EFFECT OF DIFFERENT SAMPLING METHODS ON ESTIMATING THE CONFIDENCE INTERVAL OF POPULATION COEFFICIENT OF VARIATION

This paper focuses on the effect of using simple random sample, ranked set sampling, median ranked set sampling and extreme ranked set sampling on estimating confidence intervals for population coefficient of variation using data generated from normal and lognormal distributions. Confidence intervals are constructed using fiducial inferences. Series of computer simulations are used to evaluate and compare the performance of applying different sampling methods based on empirical coverage probability and mean interval width. Data examples are used to illustrate the performance of these methods.