We consider a bivariate normal distribution with linear correlation [Formula: see text] whose random components are discretized according to two assigned sets of thresholds. On the resulting bivariate ordinal random variable, one can compute Goodman and Kruskal's gamma coefficient, [Formula: see text], which is a common measure of ordinal association. Given the known analytical monotonic relationship between Pearson's [Formula: see text] and Kendall's rank correlation [Formula: see text] for the bivariate normal distribution, and since in the continuous case, Kendall's [Formula: see text] coincides with Goodman and Kruskal's [Formula: see text], the change of this association measure before and after discretization is worth studying. We consider several experimental settings obtained by varying the two sets of thresholds, or, equivalently, the marginal distributions of the final ordinal variables. This study, confirming previous findings, shows how the gamma coefficient is always larger in absolute value than Kendall's rank correlation; this discrepancy lessens when the number of categories increases or, given the same number of categories, when using equally probable categories. Based on these results, a proposal is suggested to build a bivariate ordinal variable with assigned margins and Goodman and Kruskal's [Formula: see text] by ordinalizing a bivariate normal distribution. Illustrative examples employing artificial and real data are provided.
Goodman and Kruskal's Gamma Coefficient for Ordinalized Bivariate Normal Distributions / A. Barbiero, A. Hitaj. - In: PSYCHOMETRIKA. - ISSN 0033-3123. - 85:4(2020 Dec), pp. 905-925. [10.1007/s11336-020-09730-5]
Goodman and Kruskal's Gamma Coefficient for Ordinalized Bivariate Normal Distributions
A. Barbiero
Primo
;
2020
Abstract
We consider a bivariate normal distribution with linear correlation [Formula: see text] whose random components are discretized according to two assigned sets of thresholds. On the resulting bivariate ordinal random variable, one can compute Goodman and Kruskal's gamma coefficient, [Formula: see text], which is a common measure of ordinal association. Given the known analytical monotonic relationship between Pearson's [Formula: see text] and Kendall's rank correlation [Formula: see text] for the bivariate normal distribution, and since in the continuous case, Kendall's [Formula: see text] coincides with Goodman and Kruskal's [Formula: see text], the change of this association measure before and after discretization is worth studying. We consider several experimental settings obtained by varying the two sets of thresholds, or, equivalently, the marginal distributions of the final ordinal variables. This study, confirming previous findings, shows how the gamma coefficient is always larger in absolute value than Kendall's rank correlation; this discrepancy lessens when the number of categories increases or, given the same number of categories, when using equally probable categories. Based on these results, a proposal is suggested to build a bivariate ordinal variable with assigned margins and Goodman and Kruskal's [Formula: see text] by ordinalizing a bivariate normal distribution. Illustrative examples employing artificial and real data are provided.File | Dimensione | Formato | |
---|---|---|---|
manuscript.pdf
accesso aperto
Descrizione: articolo principale
Tipologia:
Pre-print (manoscritto inviato all'editore)
Dimensione
312.76 kB
Formato
Adobe PDF
|
312.76 kB | Adobe PDF | Visualizza/Apri |
Barbiero_et_al-2020-Psychometrika.pdf
accesso aperto
Descrizione: articolo principale online first
Tipologia:
Publisher's version/PDF
Dimensione
439.36 kB
Formato
Adobe PDF
|
439.36 kB | Adobe PDF | Visualizza/Apri |
Barbiero-Hitaj2020_Article_GoodmanAndKruskalSGammaCoeffic.pdf
accesso aperto
Tipologia:
Publisher's version/PDF
Dimensione
484.38 kB
Formato
Adobe PDF
|
484.38 kB | Adobe PDF | Visualizza/Apri |
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.