Design of experiments can make or break a good statistical analysis. The subject itself is quite rigorous and formal, with algebraic concepts that lie underneath. This book, however, is aimed at the practitioner. I used this both during my MS studies and used a few chapters when I taught statistical inference at the University of Texas at Arlington. The book is motivated by “real” examples; that is, the data used both in exercises and throughout the text is real data from industry. Most of the experiments have a manufacturing flavor to them, but are presented in a way that will be easily accessible by almost anyone. The book avoids proofs while still maintaining a rigorous treatment of the subject. It begins very simply, not presupposing a large familiarity with statistics in general. Another nice feature of the book is notations by the authors on chapters and sections that are not totally necessary for a casual or first time reader. This allows the text to be useful for beginners and professionals alike. The exercises are well designed, informative, and illuminating. The language is quite readable, and I would recommend this book for teaching upper level undergraduates, masters level work, or self-study by a data scientist or engineer.

— Rachel Traylor, Ph.D.

### Review

## Prerequisites

### high school algebra, some calculus, basic statistics (including linear regression), understanding of basic matrix algebra

## Topics Covered

- basic regression analysis
- single factor experiments
- one-way layouts
- multiple comparisons
- one-way random effects

- experiments with more than one factor
- paired comparison designs
- randomized block designs
- two-way layout
- multi-way layout
- Latin square designs
- Greco-Latin square designs
- Balanced incomplete block designs
- split plot designs
- ANCOVA

- full factorial experiments at two and three levels
- fractional factorial experiments at two and three levels
- non regular designs
- Placket-Burmann Designs
- Hall’s Designs
- Mixed-level orthogonal arrays

- Complex aliasing
- response surface methodology
- robust parameter design
- analysis of experiments with non normal data

### Attributes

Difficulty | 3 |

Good for teaching? | 5 |

Proof Quality | NA |

Quality of Exercises | 5 |

Readability | 4 |

Self-Study | 4 |