ABSTRACT
In two-stage cluster sampling, the clusters are commonly referred to as primary sampling units (PSUs) and the units selected in the second stage as the secondary sampling units (SSUs). As with cluster random sampling, two-stage cluster random sampling may lead to a strong spatial clustering of the selected population units in the study area. This may save considerable time for fieldwork, and more population units can be observed for the same budget. However, due to the spatial clustering the estimates will generally be less precise compared to samples of the same size selected by a design that leads to a much better spreading of the sampling units throughout the study area, such as systematic random sampling. Both the PSUs and the SSUs are selected with replacement. The ratio estimator of the population mean and its standard error can be computed with function svymean of package survey.
