| Abstract: | Quantifying the precision in parameter estimation from field data on times in service of equipment is an important part of assessing reliability of equipment and, hence, of the design of policies for managing such equipment. This precision is dependent on various factors, and in this paper three such factors—pattern or nature of censoring, extent of information on censored items, and sample size—are examined using the criterion of relative likelihood, which, it is argued, is preferable for small to moderate sample sizes. The relative effect of the three factors is illustrated via examples based on the two parameter Weibull distribution, although the principles of the discussion are valid for a wide variety of lifetime distributions. |
