In view of the systematic underestimation of aggregates which are associated with the richer sections of the population in the National Sample Survey and other large scale sample surveys, this paper makes a case for over- sampling the rich in such surveys.
