Observed Confidence Levels: Theory and Application

Alan M. Polansky

October 26, 2007 by Chapman and Hall/CRC
Reference - 288 Pages - 50 B/W Illustrations
ISBN 9781584888024 - CAT# C8024


Add to Wish List
FREE Standard Shipping!


  • Compares the use of observed confidence levels with other statistical methodologies such as multiple comparisons and posterior probabilities
  • Includes many practical examples from the fields of engineering, biology, and medicine as well as exercises at the end of each chapter
  • Discusses how observed confidence levels can be computed using the R statistical computing environment
  • Offers a website with many of the data sets and R code used in the book as well as links to relevant references and news about current applications of observed confidence levels
  • Summary

    Illustrating a simple, novel method for solving an array of statistical problems, Observed Confidence Levels: Theory and Application describes the basic development of observed confidence levels, a methodology that can be applied to a variety of common multiple testing problems in statistical inference. It focuses on the modern nonparametric framework of bootstrap-based estimates, allowing for substantial theoretical development and for relatively simple solutions to numerous interesting problems.

    After an introduction, the book develops the theory and application of observed confidence levels for general scalar parameters, vector parameters, and linear models. It then examines nonparametric problems often associated with smoothing methods, including nonparametric density estimation and regression. The author also describes applications in generalized linear models, classical nonparametric statistics, multivariate analysis, and survival analysis as well as compares the method of observed confidence levels to hypothesis testing, multiple comparisons, and Bayesian posterior probabilities. In addition, the appendix presents some background material on the asymptotic expansion theory used in the book.

    Helping you choose the most reliable method for a variety of problems, this book shows how observed confidence levels provide useful information on the relative truth of hypotheses in multiple testing problems.