Sensitivity of Confidence Interval to Changes in Sample Variance Parameters within Transit Time Assessment of Postal Transportation Flows

Abstract

The paper is focused on confidence interval sensitivity to changes of dispersion parameters within measurement of the transit time of postal transportation services, namely end-to-end services for single piece priority mail. The role of confidence interval is the expression of estimate accuracy for on-time probability. The accuracy is expressed as width of the confidence interval of on-time probability estimate. On-time probability is interpreted by variable called weighted estimate of on-time probability, which also includes distribution of mail flows within geographical stratification according to disjunctive set of defined postal areas. The rate of variability caused by used sample design is expressed by dispersion parameters of sample variance in various forms and by design factors regarding particular samples. The main objective of this paper is to research sensitivity of confidence interval to changes in sample variance parameters considering limit value of interval width related to acceptable accuracy level. Measurement system created by sample of test postal items flows is based on real traffic mail flows on fixed postal transportation network. The results come out from modelling of this measurement system and evaluation of postal transport network efficiency by such system is an important managerial tool for quality of service. Analysis of mentioned sensitivity can help to find the most appropriate sample design.