A sequential adaptive sampling scheme for rare populations with a network structure

The design of an efficient sample for surveying rare-trait populations still continues to be a challenging task. This contribution discusses a sampling strategy, which, integrating an adaptive component into a sequential selection, aims to intensify the detection of positive cases both by exploiting the links among population units and by spreading the sample over the whole population. Obviously, if a separate frame for a rare population exists the sample may be selected using standard methods as well as if adequate auxiliary information are available. The suggested approach refers to situations in which neither of these two conditions happen. Many factors related to the structure more or less clustered of the population and to the characteristics of the sampling design may be decisive in such a situation.