PARAMETER ESTIMATION IN MULTIPLE LINEAR REGRESSION MODELS USING RANKED SET SAMPLING

Authors : Yaprak Arzu ÖZDEMİR, Alptekin ESİN A

Pages : 7-20

Doi:10.1501/Commua1_0000000194

View : 11 | Download : 7

Publication Date : 2007-02-01

Article Type : Research Paper

Abstract :In statistical surveys, if the measurements of sampling units according to the variable under consideration is expensive in all sense, and if itis possible to rank sampling units according to the same variable by means ofa method which is not expensive at all, in those cases, Ranked Set Samplinginsert ignore into journalissuearticles values(RSS); is a more e¢ cient sampling method than the Simple Random Samplinginsert ignore into journalissuearticles values(SRS); to estimate the population mean. In this study, the eğects of using RSSin multiple linear regression analysis are considered in terms of estimation ofmodel parameters. Firstly, according to RSS and SRS the estimates of multipleregression model parameters are obtained and then the eğects concerning thevariances of the estimators are investigated by Monte Carlo simulation studybased on Relative E¢ ciency insert ignore into journalissuearticles values(RE); measure. It is shown that the estimatorsobtained based on RSS are more e¢ cient than the estimators based on SRSwhen the sample size is small
Keywords : Simple Random Sampling, Ranked Set Sampling, Order Statistics, Relative E¢ ciency, Regression Analysis