Empirical comparison of some regression and regression-type strategies

The main advantages of the regression estimator are that it can be used for both the situations of positively and negatively correlated variables and its precision is usually higher than the simple expansion (direct), ratio and product estimators. Extensive empirical studies on the properties of many types of ratio estimators have been undertaken in several papers, for example by Rao (1969), Rao and Rao (1971), Hutchinson (1971), Royall and Cumberland (1981), Wu and Deng (1983). However not much attention has been given to the use of similar methods for regression-type estimators. In this paper, an attempt has been made to compare the relative performances of some biased and unbiased regression-type strategies with the help of a wide variety of natural populations.