APPLYING STATISTICS IN THE COURTROOM A New Approach for Attorneys and Expert Witnesses Copyright 2001 by Chapman & Hall/CRC CRC Press LLC APPLYING STATISTICS IN THE COURTROOM A New Approach for Attorneys and Expert Witnesses Phillip I. Good CHAPMAN & HALL/CRC Boca Raton London New York Washington, D.C. Copyright 2001 by Chapman & Hall/CRC CRC Press LLC Library of Congress Cataloging-in-Publication Data Good, Phillip I. Applying statistics in the courtroom : a new approach for attorneys and expert witnesses / Phillip I. Good. p. cm. Includes bibliographical references and index. ISBN 1-58488-271-9 (alk. paper) 1. Forensic statistics—United States. I. Title. KF8968.75 .G66 2001 347.73′67—dc21 2001028695 Catalog record is available from the Library of Congress This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials or for the consequences of their use. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming, and recording, or by any information storage or retrieval system, without prior permission in writing from the publisher. All rights reserved. Authorization to photocopy items for internal or personal use, or the personal or internal use of specific clients, may be granted by CRC Press LLC, provided that $1.50 per page photocopied is paid directly to Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923 USA. The fee code for users of the Transactional Reporting Service is ISBN 1-58488-271- 9/01/$0.00+$1.50. The fee is subject to change without notice. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. The consent of CRC Press LLC does not extend to copying for general distribution, for promotion, for creating new works, or for resale. Specific permission must be obtained in writing from CRC Press LLC for such copying. Direct all inquiries to CRC Press LLC, 2000 N.W. Corporate Blvd., Boca Raton, Florida 33431. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation, without intent to infringe. Visit the CRC Press Auerbach Publications Web site at www.crcpressauerbach- publications.com © 2001 by Chapman & Hall/CRC CRC Press LLC St. Lucie Press Lewis Publishers Auerbach is an imprint of CRC Press LLC No claim to original U.S. Government works International Standard Book Number 1-58488-271-9 Library of Congress Card Number 2001028695 Printed in the United States of America 1 2 3 4 5 6 7 8 9 0 Printed on acid-free paper Copyright 2001 by Chapman & Hall/CRC CRC Press LLC Preface For the rational study of the law … the man of the future is the man of statistics and the master of economics. Oliver Wendell Holmes, 1897 We have written this text for two audiences with a single common goal: to ensure attorneys and statisticians will work together successfully on the application of statistics in the law. To be effective in the courtroom, a statistician must be able to think like a lawyer and present complex statistical concepts in terms a judge can understand. Thus, we present the principles of statistics and probability, not as a series of symbols, but in the words of jurists. Concepts take precedence over formula. Statisticians who may be skeptical of this approach should remember the law is what judges say it is. For the attorney, a change in point of view is also necessary. West has no categories labeled “cohort analysis” or “sample size.” Sheppardize, but do not expect to find cases that uphold or reverse the statistical points raised. The opinions in this text are included for any or all of the following reasons: (cid:1) They provide concise definitions of statistical concepts. (cid:1) They provide lucid explanations of these concepts. (cid:1) They illustrate the presentation of statistical evidence. (cid:1) They reveal grounds on which a statistical argument may be successfully attacked. Two further caveats: our coverage is not meant to be complete. If we write, “two other circuits have approved the use of statistical sampling,” Copyright 2001 by Chapman & Hall/CRC CRC Press LLC we do not mean to imply that other circuits have not approved it. Such statements imply only that our research has been limited. Finally, we have focused solely on the statistical issues in given cases. Many other issues are involved in almost every instance and, frequently, these issues, not the statistical ones, have formed the basis for a court’s ultimate determination. Hopefully, this text will serve the attorney as a comprehensive guide to the application of statistics and probability in jury selection, employment discrimination, trademark disputes, criminal law, civil law, and product liability. Sidebars and Chapters 13 and 15 indicate issues to raise during discovery and possible lines of counterattack, as well as potential areas of vulnerability. This book is divided into four parts. The opening chapters concern the relationship between a sample and the population from which it is drawn. Chapter 1 describes the courts’ gradual acceptance of samples and sampling methodology. Chapter 2 defines the representative random sam- ple. Chapter 3 compares various sampling methodologies. Chapter 4 is devoted to the use of descriptive statistics in the courtroom including measures of central tendency, precision, and percentage. Chapter 5 provides a brief introduction to probability. Chapters 6, 7, and 8 describe the varying acceptance of probability-based testimony in civil, criminal, and environmental hazard cases, respectively. Chapter 9 summarizes the courts’ responses to how large a sample must be. Chapter 10 addresses the same topic from the statistician’s point of view and describes some simple procedures for testing sta- tistical hypotheses. Chapter 11 describes correlation and regression of two variables. Chapter 12 extends this discussion to multiple variables and shows how the courts have applied multiple regression methods in cases of alleged discrimination. Chapter 13 is devoted to preventive actions that can be taken to stay out of the courtroom and discusses the challenges that can be made to bad statistics once inside. Chapter 14 prepares the statistician for some of the twists and turns the trial process can take. Chapter 15 describes how an attorney can make the most effective use of statistics and statisticians at various points in the trial process and provides a set of questions on data-related concerns for use during discovery. I would like to express my gratitude to the following libraries for the use of their facilities: the Orange County Law Library, the University of San Diego Law Library, and the Whittier Law School Library. Without these excellent community resources, this text would have been impossible. Copyright 2001 by Chapman & Hall/CRC CRC Press LLC Interpreting Case Citations For