University Publications

This study is an attempt to know to what extent the efficiency of the estimates of the upper and lower quartiles of the life time functions, median life time and cumulative intensity rate of failure using life table method, are affected by factors such as; method of grouping data, censoring degree in the data, sample size and population distribution. The study is conducted and based on three generated exponentially distributed population with different parameters levels and along with different variances levels as well. The true values of interested parameters are computed first. Then estimates are computed from three sample sizes taken from each population. For each sample size three grouping levels and five censoring degrees are adopted. Mean Square Error (MSE) is used to compare the above mentioned property, and the inter quartile range is used as another measure of the efficiency. The results lead to the fact that the efficiency of the estimates increases with the increase in sample size for a given grouping and a given censoring degree, as well as the case of increasing with the increase in grouping levels for a given censoring degree and a given sample size. And the efficiency of the estimates decreases with the increase in censoring levels for a given sample size and a given grouping. For each sample size from each population, the study revealed that they have a positive skewedness. And for some cases of population three, precisely for small sample with grouping 8, and for big sample with grouping 12, they clearly tend to have a symmetric distribution