Conference Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 489 - Quality of Survey Responses
Type: Contributed
Date/Time: Wednesday, August 10, 2022 : 2:00 PM to 3:50 PM
Sponsor: Survey Research Methods Section
Abstract #323556
Title: Propensity Models for Nonresponse Adjustments in Telephone Surveys
Author(s): Yangyang Deng* and Ronaldo Iachan and Randy ZuWallack and Adam Lee and Thomas Brassell
Companies: ICF Macro, Inc. and ICF Macro, Inc and ICF Macro, Inc. and ICF Macro, Inc. and ICF Macro, Inc.
Keywords: Telephone Surveys; Non-response analysis; Multilevel Modeling; BRFSS; Weighting
Abstract:

Telephone surveys experience large and growing levels of nonresponse, which bring the risk of potential biases. Biases may be large to the extent that nonrespondents may differ from respondents in the key survey outcome variables. Survey nonresponse adjustments typically capitalize on external data that may be available for nonrespondents, data that are usually lacking for telephone surveys. This paper investigates the development of propensity models using both multilevel logistic regression and machine learning methods. The methods take advantage of auxiliary data which are merged to sample files for the Behavioral Risk Factor Surveillance System (BRFSS) state survey samples. The study is focused on listed cell phone samples for a couple of states linked to socio-demographic data from the American Community Survey (ACS) at the Zip Code or county level. We also explore the use of the county-level Social Vulnerability Index (SVI) variables as simple predictors in the models. The models may lead not only to more effective nonresponse weight adjustments but also to efficient targeting of subgroups that may be more likely or less likely to participate in telephone health surveys.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2022 program