those unfamiliar with legal research, here is a brief explanation of case citations such as U.S. v. Two Bulls, 918 F.2d 56, 61 (8th Cir. 1990). This citation refers to a decision handed down in 1990 by the court of appeals for the Eighth Circuit which embraces the Dakotas, Nebraska, Minnesota, Iowa, Missouri, and Arkansas. The appeal was brought by federal prosecutors (the U.S.) unhappy with the results of the earlier trial of Two Bulls (a Navaho Indian). This decision was reported in Vol- ume 918 of the second series of the Federal Reporter beginning on page 56; the cited quotation will be found on page 61. Copyright 2001 by Chapman & Hall/CRC CRC Press LLC Prologue: An Introduction to the U.S. Court System for the Statistician The purpose of this prologue is to describe for the statistician the court system in the U.S. and the roles the statistician will be expected to play. The majority of the statistician’s participation will take place behind the scenes as an advisor to the attorney. He or she may provide an estimate or a significance level and be asked to comment on and find weaknesses in the statistical procedures the attorney’s opponent has used or is likely to use. The statistician’s work product may appear in court as a portion of an attorney’s brief (written motion), or it may never appear at all. A trial is preceded by a series of pretrial motions and a process known as discovery, in which depositions are taken and the facts of the case are more or less agreed upon. The purpose of discovery is to reduce the time actually spent in the courtroom and, hopefully, to bring both parties to a settlement without having to hold a trial. Consider yourself a success if your efforts as a statistician contribute to a pretrial settlement.1 As a statistician, you may be asked to provide testimony under oath in the form of a deposition. This testimony will not be given in court and may be taken in an attorney’s office or on neutral ground. You may be asked to testify by the attorney who hired you originally or by the attorney 1 Assuming, of course, you don’t persuade the attorney you work for that he or she has no case when, in fact, the case is an excellent one. Copyright 2001 by Chapman & Hall/CRC CRC Press LLC who is his or her opponent. In either case, you may be subject to a detailed (and unpleasant) examination or cross-examination by the oppos- ing attorney. You can get revenge of a sort by counseling your attorney thoroughly so that he or she can subject the opposing party’s statistician to an equally unpleasant afternoon of questions. In a third, relatively rare alternative, you may be summoned and employed by a judge to act as an impartial expert witness. Many cases go to trial in a state or a federal court. During the trial, the judge may be asked to rule on various points of law that arise. An example would be to decide whether the results of a survey may be introduced in court. To guide his or her ruling, the judge may review rules that his or her own and higher courts have published, the published opinions of appeals and higher courts (common law), statutes (laws passed by legis- latures), and other sources of information he or she considers helpful. A ruling made by a federal district court may be reviewed by a federal circuit appeals court, and the appeals court ruling may be further reviewed by the U.S. Supreme Court. Not all cases are subject to review; the Supreme Court reviews only a handful of the appeals that come to it. A ruling by the Supreme Court is binding on all federal courts. A ruling by a circuit appeals court is binding only on the courts within its circuit, and its ruling is subordinate to whatever the Supreme Court later decides. A ruling by a state court may be reviewed by a state appeals court and (depending on the state) may be reviewed subsequently by still higher level courts in that state. A ruling by the highest state court is binding on all the courts within that state, but it is not binding on the courts in other states.2 Nor is a ruling in a federal court binding on any of the state courts.3 Decisions by state trial courts are almost never published and thus cannot be cited by attorneys. Some federal trial court opinions are pub- lished; most are not. Almost all appeals court and higher court opinions are published. Trial court judges look to those opinions for guidance, and we have looked to them for material on the application of statistics in the courtroom. 2 Nonetheless, a judge in one state might well be interested in what the appeals courts in other states have decided if a ruling on a particular point of law has not yet been made in his or her own state. 3 The exception is an appeal of a state court decision based on an alleged violation of federal law. An example would be the overturning by a federal court of a decision in a state that did not allow African-Americans to serve on its juries. Copyright 2001 by Chapman & Hall/CRC CRC Press LLC The Author Phillip Good, Ph.D., a graduate of the University of California at Berkeley’s program in mathematical statistics and a paralegal, is the author of 12 published books including Permutation Tests (Springer, 1994, second edition, 2000), and Resampling Methods (Birkhauser, 1999, second edition, 2001), as well as 34 scholarly publications, 12 published short stories, and over 600 popular articles. A former Calloway Professor of Computer Science at the University of Georgia, Division Head and Professor of Biology and Physics at West Coast University, and Associate Professor of Applied Mathematics at Claremont College, he has worked in the aero- space, computer, energy, marketing, pharmaceutical, and telecommunica- tion industries. He has worked with attorneys on cases involving audits, jury selection, personal injury, and product liability. Copyright 2001 by Chapman & Hall/CRC CRC Press LLC Contents PART I: SAMPLES AND POPULATIONS 1 Samples and Populations 1.1 Audits 1.1.1 Validity of Using Sample Methods 1.1.2 Basis for Objection 1.1.3 Is the Sample Size Adequate? 1.2 Determining the Appropriate Population 1.2.1 Jury Panels 1.2.2 Criminal Universe 1.2.3 Trademarks 1.2.4 Discrimination 1.2.5 Downsizing 1.3 Summary 2 Representative Samples and Jury Selection 2.1 Concepts 2.2 Issues 2.2.1 Burden of Proof 2.2.2 The Right to be Eligible to Serve 2.2.3 Cognizable, Separate, Identifiable Groups 2.2.4 Voir Dire Rights of Litigants and Jurors 2.3 Composition of the Jury Pool 2.3.1 True Cross-Section 2.3.2 Snapshot in Time 2.3.3 Composition of the Individual Panel 2.3.4 Standing 2.4 Random Selection 2.4.1 Errors in Sampling Methodology 2.5 Summary 2.6 To Learn More Copyright 2001 by Chapman & Hall/CRC CRC Press LLC