Balanced Sampling for Multi-Way Stratification (Method)


Balanced sampling is a class of techniques using auxiliary information at the sampling design stage. Many types of sampling designs can be interpreted as balanced sampling, such as simple random sampling with fixed size, stratified simple random sampling and unequal probability sampling.

Furthermore, the balanced sampling can be applied to define a multi-way stratification design also known as incomplete stratification or marginal stratification. Multi-way stratification allows to plan the sample sizes of the domains of interest belonging to two or more non-nested partitions of the population in question without using the standard solution based on a stratified sample in which strata are identified by cross-classifying the variables defining the different partitions (one-way stratified design). The standard solution in many Structural Business Surveys (SBSs) may have drawbacks from the view-point of cost-effectiveness. In fact, SBSs produce typically estimates for a great number of very detailed domains forming several non-nested partitions of the population and creating really small cross-classified strata.


To read the entire document, please access the pdf file (link under "Related Documents" on the right-hand-side of this page).


Your feedback is appreciated. Please send your remarks, suggestions for improvement, etc. to memobust@cbs.nl.