Exploring Data Tables, Trends, and Shapes

WILEY-INTERSCIENCE PAPERBACK SERIES The Wiley-Interscience Paperback Series consists of selected books that have been made more accessible to consumers in an effort to increase global appeal and general circulation. With these new unabridged softcover volumes, Wiley hopes to extend the lives of these works by making them available to future generations of statisticians, mathematicians, and scientists. "e;Exploring Data Tables, Trends, and Shapes (EDTTS) was written as a companion volume to the same editors' book, Understanding Robust and Exploratory Data Analysis (UREDA). Whereas UREDA is a collection of exploratory and resistant methods of estimation and display, EDTTS goes a step further, describing multivariate and more complicated techniques . . . I feel that the authors have made a very significant contribution in the area of multivariate nonparametric methods. This book [is] a valuable source of reference to researchers in the area."e; Technometrics "e;This edited volume . . . provides an important theoretical and philosophical extension to the currently popular statistical area of Exploratory Data Analysis, which seeks to reveal structure, or simple descriptions, in data . . . It is . . . an important reference volume which any statistical library should consider seriously."e; The Statistician This newly available and affordably priced paperback version of Exploring Data Tables, Trends, and Shapes presents major advances in exploratory data analysis and robust regression methods and explains the techniques, relating them to classical methods. The book addresses the role of exploratory and robust techniques in the overall data-analytic enterprise, and it also presents new methods such as fitting by organized comparisons using the square combining table and identifying extreme cells in a sizable contingency table with probabilistic and exploratory approaches. The book features a chapter on using robust regression in less technical language than available elsewhere. Conceptual support for each technique is also provided.

WILEY-INTERSCIENCE PAPERBACK SERIESThe Wiley-Interscience Paperback Series consists ofselected books that have been made more accessible to consumers inan effort to increase global appeal and general circulation. Withthese new unabridged softcover volumes, Wiley hopes to extend thelives of these works by making them available to future generationsof statisticians, mathematicians, and scientists."Exploring Data Tables, Trends, and Shapes (EDTTS) waswritten as a companion volume to the same editors' bookUnderstanding Robust and Exploratory Data Analysis (UREDA). WhereasUREDA is a collection of exploratory and resistant methods ofestimation and display, EDTTS goes a step further, describingmultivariate and more complicated techniques . . . I feel that theauthors have made a very significant contribution in the area ofmultivariate nonparametric methods. This book [is] a valuablesource of reference to researchers in the area."Technometrics"This edited volume . . . provides an important theoretical andphilosophical extension to the currently popular statistical areaof Exploratory Data Analysis, which seeks to reveal structure, orsimple descriptions, in data . . . It is . . . an importantreference volume which any statistical library should considerseriously."The StatisticianThis newly available and affordably priced paperback version ofExploring Data Tables, Trends, and Shapes presents majoradvances in exploratory data analysis and robust regression methodsand explains the techniques, relating them to classical methods.The book addresses the role of exploratory and robust techniques inthe overall data-analytic enterprise, and it also presents newmethods such as fitting by organized comparisons using the squarecombining table and identifying extreme cells in a sizablecontingency table with probabilistic and exploratory approaches.The book features a chapter on using robust regression in lesstechnical language than available elsewhere. Conceptual support foreach technique is also provided.

1. Theories of Data Analysis: From Magical Thinking Through Classical Statistics 1Peris Diaconis1A. Intuitive Statistics--Some Inferential Problems 4IB. Multiplicity--A Pervasive Problem 91C. Some Remedies 12ID. Theories for Data Analysis 22IE. Uses for Mathematics 29IF. In Defense of Controlled Magical Thinking 312. Fitting by Organized Comparisons: The Square Combining Table 37Katherine Godfrey2A. Combining Comparisons 372B. Two-Way Tables 392C. Paired Comparisons 472D. Analyzing Tables Containing Holes 492E. Summary 613. Resistant Nonadditive Fits for Two-Way Tables 67John D. Emerson and Gregory Y. Wong3A. The Simple Additive Model and Median Polish 683B. One Step Beyond an Additive Fit 713C. Assessing and Comparing Fits 793D. Multiplicative Fits 833E. Techniques for Obtaining Simple Multiplicative Fits 923F. Additive-Plus-Multiplicative Fits 1003G. Some Background for Nonadditive Fits 1133H. Summary 1174. Three-Way Analysis 125Nancy Cook4A. Structure of the Three-Way Table 1264B. Decompositions and Models for Three-Way Analysis 1284C. Median-Polish Analysis for the Main-Effects-Only Case 1304D. Nonadditivity and a Diagnostic Plot in Main-Effects-Only Analysis 1454E. Analysis Using Means 1584F. Median-Polish Analysis for the Full-Effects Case 1644G. Diagnostic Plots for the Full-Effects Case 1764H. Fitting the Full-Effects Model by Means 1804I. Computation, Other Polishes, and Missing Values 1824J. Summary 1835. Identifying Extreme Cells in a Sizable Contingency Table: Probabilistic and Exploratory Approaches 189Frederick Mosteller and Anita Parunak5A. The Hypergeometric Distribution 1925B. Assessing Outliers 1955C. The Simulation Approach 1995D. Applying the Simulation Approach to the Table of Archaeological Data 2065E. An Exploratory Approach, Based on Deviations from Independence 2125F. A Logarithmic Exploratory Approach 2145G. Illustrations of the New Standardization 2175H. Summary 22151. Conclusion 2236. Fitting Straight Lines By Eye 225Frederick Mosteller, Andrew F. Siegel, Edward Trapido, and Cleo Youtz6A. Method 2266B. Results 2296C. Summary 2387. Resistant Multiple Regression, One Variable at a Time 241John D. Emerson and David C. Hoaglin7A. Resistant Lines 2427B. Sweeping Out 2467C. Example 2507D. When Carriers Come in Blocks 2637E. Summary 2738. Robust Regression 281Guoying Li8A. Why Robust Regression? 2828B. M-Estimators and W-Estimators for Regression 2918C. Computation 3048D. Example: The Stack Loss Data 3108E. Bounded-Influence Regression 3228F. Some Alternative Methods 3288G. Summary 3359. Checking the Shape of Discrete Distributions 345David C. Hoaglin and John W. Tukey9A. A Poissonness Plot 3489B. Confidence Intervals for the Count Metameter 3589C. When Is a Point Discrepant? 3709D. Overall Plots for Other Families of Distributions 3769E. Frequency-Ratio Alternatives 3899F. Cooperative Diversity 3969G. Double-Root Residuals 4069H. Summary 40910. Using Quantiles to Study Shape 417David C. Hoaglin10A. Diagnosing Skewness 41910B. Diagnosing Elongation 425IOC. Quantile-Quantile Plots 43210D. Plots for Skewness and Elongation 44210E. Pushback Analysis 45010F. Summary 45410G. Appendix 45611. Summarizing Shape Numerically: The g-and-h Distributions 416David C. Hoaglin11 A. Skewness 46211B. Elongation 47911C. Combining Skewness and Elongation 48511D. More General Patterns of Skewness and Elongation 490HE. Working from Frequency Distributions 49611F. Moments 50111G. Other Approaches to Shape 50411H. Summary 508Index.