A New Class of L-Moments Based Calibration Variance Estimators

Variance is one of the most important measures descriptive statistics and commonly used for statistical analysis. The traditional second-order central moment based variance estimation a widely utilized methodology. However, estimator highly affected in presence extreme values. So this paper initially, proposes two classes calibration estimators on an adaptation recently proposed by Koyuncu then presents new class L-Moments utilizing characteristics (L-location, L-scale, L-CV) auxiliary information. It demonstrated that are more efficient than adapted ones. Artificial data considered assessing performance estimators. We also application related to apple fruit purposes article. Using artificial real sets, percentage relative efficiency (PRE) with respect ones calculated. PRE results indicate superiority over In manner, could be applied expansive range survey sampling whenever information available