Optimum Stratified Sampling Using Prior Information

On Sampling of Very Rare Human Populations

“…Ericson [1] has given the solution to the problem of finding an optimum stratified sample when costs, variances, and prior information differ by strata. Suppose that 0"7 is the known within stratum variance in the ith stratum, 11"; is the known proportion of the population in the ith stratum, vI is the Bayesian prior variance in the ith stratum, (l/vI is a measure of prior information) and c; is the variable per interview cost of screening and interviewing in the ith stratum.…”

“…This can be justified even by non-Bayesian samplers by estimating the maximum bias resulting from such a procedure which would usually be small. The Bayesian approach [1] which is followed here indicates that under the weakest of priors, an optimum sampling design omits clusters where P<Pc because of cost reasons.…”

Optimum allocation in multivariate stratified random sampling with overhead cost

“…Ericson [1965] has stated the problem as to "minimize the posterior variance of the over all population mean subject to a total budgetary constraint". He also discussed the case when more than one population characteristics are to be estimated, under the assumption that the strata are sufficiently similar with respect to the various characteristics.…”

Compromise Allocation for Combined Ratio Estimates of Population Means of a Multivariate Stratified Population Using Double Sampling in Presence of Non-Response

“…In Phase I, we obtain the sample size h n in each stratum by minimizing variance given in (5) for fixed cost given in (7). At Phase II subsample size from non-respondents group has been obtained by minimizing the sampling variance in (5) for given cost in (7).…”