Response Models in RDD Surveys

• Activity Number: 79
• Type: Contributed
• Date/Time: Monday, August 9, 2004 : 8:30 AM to 10:20 AM
• Sponsor: General Methodology
• Author(s): Stephen Williams*+ and John W. Hall and Ronghua Lu
• Companies: Mathematica Policy Research, Inc. and Mathematica Policy Research, Inc. and Mathematica Policy Research, Inc.
• Keywords: nonresponse ; propensity modeling ; RDD telephone surveys ; Community Tracking Study

Abstract

For random digit dialing (RDD) household surveys, for example, the resolution of whether or not a number identifies a household is becoming more challenging. Likewise, obtaining household participation in a survey is becoming more difficult. This paper investigates the use of logistic regression models to predict for each sampled RDD number in a recently completed household survey, the likelihood of resolving the number and the likelihood, if the number is a household telephone number, that the household will complete the screening questions, and finally that the household will complete the questionnaire. It builds on earlier analyses by the same authors. We apply these methods to compute weights for round three of the Community Tracking Study (CTS) Household Survey. The RDD numbers in this example are generated through the GENESYS system, which provides a substantial amount of auxiliary demographic information at the telephone-exchange level. Other information includes prior responses from the longitudinal panel for some numbers. We develop these models for a single stage national survey sample and also for samples in a number of the communities studied in the CTS.