449 – Cell Phone-Landline Dual Frame (RDD) Surveys: Weighting Issues
Venn Diagrams, Probability 101, and Sampling Weights Computed for Dual-Frame Telephone RDD Designs
Jonathan Best
Princeton Survey Research International
Trent Buskirk
The Nielsen Company
The continued rise in cell phone penetration creates a real potential for undercoverage bias in many RDD sample surveys. To respond to such threats researchers have begun implementing dual frame RDD sampling strategies. In this paper we present a method for constructing first-stage sampling weights derived under an overlapping, dual frame design (e.g. cell and landline RDD numbers) based on probability 101 fundamentals. Because these two frames potentially overlap at the user level, selection probabilities must be adjusted for multiplicity of selection. Our method resembles weighting strategies consistent with a "single frame" approach and does not require estimation of a compositing factor traditionally used in the "dual frame" approach.
