Efficiency Comparison of Conventional Estimators and Modified Searls Estimators

Abstract

The objective of this research was to compare the mean square errors (MSEs) of the means estimated from simple random sampling by seven estimators: ratio estimator, two linear regression estimators based on the modified Searls estimators as well as conventional estimators proposed by Searls, Sisodia and Dwivedi, Koyuncu and Kadilar and Shabbir and Gupta. The study utilized real-life data on the number of teachers per school of 1,009 schools in Bangkok and its surrounding provinces. It was found
that two linear regression estimators based on the modified Searls estimators outperformed the rest. This observation is in agreement with that of the theoretical study in the past.