Unbiased Ratio and Regression Type Estimators

Abstract

Simple random sampling is by far the most commonly used method of sampling in surveys. It is simple, operationally convenient and gives equal chance of selection for all the units in the population. When, however, the units vary considerably in size, as is often the case, simple random sampling does not take into account the possible importance of larger units in the population. Under such circumstances, without foregoing the operational convenience of simple random sampling, it is desirable to use auxiliary information, such as size of unit, at the estimation stage, for obtaining more efficient estimators of the population value in the sense of giving estimators with smaller standard errors. Two examples of such estimation procedures are the 'Ratio and Regression’ methods of estimation.