ABSTRACT
To obtain estimates of totals, the sample observations have to be multiplied by factors, variously called weights, multipliers, or weighting, raising or inflation factors, which depend on the particular sample design and the method of estimation adopted. A sample design becomes self-weighting with respect to the particular linear estimator of a total (or an average) when the multipliers of the sample units are all equal: with only one multiplier, the tabulation becomes simpler, speedier and more economical, as the unbiased estimator of the universe total is obtained on multiplying the sample total by the constant, overall multiplier, and the estimator of the ratio of the universe totals of two variables is obtained on taking the ratio of the corresponding sample totals; the estimation of the variances, covariances, etc., is also simplified.
