The problem of optimum allocation in multivariate stratified sampling

Abstract

Generally, sample surveys are multivariate in nature where multiple response are obtained on every unit selected in a sample, that is, more than one characteristics are defined on each and every unit of the population. While dealing with a multivariate stratified population, to workout an allocation that is optimum for all characteristics is almost impossible unless the characteristics are highly correlated. Some compromise must be allowed to obtain an allocation that is optimum, in some sense, for all the characteristics. Since such allocations are based on some compromise criteria they are known as compromise allocations. This paper deals with the problem of obtaining an optimum allocation in multivariate stratified sampling design.