Presents theory and methods for recognizing and handling the more complicated cases often encountered in practice. This work integrates a physical understanding of underlying phenomena and the statistical modeling of observation 'noise' to provide a single theoretical framework for accelerated testing.