ABSTRACT
This chapter argues that QQ-plots and hazard plots can be of assistance. A probability plot provides evidence as to whether data have come from a population having a stated probability model S. Since survival times are often transformed by logarithms to create a natural measurement scale against which the data may appear symmetrically distributed, normal probability plots will be relevant for the transformed data; this would mean that the original data follows a lognormal distribution. To to assess whether the inter-arrival times of cars at a service station follow an exponential distribution, the analysts first obtain a data set by direct observation of the service station and then decide through a stem-and-leaf plot whether an exponential distribution is a possible model. In engineering applications, the lifetime of electronic components is of interest; in medical applications, the lifetime of cancer patients is of interest.
