Weighting Approach for a Statewide Dual-Frame RDD Survey

• Activity Number: 445
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Survey Research Methods Section
• Author(s): Liza Nirelli McNamee* and Monica Reyna and Snehalatha Gantla and Priscilla Salant and J.D. Wulfhorst
• Companies: University of Idaho and University of Idaho and University of Idaho, Social Science Research Unit and University of Idaho and University of Idaho
• Keywords: design weights ; raking ; trimming ; Raking SAS Macro with weight trimming ; sample balancing ; iterative proportional fitting

Abstract

This paper describes a statewide dual-frame RDD survey weighting procedure developed in SAS using the Raking SAS Macro with weight trimming (Izrael, 2009). The AAPOR Cell Phone Survey Task Force Report states, "There is no consensus regarding how RDD cell phone samples should be weighted, especially when combining them with RDD landline samples." During the winter of 2014, a dual-frame, stratified random sample, including cell and landline telephone users, was utilized to ensure coverage of a statewide population. A total of 12 sub-sampling frames were created by using six non-overlapping strata representing six geographic regions and two frames, cell and landline, for each region. The weighting process consisted of developing design weights for selection probability, eligibility, non-response, multiplicity due to multiple phones, respondent selection procedure, and frame overlap. Raking and trimming techniques were used to adjust the sample to reflect the statewide telephone service benchmarks and demographics of interest. This paper presents the weighting methodology along with estimates at each stage, the limitations of the procedure, and the methodological lessons learned.