Moment (in-)equalities for differences of order statistics with different sample sizes (Q1082007)

Let \(X_{i:n}\) denote the \(i^{th}\) order statistic of a random sample of size n from a continuous distribution function F. Let \(q_{\alpha:n}(\equiv X_{[n\alpha]:n})\) and \(q_{\alpha}\) denote \(\alpha^{th}\) sample and population quantiles, respectively. Further let \(\| X\|\) denote \((E| X^ r|)^{1/r}\), \(r\geq 1\), for a random variable X. When F is uniform over (0,1) an expression for \(\| X_{i_ 1:n_ 1}-X_{i_ 2:n_ 2}\|\) is given in terms of hypergeometric probabilities. Using this it is shown under some mild conditions that for \(n_ 1<n_ 2\), \[ \| q_{\alpha:n_ 1}- q_{\alpha:n_ 2}\| \leq c(n_ 1^{-1}-n_ 2^{-1})^{1/2} \] and \[ \| n_ 1(q_{\alpha:n_ 1}-q_{\alpha})-n_ 2(q_{\alpha:n_ 2}-q_{\alpha})\| \leq c\sqrt{(n_ 2-n_ 1)}, \] where c depends on \(\alpha\), r and F.