A Dual Frame Sampling Design for an RDD Survey that Screens for a Rare Population

• Activity Number: 183
• Type: Topic Contributed
• Date/Time: Tuesday, August 10, 2004 : 8:30 AM to 10:20 AM
• Sponsor: Section on Survey Research Methods
• Author(s): Kadaba Srinath*+ and Michael P. Battaglia and Meena Khare
• Companies: Abt Associates Inc. and Abt Associates Inc. and National Center for Health Statistics
• Keywords: screening sample size ; list frame ; optimum allocation

Abstract

In a random-digit-dialing survey, the sample from a rare population is selected through screening a sample from the general population. Generally, a large screening sample size is required to identify an adequate sample from the rare population. The effort and cost of screening could be high if the required sample size from the rare population is large. Alternatively, if a partial list of members of the rare population is available, then a method of minimizing the screening cost is to use a dual frame sampling design with samples from both the RDD frame and the list frame. We investigate the use of a dual frame sampling design for the National Immunization Survey (NIS), in which the population of interest is children aged 19 to 35 months. We consider a Hartely-type ratio estimator for population proportions based on samples from both the frames. We determine the optimum allocation of the screening sample between the RDD frame and the list frame which minimizes the screening cost for a given sample size from the rare population. We compare the efficiency of the estimates from the RDD frame design and the dual frame desgin in the NIS.