 |
 |
Dear Allstaters Looking for some clarification on the definition of cluster / multi-stage sampling. Discrepancies in texts. Cluster sampling - random sample of clusters and then everybody in cluster is sampled. (happy with that unless otherwise informed) Multi-stage : Consider following example. A random sample of zones within a city is obtained, followed by a random sample of households within each zone (two-stage). Assume no further naturally occurring clusters in the house. In multi-stage sampling - would I survey all individuals in each household or take a sample of them? If the latter - is sample random or not? If random sample - would it be three-stage sampling? I always assumed the number of stages = number of natural clusters sampled embedded within other. Thanks Philip Sedgwick St. George', University of London