Single Frame versus Multiple Frame Designs for Telephone Samples

• Activity Number: 499
• Type: Invited
• Date/Time: Thursday, August 11, 2005 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Charlotte Steeh*+
• Companies: Georgia State University
• Keywords: Mobile Phone Sampling ; Multiple Frame Designs ; Wireless Phone Surveys

Abstract

As new telephone technologies are developed, RDD sampling methods must be modified if telephone surveys are to remain representative of the general population. The introduction of Voice over Internet Protocol (VoIP) adds another option to a landscape that already includes mobile and landline. Obviously, a multiple frame design could handle this diversity, but as the number of modes rises beyond two, this design becomes complicated to implement. In this paper, we propose a design that will cover all telephone modes, thus giving maximum coverage to the general population. The design treats all telephone numbers (fixed line, cellular, and VoIP) as one frame. The sample unit changes from the household to the individual. The problems discussed involve calculating weights that adjust for multiple telephone access, defining stratification criteria prior to selecting sample units, ensuring geographic representation when area codes are no longer tied to specific localities, taking into account the probability of choosing two or more individuals from the same household, and deciding how to choose a respondent when the number reaches a fixed line, household telephone.