Postprint version. Published in Stochastic Analysis and Applications, Volume 18, Issue 1, January 1, 2000, pages 87-99.
Copyright © 2000 Taylor & Francis. This is an electronic version of an article published in Stochastic Analysis and Applications. The definitive version is available at http://dx.doi.org/10.1080/07362990008809656.
NOTE: At the time of publication, the author Theodore P. Hill was not yet affiliated with Cal Poly.
A class of empirical distributions is introduced which are based on various weighted linear combinations of order statistics, and which have convergence properties the classical empirical distribution does not, or which stochastically or convexly dominate the classical empirical distribution