In manpower planning the data are often incomplete due to left truncation as well as right censoring occurs when a number of people have not yet left when data collection is terminated. Left truncation arises when some people are already in service at the commencement of data collection. Several models have been suggested to describe the internal and external movements of staff in a commercial or industrial organization relating to censored data. In manpower planning, one of the most important variables is completed length of service on leaving a job, since it enables us to predict staff turnover. The most widely used distributions for completed length of service until leaving are the mixed exponential distribution and the lognormal distribution. The Weibull distribution is a particularly important life distribution and a large body of literature on statistical methods has evolved for it. In place of the Weibull distribution, it is often more convenient to work with the equivalent extreme value distribution when the observation are censored. In this paper a parametric and non-parametric estimate of the survivor function which extends Kaplan and Meier’s estimate to include left truncation as well as right censoring has been discussed. A suitable model is also developed, to analyze and predict the pattern of manpower wastage and the basis of all possible individual characteristics responsible for the wastage on the basis of Cox’s approach using Weibull and Extreme value distribution. A real industrial data has been used to validate the above models.
