Graphical methods for data analysis. (English) Zbl 0532.65094

Chambers, John M.; Cleveland, William S.; Kleiner, Beat; Tukey, Paul A.

The Wadsworth Statistics/Probability Series. Belmont, California: Wadsworth International Group; Boston: Duxbury Press. XIV, 395 p. $ 37.75 (1983).

New and old graphical methods for analyzing data are presented for users of all sciences. Chapters 2 to 5 show exploratory methods in which the raw data themseles are displayed, as methods for portraying the distribution of a single set, for comparing several data sets. In chapter 4 paired measurements are treated, and in chapter 5 more than two variables. In chapter 6 and 7 original data are studied with statistical models as soon as derived statistics in order to find appropriate statistical analyses. The plots allow to probe the results of analyses and to judge whether the data support the assumptions (e.g. of the regression model).
The general discussion in chapter 8 presents principles that help to judge the strength and weaknesses of graphical displays and help for designing new ones. The appendix presents 35 data sets which are used as examples in the text and in the exercises.

Reviewer: H.Grimm

MSC:

65S05	Graphical methods in numerical analysis
65C99	Probabilistic methods, stochastic differential equations
65-01	Introductory exposition (textbooks, tutorial papers, etc.) pertaining to numerical analysis
68U99	Computing methodologies and applications
62J99	Linear inference, regression

Keywords:

exploratory data analysis; textbook; statistical models; regression model; graphical displays; examples