An extension of the Stein/Malliavin-Stein method to independent random variables

Abstract: Stein's method is a powerful method for characterizing laws of random variables and deriving estimates on the distance in law between a given random variable and a target random variable. In recent work, Stein's method for normal distributions has been extended to provide a characterization of normality together with independence with respect to an auxiliary random variable. We provide here an extension of Stein's method to include independence with respect to an auxiliary random variable, for any law for which a Stein characterization does exist.