Upon surveying the distributions examined in this book, the reader is apt to seek

answers to the question, Given a data set, what distribution model should one use

to fit the data? To assist the analyst in this area, formal procedures for testing the

goodness-of-fit of a particular distribution are introduced. However, the efficiency

of such procedures is limited. As such, reliability professionals usually settle for

the use of one or two distribution models that make sense for the kind of data they

see in their industry. For example, reliability engineers who work with mechanical

systems often make use of a Weibull distribution for modeling wearout phenom-

ena. This is not exclusively true, as others, depending on their specific industry or

company, make use of the lognormal distribution. On the other hand, reliability

engineers who work in the electronics industry often rely on the exponential

distribution for modeling time-to-failure phenomena. This has begun to change in

certain segments of the electronics industry that have started to use distributions

such as the Weibull for modeling time-dependent hazard rates.