Abstract:
|
Surveys concerned with substance use or the etiology of medical conditions often want to have larger samples of persons who are in specific subpopulations, while maintaining the ability to monitor transitions from one subpopulation to another. We consider optimal survey design using a two-phase sample, in which persons are selected for the first phase using a screening instrument that misclassifies some individuals, and the second-phase sample is selected using a more accurate classification instrument. We discuss software that can be used to solve the optimization problem with constraints, and illustrate with an example.
|