Date of this Version
Journal of the Royal Statistical Society: Series B
Cox and Wermuth proposed that the partial derivative of the conditional distribution function of a random variable Y given another X is used for measuring association between two variables with arbitrary distributions. This paper shows the condition for collapsibility of the association measure.
This is the peer reviewed version of the following article: Ma, Z., Xie, X., Geng, Z., Uniform Collapsibility of Distribution Dependence Over a Nominal, Ordinal or Continuous Background, which has been published in final form at http://onlinelibrary.wiley.com/doi/10.1111/j.1467-9868.2005.00536.x/abstract. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Self-Archiving [link to http://olabout.wiley.com/WileyCDA/Section/id-820227.html#terms].
collapsibility, distribution dependence, Yule-Simpson paradox
Ma, Z., Xie, X., & Geng, Z. (2006). Uniform Collapsibility of Distribution Dependence Over a Nominal, Ordinal or Continuous Background. Journal of the Royal Statistical Society: Series B, 68 (1), 127-133. http://dx.doi.org/10.1111/j.1467-9868.2005.00536.x
Date Posted: 27 November 2017
This document has been peer reviewed.