COMPARISON OF SOME RANKED SET SAMPLING ESTIMATORS

Abstract

The method of Ranked Set Sampling has been proved to be one of the efficient methods of sampling technique. This was suggested as an efficient method compared to the well-known Simple Random Sampling (SRS) method. However, a lot of these methods had been developed over the year. This research work focused on comparison of some classes of Ranked set Sampling Estimators. Four classes of RSS estimators were considered and these were grouped into two, based on their mode of selection. The selection’s modes considered in this research work were the odd and even respectively. The variance and the Mean Square Error (MSE) under both even and odd cases of these classes of RSS estimators were obtained and compared under the different distributions through simulation study in order to know their performances. It was discovered that, at the small sample size of 3 and 4, the Extreme Ranked Set sampling (ERSS) estimator performed better than the Median Ranked set sampling (MRSS) estimator. Also, when an increased sample size of 6 and 9 are considered, it was observed that the Balanced Group Ranked set sampling (BGRSS) estimator performed better than the Two-Stage Ranked Set Sampling (TSRSS) estimator.