Dual-Frame Sample Weighting for a Telephone Survey of Children's Vaccination Status

• Activity Number: 633
• Type: Invited
• Date/Time: Thursday, August 2, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Kennon R. Copeland*+ and Kirk M. Wolter and N. Ganesh and Meena Khare and Stacie M. Greby
• Companies: NORC and NORC and NORC and National Center for Health Statistics and National Center for Immunization and Respiratory Diseases
• Keywords: Dual-frame weighting ; Bias ; Mean-squared error

Abstract

Rapid increases in the US cell-phone-only population have motivated use of dual frame, landline and cell phone random-digit-dial, sample designs to minimize potential coverage error. Given cost considerations, the size of the cell sample is not typically proportional to the size of the cell phone population, yielding variability in survey weights that increases standard errors for estimates. As a result, cell sample sizes may be small for states and local areas, which can also serve to increase the size of standard errors. We provide a framework for weighting a dual frame sample design from the National Immunization Survey, which monitors the vaccination rates of children 19-35 months and adolescents 13-17 years for the nation, states, and selected local areas. This framework addresses overlap between the landline and cell phone samples through a composite estimation approach, and the variability in survey weights through weight attenuation for the cell phone only domain, while controlling impact on bias and mean-squared error. Bias, variance, and mean-squared error of the weighting approach are examined and contrasted with those of other common dual frame weighting approaches.