variance estimation

The presence of extreme events gives rise to outrageous results regarding population parametersand their estimates using traditional moments. Traditional moments are usually influenced by extremeobservations. In this paper, we propose some new calibration estimators under L-Moments scheme for variance which is one of the most important population parameters. Some suitable calibration constraints under double stratified random sampling are also defined for these estimators. Our proposed estimators based on L-Moments are relatively more robust in presence of extreme values. The empirical efficiency of proposed estimators is also calculated through simulation. Covid-19 pandemic data from January 22, 2020, to August 23, 2020, is considered for simulation study.

In this study, we have proposed a new generalized estimator using auxiliary information for the estimation of population variance in presence and absence of measurement error. The expressions approximate bias and mean square error of the proposed generalized estimator are derive up to the first-order. Several new and existing estimators are found as the special cases of proposed estimator and expressed on various values of optimizing and generalized constants. The proposed estimator is compared mathematically with some existing estimators under certain conditions. The performance of proposed estimator is observed by simulation and real data application under different sample sizes. It is observed that the proposed estimator performed well than other competing estimators in the presence and absence of measurement error.