The statistics of aging models and practical reality

The consequences of the inherent high variability of life data are described, and some objective methods to aid in the design and analysis of aging experiments are reviewed. It is argued that before accelerated aging tests and proposed aging models can acquire credibility with high-voltage (HV) equipment designers, testing of many more specimens and improved analysis methods are needed. The recent availability of sophisticated computer programs will aid in the analysis of aging data.<
>