Approximately Optimum Stratification on the Auxiliary Variable
Ravindra Singh · Journal of the American Statistical Association · 1971
Serfing [8] has obtained optimum stratification on the auxiliary variable x making use of the cum. method when the regression of the estimation variable y on x is linear and the correlation is nearly perfect. In the present article another procedure, applicable in more general situations, is suggested to obtain optimum stratification on x. The method makes use of a new cum. rule for obtaining approximately optimum strata boundaries. A priori knowledge of the form of regression of y on x and the form of the conditional variance function V(x|y) is assumed. The article concludes with a numerical example.