A great many studies in statistics deal with deaths or with failures of components: they involve the numbers of deaths, the timing of death, or the risks of death to which different classes of individuals are exposed. The analysis of survival data is a major focus of the statistics business (see Kalbfleisch and Prentice, 1980; Miller, 1981; Fleming and Harrington 1991), for which R supports a wide range of tools. The main theme of this chapter is the analysis of data that take the form of measurements of the time to death, or the time to failure of a component. Up to now, we have dealt with mortality data by considering the proportion of individuals that were dead at a given time. In this chapter each individual is followed until it dies, then the time of death is recorded (this will be the response variable). Individuals that survive to the end of the experiment will die at an unknown time in the future; they are said to be censored (as explained below).
27.1 A Monte Carlo experiment
With data on time to death, the most important decision to be made concerns the error distribution. The key point to understand is that the variance in age at death is almost certain to increase with the mean, and hence standard models (assuming constant variance and normal errors) will be inappropriate. You can see this at once with a simple Monte Carlo experiment. Suppose that the per-week probability of failure of a component is 0.1 from one factory but 0.2 from another. ...