Online Program

Can Information from Market Research Companies be Used to Develop an Efficient Sampling Strategy for a Rare Population?

*Kirk M. Wolter, NORC and University of Chicago 

Keywords: National Immunization Survey, two-phase sampling, stratification, disproportionate sampling

We discuss a new sampling strategy for the National Immunization Survey (NIS) conducted by the Centers for Disease Control and Prevention. The NIS targets the relatively rare population of children 19-35 months and adolescents 13-17 years. The survey is conducted in two phases: 1) a survey of over 9,000,000 telephone numbers per year that identifies households with age-eligible children and collects socio-demographic characteristics, childhood- and adolescent-administered vaccination histories, and contact information for the child’s vaccination providers; and 2) a mail survey that collects vaccination histories from the identified providers. Market research companies have developed large databases containing demographic and economic information on American households. Using exact or modeled age data, such companies are able to develop lists of households that are presumed to have at least one resident in a user-specified age range. The age-targeted lists are imperfect, with both erroneous inclusions and exclusions. We consider the sampling strategy of 1) stratifying the NIS sampling frame by presence of the telephone number on the age-targeted list, and 2) oversampling the age-targeted stratum and under-sampling the remainder stratum. We examine whether the methods yield worthwhile cost savings while preserving the precision of survey estimators.