Estimation of the Finite Population Mean by Judgment Post-stratification

• Activity Number: 180
• Type: Topic Contributed
• Date/Time: Tuesday, August 10, 2004 : 8:30 AM to 10:20 AM
• Sponsor: General Methodology
• Author(s): Xiaobai Li*+
• Companies: Ohio State University
• Keywords: ranked set sampling ; stratification ; post-stratification ; judgement ; finite population ; variance reduction

Abstract

It is well known that stratified sampling plan and ranked set sampling plan both give better estimator of population mean than simple random sample. We need to know the fraction of the population that falls into each stratum and specify the rules for stratification before the data is analyzed. And we need to make the judgement before the sample is takenor before quantification. Judgement post-stratification is a new method that keeps the benefits of post-stratification and ranked set sampling while overcomes their shortcomings. It is based on the collection of a simple random sample of units to be measured, while creating the strata by post-stratification and comparing the units in the finite population. The comparison can be based on subjective judgment or objective concomitant variables. This method is most efficient when te ranking of the units in the population can be carried out relative easily and cheaply compared to the effort and expensie required for measuiring the units. Simulation results and real dataset analysis demonstrate that judgment post-stratification yield unbiased estimator of the population mean with smaller mean squared error.