Stein's method and the zero bias transformation with application to simple random sampling (Q1379715)

Let \(W\) be a random variable with mean \(0\) and variance \(\sigma^2\). Then there exists a unique distribution \(\mu\) such that \(\sigma^{-2}{\mathbf E}\{Wf(W)\}= \mu(f')\) for all smooth bounded functions \(f\), the zero bias transformation of \({\mathcal L}(W)\). The transformation shares many properties with the size bias transformation of nonnegative random variables, but is also applicable to distributions on the whole real line. Its use in conjunction with Stein's method is examined.