Abstract

We, in this paper, extending the fabric of hierarchic estimation for mean introduced by Agrawal and Sthapit (1997) and carried forward successively by Panda and Sahoo (2015) and Panda and Das (2018) to the dimension of variance estimation, propose a new ratio estimator for the estimation of population variance using simple random sampling without replacement (SRSWOR) scheme. The novelty of the proposed estimator of order is that it is predictive in form under the assumption of balanced sampling. Conditions under which the proposed estimator with optimal fares better than the usual variance estimator and the ratio estimator due to Isaki (1983) have been arrived at. In addition to this, the supremacy of the proposed estimator over the one due to Kadilar and Cingi (2006) is established numerically. Theoretical findings are supported by numerical illustration.

Author: K. B. Panda and P. Das

Received on: May, 2020

Accepted on: July, 2021