Model-based estimation of finite population total in stratified sampling

Abstract

The work presented in this paper concerns the estimation of finite population total under model – based framework. Nonparametric regression approach as a method of estimating finite population total is explored. The asymptotic properties of the estimators based on nonparametric regression are also developed under some specified conditions. An empirical study is undertaken to test the practicability of the theories developed using real data from the Kenya Population and Housing Censuses of 1989 and 1999

Keywords: kernel smoothers, nonparametric, separate ratio estimator

> East African Journal of Statistics Vol. 1 (1) 2005: pp. 23-40