Approximation Theorems of Mathematical Statistics ROBERT J. SERFLING Professor of Mathematical Sciences The Johns Hopkins University Jobn Wiley & Sons [New York * Chichester Brisbane + Toronto + SingaporeCopyright © 1980 by Jobe Wey & Son, ae Reproduction oF trantion of any ptf this work eyon? that pert by Sections 107 oF 108 of he 1996 Unite Sates Copyright Act without the permission the copys omer nual Request for fermicson ofthe information shoud be ares to the Permisions Department, Jot Wey & Sons. Ie Liters of Congres Cataloging in Pablcaton Date: Sexi. RI, ‘Appronimation theorems of mathemati statis, (Wy series in protaiiy and mathematical tats) Includes biographical efeenes and indexes Mathematical static. 2 Lin theorems Probab theory) Tie arTests sto gnsaes ISeNna7ianas1 To my parents and 10 the memory of my wife's parentsPreface ‘This book covers a broad range of limit theorems useful in mathernatical statistics, alang with methods af proof and techniques of application, The manipulation of “probability” theorems to obtain “statistical” theorems is emphasized. It is hoped that, besides a knowledge ofthese basic statistical ‘theorems. an appreciation oa the instrumental role of probability theory and ‘a perspestive on practical needs for its further development may be gained. ‘A one-semester course each on probability theory and mathematical statistics at the beginning graduate level is presupposed. However, highly polished expertise is not necessary, the treatment here being self-contained. at an clementary level. The content is readily accessible to students in statistics, general mathematics, operations research, and selected engineering fields. ‘Chapter I lays out a variety of tools and foundations basic to asymptotic theory in statistics as treated in this book. Foremost are: modes of conver gence of a sequence of random variables (convergence in distribution, con- vergence in probability, convergence almost surely, and convergence in the rth mean); probability limit laws (the law of large numbers, the central limit theorem, and related results), ‘Chapter 2 deals systematically with the usual statistics computed from a sample: the sample distribution function, the sample moments, the sample ‘quantiles, the order statistics, and cell frequency vectors. Properties such as ‘asymptotic normality and almost sure convergence are derived, Also, deeper insights ore pursued, inckiding R. R. Bahadur’ fruitfal almost sure repre- sentations for sample quantiles and order statistics, Building on the results of Chapter 2, Chapter 3 treats the asymptotics of statistics concocted as transformations of vectors of more basic statistics. Typical examples are the sample coefficient of variation and the chi-squated statistic. Taylor seties approximations play a key role in the methodotogy. ‘The next six chapters deal with important special clases of statistics. ‘Chapter 4 concerns statistics arising in classical parametric inference and ‘contingency table analysis. These include maximum likelihood estimates,