SIXTH EDITION A NEW EDITION OF THE CLASSIC WORkK IN THE FIELD OF BA DIALS TICS (4 W 7B 1994 Digitized by the Internet Archive In 2021 with funding from Kahle/Austin Foundation https://archive.org/details/sayitwithfiguresO000Zeis SAY IT WITH FIGURES ABOUT THE AUTHOR Hans Zeisel is Professor of Law and Sociology Emeritus at the University of Chicago where he pioneered the application of social science to law. Earlier he had a distinguished career in public opinion and market research; in 1980 he was inducted into the Market Research Hall of Fame. He has written on a wide variety of topics, ranging from research methodology and history to law enforcement, juries, and Shakes- peare. He has been elected fellow of the Amer- ican Statistical Association and of the American Association of the Advancement of Science. SAY IT WITH FIGURES SIXTH EDITION HANS ZEISEL r 1817 HARPER & ROW, PUBLISHERS, New York Cambridge, Philadelphia, San Francisco, London, Mexico City, SAo Paulo, Singapore, Sydney To the Memory of Paul Felix Lazarsfeld SAY IT WITH FIGURES, (Sixth Edition). Copyright © 1985 by Harper & Row, Publishers, Inc. All rights reserved. Printed in the United States of America. No part of this book may be used or reproduced in any manner whatsoever without written permission except in the case of brief quotations embodied in critical articles and reviews. For information address Harper & Row, Publishers, Inc., 10 East 53rd Street, New York, N.Y. 10022. Published simultaneously in Canada by Fitzhenry & Whiteside Limited, Toronto. Designer: C. Linda Dingler a Library of Congress Cataloging in Publication Data Zeisel, Hans. Say it with figures. Includes index. 1. Social sciences—Statistical methods. 2. Statistics. I. Title. HA29.Z4 1985 001.4'22 84-48207 ISBN 0-06-181982-4 85 86 87 88 8910987654321 ISBN 0-06-131994-5 (pbk.) 85 86 87 88 8910987654321 Contents Preface Introduction to the Fifth Edition by Paul F. Lazarsfeld From the Introduction to the First Edition by Paul F. Lazarsfeld xiv PART |. THE PRESENTATION OF NUMBERS 1. The Function of Percent Figures Facilitating Comparisons 3 Deemphasizing the Numbers 6 Causal Implications 7 Percents and Percentage Points 9 WrongEnd Up 9 __ The Ceiling Effect 10 Summary 13 2. Presentation Problems 14 Numbers and Percents 14 Percents Running Over 100 15 Subtotals 16 Decimals 16 Percent, Per Mille, Per 100,000 18 Special Ratios 19 When Captions Need Explaining 21 The Percentage Chain 22 Of100 People 23 Graphs 24 Tables as Bar Graphs 24 Computations at a Glance 25 The “Gestalt” 25 One Picture Is Worth a Thousand Numbers 29 Summary 33 3. In Which Direction Should Percents Be Run? 34 The Cause-and-Effect Rule 34 The Ambiguity of “Cause” 36 Digging Out the Information 38 The Proviso of Representative Sampling 39 The Total Column 45 Summary 46 vi Contents How to Handle Don’t Knows and No Answers 48 The Legitimate Don’t Knows 48 The Failure Don’t Knows 49 What Not to Do with the Failure Don’t Knows 50 What to Do with the Failure Don’t Knows 51 Reducing the Numbers of Don’t Knows 53 Asking for Numbers 54 Indefinite Numerals 54 Don’t Knows with a Special Meaning 56 Facilitating Recali 58 Reducing the Number of Legitimate Don’t Knows 58 Statistical Lie Detection 59 Census Eliminates 207,000 Don’t Knows 62 Summary 63 Tables of More than Two Dimensions 65 The Problem of Reduction 65 The Principle Illustrated 68 Making a Dichotomy 69 An Average Representing the Column 71 Rank Orders 73 Reducing a Trichotomy 75 A Four-Dimensional Table a Summary 82 Indices 83 Judgment Indices 84 A Complex Average 86 Index Object and Index Formula 88 = Ambiguous Labels 89 Effect of “Aging” 90 Baseball Indices 91 Olympic Scoring 95 Interrelated Percentages 97 Sociometric Indices 100 Spearman Coefficient of Rank Correlation 104 | Custom-Made Indices 108 Summary 110 PART Il. THE TOOLS OF CAUSAL ANALYSIS 111 The Cross-Tabulation Refines 115 Purposes of Cross Tabulating 115 Different Types 117. _—‘T he Additional Factor Refines 118 Correlations near Zero 120 Additional Factor Reveals Limiting Conditions 123 Additional Factor Has an Independent Effect 124 Summary 126 Contents vil 8. Experimental Evidence mw 188 The Problem 128 The Controlled Randomized Experiment 130 The Miracle of Random Selection 131 Discrimination 133 Generalizing from an Experiment 134 Designating Experimental and Control Subjects 136 Statistical Error 137 The Natural Experiment 137 The Half-a-Loaf Experiment 139 Summary 142 Analysis of Nonexperimental Data 143 A Different Problem 143 Full Explanation 145 Partial Explanation 148 Spurious Correlations 151 Partly Spurious Correlation 152 The Correlation Is Reversed 155 Spurious Noncorrelation 156 True and Spurious Correlations 157 “Before” and “After” Comparison 159 Lateral Analysis 162 Summary 165 166 10. Regression Analysis The Scatter Diagram 166 The Correlation Coefficient 166 Regression toward the Mean 169 The Regression Fallacy 171 How Big a Change? 172 How Much Is Explained? 172 Multiple Regression 177. The Purpose of Regression Analysis 179 Causal Analysis of Observational Data 180 Traps 182 Summary 185 11. Reason Analysis |: The Accounting Scheme 186 The Art of Asking, “Why?” 186 Formulating the Problem 188 The Exploratory Interview 189 Developing the Accounting Scheme 191 The Push-and- Pull Model 194 Reason Assessment 195 Multidimensional Models 195 The Art of Asking, “Why Not?” 198 The Time Dimension 199 _ Precipitating Events 200 Phases of Decision 201 Narrowing the Choice 202 Summary 203 viii Contents 12. Reason Analysis II: Data Collection and Interpretation 204 Accounting Scheme and Questionnaire 204 Probing 205 Verifying Answers 207 How Far to Search? 208 Primary and Secondary Reasons 210 Summary 215 13. The Panel 216 Overview 216 Concepts Covering Time 218 Shifts and Changes 218 Turnover and Net Change 221 Multiple Shifts 223 Shifts at Several Points in Time 224 Operational Analysis 226 Measuring Loyalty 228 Who Shifted and Why 233 The Bandwagon Effect 235 Effects of Advertising 236 Reversal of Cause and Effect 242 Bias from Prior Interviewing 243 ‘Panel Mortality 249 Summary 250 14. Triangulation 252 Origin of the Concept 252 Repeated Observations 253 Gauging Imperfect Samples 253 Triangulating Simulated Results 254 Experiences in Different Places 255 Cross Examination 257 Failure to Aggregate the Evidence 259 Correcting False Counts 259 The Conviction-Prone Jurors 261 Summary 262 Author Index 263 Subject Index 265