This book is dedicated to the memory of
Martin and Linda Gardner.
Chapter 18 Martin J Gardner 1989, 2000
CIA Software Trevor N Bryant 2000
All else British Medical Journal 1989, 2000
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording and/or otherwise, without the prior written permission of the publishers.
First published 1989
Second edition 2000
6 2011
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British library.
ISBN 978 0 7279 1375 3
Contributors
Douglas G Altman,Director, Imperial Cancer Research Fund Medical Statistics Group and Centre for Statistics in Medicine, Institute of Health Sciences, Oxford
Trevor N Bryant,Senior Lecturer in Bio computation, Medical Statistics and Computing (University of Southampton), Southampton General Hospital, Southampton
Michael J Campbell,Professor of Medical Statistics, Institute of Primary Care, University of Sheffield, Northern General Hospital, Sheffield
Leslie E Daly,Associate Professor of Public Health Medicine and Epidemiology, University College Dublin, Ireland
Martin J Gardner,former Professor of Medical Statistics, MRC Environmental Epidemiology Unit (University of Southampton), Southampton General Hospital, Southampton
Sheila M Gore,Senior Medical Statistician, MRC Biostatistics Unit, Cambridge
David Machin,Director, National Medical Research Council Clinical Trials and Epidemiology Research Unit, Ministry of Health, Singapore
Julie A Morris,Medical Statistician, Department of Medical Statistics, Withington Hospital, West Didsbury, Manchester
Robert G Newcombe,Senior Lecturer in Medical Statistics, University of Wales College of Medicine, Cardiff
Stuart J Pocock,Professsor of Medical Statistics, London School of Hygiene and Tropical Medicine, London
Now Professor of Clinical Trials Research, University of Sheffield
Source of contents
INTRODUCTION Specially written for second edition
PART I Estimation and confidence intervals
1 Gardner MJ, Altman DG. Estimating with confidence. BMJ 1988;:12101 (revised)
2 Specially written for second edition
3 Gardner MJ, Altman DG. Confidence intervals rather than P values: estimation rather than hypothesis testing. BMJ 1986;:74650 (revised)
4 From appendix 2 of source reference to chapter 3 (revised and expanded)
5 Campbell MJ, Gardner MJ. Calculating confidence intervals for some non-parametric analyses. BMJ 1988;:14546 (revised and expanded)
6 Specially written for second edition (replaces chapter 4 of first edition)
7 Morris JA, Gardner MJ. Calculating confidence intervals for relative risks (odds ratios) and standardised ratios and rates. BMJ 1988;:13136 (revised and expanded)
8 Altman DG, Gardner MJ. Calculating confidence intervals for regression and correlation. BMJ 1988;:123842 (revised and expanded)
9 Machin D, Gardner MJ. Calculating confidence intervals for survival time analyses. BMJ 1988;:136971 (revised and expanded)
10 Specially written for second edition
11 Specially written for second edition. Includes material from Altman DG. Confidence intervals for the number needed to treat. BMJ 1998;:130912.
12 Daly LE. Confidence intervals and sample sizes: dont throw out all your old sample size tables. BMJ 1991;:3336 (revised)
13 Specially written for second edition
PART II Statistical guidelines and checklists
14 Altman DG, Gore SM, Gardner MJ, Pocock SJ. Statistical guidelines for contributors to medical journals. BMJ 1983;:148993 (revised)
15 Gardner MJ, Machin D, Campbell MJ. Use of checklists in assessing the statistical content of medical studies. BMJ 1986;:8102 (revised and expanded)
PART III Notation, software and tables
16 Specially written for this book. Minor revisions for second edition
17 Specially written for second edition
18 Specially prepared for this book. Minor revisions for second edition
Introduction
DOUGLAS G ALTMAN, DAVID MACHIN, TREVOR N BRYANT
In preparing a new edition of a book, the editors are usually happy in the knowledge that the first edition has been a success. In the current circumstances, this satisfaction is tinged with deep personal regret that Martin Gardner, the originator of the idea for Statistics with Confidence, died in 1993 aged just 52. His achievements in a prematurely shortened career were outlined in his obituary in the BMJ.1
The first edition of Statistics with Confidence (1989) was essentially a collection of expository articles concerned with confidence intervals and statistical guidelines that had been published in the BMJ over the period 1986 to 1988. All were coauthored by Martin. The other contributors were Douglas Altman, Michael Campbell, Sheila Gore, David Machin, Julie Morris and Stuart Pocock. The whole book was translated into Italian2 and the statistical guidelines have also appeared in Spanish.3
As may be expected, several developments have occurred since the publication of the first edition and Martin had discussed and agreed some of the changes that we have now introduced into this new and expanded edition. Notably, this second edition includes new chapters on Diagnostic tests (chapter 10); Clinical trials and meta-analyses (chapter 11); Confidence intervals and sample sizes (chapter 12); and Special topics (substitution method, exact and mid-P confidence intervals, bootstrap confidence intervals, and multiple comparisons) (chapter 13). There is also a review of the impact of confidence intervals in the medical literature over the ten years or so since the first edition (chapter 2). All the chapters from the first edition have been revised, some extensively, and one (chapter 6 on proportions) has been completely rewritten. The list of contributors has been extended to include Leslie Daly and Robert Newcombe. We are grateful to readers of the first edition for constructive comments which have assisted us in preparing this revision.
Alongside the first edition of Statistics with Confidence, a computer program, Confidence Interval Analysis (CIA), was available. This program, which could carry out the calculations described in the book, had been written by Martin, his son Stephen Gardner and Paul Winter. An entirely new Windows version of CIA has been written by Trevor Bryant to accompany the book, and is packaged with this second edition. It is outlined in chapter 17. The program reflects the changes made for this edition of the book and has been influenced by suggestions from users.
Despite the enhanced coverage we would reiterate the comment in the introduction to the first edition, that this book is not intended as a comprehensive statistical textbook. For further details of statistical methods the reader is referred to other sources.47
We were all privileged to be colleagues of Martin Gardner. We hope that he would have approved of this new edition of
Next page