:163–170
Yerkes RM, Dodson JD (1908) The relation of strength of stimulus to rapidity of habit-formation. J Comp Neurol Psychol 18:459–482
Zayed H, Quade D (1997) On the resistance of rank correlation. J Stat Comput Simul 48:59–81
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by: Michel van de Velden.
About this article
Cite this article
Tarsitano, A., Amerise, I.L. Effectiveness of rank correlations in curvilinear relationships. Behaviormetrika 44, 351–368 (2017). https://doi.org/10.1007/s41237-017-0020-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s41237-017-0020-1