Logarithmic-Type Estimators of Population Coefficient of Variation Under Successive Sampling in the Presence of Nonresponse and Measurement Errors

This paper considered the effects of random non-response and measurement error on the estimation of the population coefficient of variation of the characteristic studied character in two-occasion successive sampling. To lessen the effects of non-response and measurement errors, logarithmic-type estimators for estimating the population coefficient of variation have been proposed. The theoretical properties of the estimators (biases and MSEs) were obtained using Taylor’s series approximation method. Empirical studies have been conducted with simulated data and the results revealed that the proposed estimators demonstrated high level of efficiency, and appropriate recommendations have been made to survey statisticians for use in real-world scenarios.