Using extreme values and fractional raw moments for mean estimation in stratified random sampling

Authors : Shoaib ALİ, Manzoor KHAN, Javid SHABBİR

Pages : 383-402

View : 13 | Download : 5

Publication Date : 2018-04-01

Article Type : Research Paper

Abstract :Unusual observations can occur in sample survey data. Mean estimator is sensitive to very large and/or small values, if included in sample. It can provide biased results and ultimately, tempted to delete from the sample data. Extreme values, if known, can be retained in data and used as the auxiliary information to increase the precision of estimate. Similarly, a known auxiliary variable is always source of improvement in precision of estimates. A transformation can be used for the auxiliary variable to get even more precised estimates. In this article, we have suggested modified estimators for finite population mean when a sample is drawn under stratified random sampling design. We used extreme values and fractional raw moments of the auxiliary variable and suggested improved ratio, product and regression type estimators. By theoretical comparison, efficiency of proposed estimators is established and numerical and simulation studies are conducted to support the theoretical results.
Keywords : Use of Extreme Values, Fraction Raw Moments, Estimation of Mean, Stratified Random Sampling