Early approaches to accelerated testing were based on the assumption that there was a simple acceleration factor dependent on stress that would correspond to a linear scaling of time from the operating stress to the accelerating stress. This corresponds to the simplest physical model of the kinetics governing the underlying degradation. This book provides theory and methods for recognizing and handling the more complicated cases that quite often occur in practice. It provides an integration of chemical kinetics and statistics aimed at designing accelerated tests and modeling the results. The book includes general approaches that can be used in various computational tools, an explicit computing environment written in Splus, practical applications, exercises for teachers and students, and intriguing open questions for researchers.