Statistical Analysis and Data Display: An Intermediate Course with Examples in R (2nd Edition)

Statistical Analysis and Data Display An Intermediate Course with Examples in R – Heiberger – 2015



Series: Springer Texts in Statistics
Publisher: Springer; 2nd ed. 2015 edition (December 24, 2015)
Language: English
ISBN-10: 1493921215
ISBN-13: 978-1493921218

This contemporary presentation of statistical methods features extensive use of graphical displays for exploring data and for displaying the analysis.  The authors demonstrate how to analyze data―showing code, graphics, and accompanying tabular listings―for all the methods they cover. They emphasize how to construct and interpret graphs. They discuss principles of graphical design. They identify situations where visual impressions from graphs may need confirmation from traditional tabular results. All chapters have exercises.

The authors provide and discuss R functions for all the new graphical display formats. All graphs and tabular output in the book were constructed using these functions. Complete R scripts for all examples and figures are provided for readers to use as models for their own analyses.

This book can serve as a standalone text for statistics majors at the master’s level and for other quantitatively oriented disciplines at the doctoral level, and as a reference book for researchers. In-depth discussions of regression analysis, analysis of variance, and design of experiments are followed by introductions to analysis of discrete bivariate data, nonparametrics, logistic regression, and ARIMA time series modeling. The authors illustrate classical concepts and techniques with a variety of case studies using both newer graphical tools and traditional tabular displays.

The Second Edition features graphs that are completely redrawn using the more powerful graphics infrastructure provided by R’s lattice package. There are new sections in several of the chapters, revised sections in all chapters and several completely new appendices.

New graphical material includes:

• an expanded chapter on graphics

• a section on graphing Likert Scale Data to build on the importance of rating scales in fields from population studies to psychometrics

• a discussion on design of graphics that will work for readers with color-deficient vision

• an expanded discussion on the design of multi-panel graphics

• expanded and new sections in the discrete bivariate statistics capter on the use of mosaic plots for contingency tables including the n×2×2 tables for which the Mantel–Haenszel–Cochran test is appropriate

• an interactive (using the shiny package) presentation of the graphics for the normal and t-tables that is introduced early and used in many chapters

The new appendices include discussions of R, the HH package designed for R (the material in the HH package was distributed as a set of standalone functions with the First Edition of this book), the R Commander package, the RExcel system, the shiny package, and a minimal discussion on writing R packages. There is a new appendix on computational precision illustrating and explaining the FAQ (Frequently Asked Questions) about the differences between the familiar real number system and the less-familiar floating point system used in computers. The probability distributions appendix has been expanded to include more distributions (all the distributions in base R) and to include graphs of each. The editing appendix from the First Edition has been split into four expanded appendices―on working style, writing style, use of a powerful editor, and use of LaTeX for document preparation.

About the Author

Richard M. Heiberger is Professor Emeritus in the Department of Statistics of Temple University, an elected Fellow of the American Statistical Association, and a former Chair of the Section on Statistical Computing of the American Statistical Association.  He was Graduate Chair for the Department of Statistics and Acting Associate Vice Provost for the University.  He participated in the design of the linear model and analysis of variance functions while on research leave at Bell Labs.  He has taught short courses at the Joint Statistics Meetings, the American Statistical Association Conference on Statistical Practice, the R Users Conference, and the Deming Conference on Applied Statistics.  He has consulted with several pharmaceutical companies.

Burt Holland was Professor in the Department of Statistics of Temple University, an elected Fellow of the American Statistical Association, Chair of the Department of Statistics of Temple University, and Chair of Collegial Assembly of the Fox School.  He has taught short courses at the Joint Statistics Meetings and the Deming Conference on Applied Statistics.  He has made many contributions to linear modeling and simultaneous statistical inference.  He frequently served as consultant to medical investigators.  He developed a very popular General Education course on Statistics and the News.