New estimators of population variance based on logarithmic transformation in the presence of random non-response and measurement errors under successive sampling

This paper considered the problem of estimation of population variance of the study character in two-occasion (successive) sampling in absence and presence of non-response and measurement error. Logarithmic type estimators have been developed to reduce the nuisance effect of non-response in sample surveys. The proposed logarithmic based estimators aim to enhance the precision and validity of inferences drawn from successive sampling surveys, where non-sampling errors can significantly affect the quality of the data and the resulting population parameter estimates. The expressions for the biases and mean squared errors of the proposed estimators were derived to quantify their statistical properties and performance. The empirical results through simulation studies demonstrated that the proposed logarithmic-type estimators outperform the traditional variance estimator obtained through linear combination of sample variance and conventional ratio estimator in terms of relative absolute bias and mean squared error, especially when non-response and measurement errors are substantial.