ACSPRI Conferences, RC33 Eighth International Conference on Social Science Methodology

Font Size:  Small  Medium  Large

Propensity Model for Weighting Dual Frame Telephone Samples

Robert Benford, Trevor Tompson, Julian Baim, John Lien, Lancey Heyman

Building: Law Building
Room: Breakout 7 - Law Building, Room 028
Date: 2012-07-10 11:00 AM – 12:30 PM
Last modified: 2012-06-19


Propensity models are a potential solution to compensate for non-coverage and non-response, problems that threaten telephone surveys even when frames do cover the entire population of inference and where response tendencies by frame may lead to differential response or potential bias by frame. Dual-frame RDD (landline and cell) surveys improve coverage of the entire adult population and possibly improve non-response when dual phone service households are included regardless of sample frame. In this paper we model cell only versus dual service persons as a function of demographic and attitudinal variables found in Mediamark Research & Intelligence (GfK-MRI). This reference source provides a robust set of potential covariates for modeling purposes. In the paper we develop the model, compare the application of this model to our current first stage pre-weight practice over two AP-GfK polls, and apply the model to the landline sample only. Further, the paper addresses the bias and variance components that result from different approaches.