SOME NEW DEVELOPMENTS IN CHAIN TYPE ESTIMATORS FOR ESTIMATING POPULATION MEAN IN CLUSTER SAMPLING.

In this paper we suggest a ratio type exponential estimator, a chain ratio-type estimator, a chain ratio-type exponential estimator and a chain ratio-ratio-type exponential estimator in cluster sampling. Further we consider a generalized class of chain ratio-ratio-type exponential estimator. The properties of the suggested class of estimators are studied under large sample approximation. Conditions are obtained under which these estimators are more efficient than the usual unbiased estimators. [ABSTRACT FROM AUTHOR]