Statistical Distributions in Reliability

This paper gives an introduction to the important probability models and statistical techniques that can be used in reliability studies. In Section I several important ideas are stressed. They are 1) That it is generally impossible to fit a statistical distribution from data alone, but that one must usually select first a family of distributions based on some theoretical and/or physical considerations and then estimate the parameters in this family of distributions. 2) That it is important to know something of the properties of the estimators used since not all estimators are necessarily good and could possibly be worse than just picking a value for the estimate at random. In addition to the above points important statistical quanities are defined and discussed. Also, in Section I the ideas of testing hypothesis are discussed and illustrated. In Section II a very brief discussion of some of the important discrete distributions is made. In Section III the probability distributions under discussion are briefly introduced. In Section IV some guides and examples are presented on how the different family of distributions may arise.