Observational Before-After Studies in Road Safety

Rating:
Summary: Must-reading for those conducting safety studies
Review: This is one of the best technical books I've ever read on any subject--and I've read many of them! It should be must reading not only for those who conduct transportation safety studies--the subject of the book-- but also for those conducing any observational study ("before/after study"), whether safety-related or not.

This book is a valuable and much-needed contribution to the technical literature. The book begins by showing how classical statistical techniques are often misused in observational studies. Classical techniques assume a rigorous experimental design, with random assignment to treatment and non-treatement groups. These conditions do not apply to observational studies, such as those determining the influence of smoking on health. Errors such as assuming safety in the immediate "before" period is a good predictor of safety in the "after" period are discussed. The book then goes on to provide valid techniques that work. It shows how to compare the what would have been the safety in the after period without treatment, to the safety that was achieved. Complications of assessing the safety in the before and after periods are addressed.

The book is not only an important technical contribution, it is also extremely well-written. The author's writing style is extremely lucid and should serve as an example to others. The first section of the book provides a qualitative description of the material. Subsequent sections then repeat the material, but with mathemtatical rigor. Numerous numerical examples (some with real-world data; others with hypothetcial data for simplicity sake) make the material clear. Mathematical dervivations are provided for rigor, but isolated from the rest of the text to keep from overwhelming the reader.

The author writes with a conversational style, frequently re-emphasizing important points. He takes care to show at all points where he is in the logical discussion and how the current topic fits in to the overall picture. Some of the words are written tongue in cheek. My favorite is his description that
"they do not have a problem of congnition, but only of brittle bones".

In short, the book is outstanding for both its technical content and for the way the material is presented.