ABSTRACT
This chapter continues exploring topics in survival analysis when, in addition to data modification caused by censoring and truncation, some observations may be missed. Censoring, truncation and missing may imply biased data, and hence it will be of interest to understand how these modifications act together and how a consistent estimator should take into account the effect of these modifications. Some modifications may be destructive. Further, missing always decreases the size of available observations. All these and other related issues are discussed in this chapter. The chapter is devoted to estimation of distributions, while remaining sections explore regression problems. It explores estimation of the distribution of a right censored (RC) lifetime when the indicator of censoring may be missed according to a MAR (missing at random) mechanism. In particular, the chapter presents the estimation of the cumulative hazard, hazard rate, survival function and probability density.
