Ratio estimation in poststratified sampling over two occasions

Abstract

Two estimators are proposed for the estimation of the second occasion population ratio, R2, of two characters of study in poststratified sampling over two occasions. The estimators are proposed along the line of Tripathi and Sinha (1976) and Okafor (1985). One of the estimators, d1, is a ratio-cum-product type estimator, while the other, d2, is a product-cum-ratio type estimator. Both estimators do not assume knowledge of the first occasion population ratio, R1. Expressions for the optimum matching or replacement fractions of both estimators are obtained since the estimators are based on a partial replacement of sample units on the second occasion. Conditions under-which one estimator is to be preferred to the other estimator are obtained for repeated samples of fixed sizes.


Keywords: Successive sampling, poststratified sampling, ratio estimation, matching or replacement fraction, repeated samples.


[Global Jnl Pure & Appl. Sci. Vol.9(3) 2003: 407-414]