Trends in sampling designs

Abstract

Sampling designs should be theoretically based, once the investigators have decided on the purpose of their study and the comparisons needed. Designs have included traditional ‘descending’ designs such as probability sampling, as well as ‘ascending’ designs, such as snowball sampling. No longer are treatment-based samples the norm in psychiatric epidemiologic studies. Sampling designs have a profound effect on the rates of disorder as well as risk factors analyses; thus, these designs need careful consideration. This chapter introduces diverse options which have been used in epidemiologic studies and discusses the advantages and disadvantages of each of these designs.