Introduction to design and analysis of experiments

Publication date 26 April 1998

analysis of variance computer package MINITAB exercises factorial crossing fixed and random effects multiple comparisons nested factors randomization real examples

Mathematics Subject Classification ID

Introductory exposition (textbooks, tutorial papers, etc.) pertaining to statistics (62-01) Statistical block designs (62K10) Factorial statistical designs (62K15) Design of statistical experiments (62K99) Design of statistical experiments (62Kxx)

Review text

According to the author, this book is aimed at students and practitioners with no previous background in statistics. The book has fifteen chapters, a brief set of statistical tables, a list of the sources for the data used or referenced in the book, a data index, and a subject index. Chapters 1, 2 and 3 attempt to give an overview of the entire book for those unfamiliar with statistical ideas. Many examples are used to motivate and illustrate fundamental concepts underlying planning an experiment. Ideas of randomization, blocking, and factorial crossing are introduced. An informal discussion of analysis of variance is presented in Chapter 2. Assumptions that justify the standard analyses are presented in rigorous but nonmathematical terms. Chapter 3 provides a more formal introduction to analysis of variance and associated inference procedures. Hypothesis testing and confidence interval procedures are explained in easily understandable terms. Chapter 4 discusses three key decisions the experimenter must make to determine the content of an experiment -- what to measure, what conditions to compare, and what experimental material (units) to use. Many interesting examples are provided to reinforce these basic concepts. Chapter 5 first provides an overview of different basic design strategies available to an investigator and then discusses the completely randomized design in an informal manner. The role of randomization in designed experiments is considered. Again, several examples are provided. Chapter 6 introduces the factorial treatment structure and discusses reasons for factorial crossing. The notion of interaction is explained in words as well as using interaction graphs. The ANOVA decomposition for the two-way design is discussed. Chapter 7 is devoted to the idea of blocking in general. Two particular designs -- the Randomized Complete Block Design and the Latin Square Design -- are explained using several examples. The Split-Plot/Repeated Measures design is also presented in this chapter. In Chapter 8, the author provides guidelines and tips that will help an investigator or data analyst identify the type of design that was used in a study by systematically identifying what the factors are, what the experimental units are, what the response is, whether or not there are blocks, whether or not split plots are involved, etc. Again, all the concepts are well explained with plenty of real examples. Chapter 9 describes more complex designs obtained from the basic designs using the principle of factorial crossing. Expected mean squares in ANOVA tables are introduced and rules are given for choosing appropriate denominators for \(F\)-tests. Chapter 10 explains a general procedure for ANOVA decomposition for balanced designs and illustrates the procedure with many examples. The SWEEP operation is introduced and explained in detail. Chapter 11 considers comparisons among factor level effects, inference for contrasts, various hypothesis tests, and multiple comparisons. Chapter 12 discusses the standard ANOVA assumptions and methods for checking them. Chapter 13 introduces two new ideas -- the idea of nested factors and the concept of fixed and random effects. Complex factor structures obtained by crossing and nesting are discussed. Additional discussion is provided for calculating expected mean squares and for determining appropriate denominator mean squares for ANOVA tests. An illuminating discussion is given which provides an intuitive explanation for why the rules for computing expected mean squares work. Chapter 14 gives a brief but general overview of analyses that are appropriate when some of the factors are ``continuous rather than ``categorical. Simple regression and correlation are explained and the method of least squares is introduced. The chapter concludes with a section on Analysis of Covariance. The final chapter is devoted to a discussion of sampling distributions of test statistics. Four fundamental families of sampling distributions -- the normal, \(\chi^2\), \(t\), and \(F\) -- are introduced and relationships among them are discussed. Chapters 5, 6, and 7 each have a separate section devoted to the use of the computer package MINITAB for obtaining ANOVA calculations and another section devoted to the description of the algebraic formulation of the underlying linear model commonly used in mathematical treatments of experimental design and analysis. There are numerous exercises distributed throughout the book. An abundance of real examples is provided. The book is extremely well written and should be a pleasure to read even for readers with no previous background in statistics. It is highly suited for use as a text book for undergraduate statistics majors as well as undergraduate and graduate students in statistically oriented disciplines. It appears to be particularly well suited for biology and psychology majors. It can also be used in an experimental design course for beginning graduate students in statistics as a supplement to more theoretical text books.

Recommendations

Cited in

(11)

Describes a project that uses

Uses Software

MINITAB

This page was built for publication: Introduction to design and analysis of experiments

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4384479)