Online Program Home
My Program

Abstract Details

Activity Number: 354 - SPEED: Big Data, Small Area Estimation, and Methodological Innovations Under Development, Part 2
Type: Contributed
Date/Time: Tuesday, July 30, 2019 : 10:30 AM to 11:15 AM
Sponsor: Survey Research Methods Section
Abstract #307755
Title: Benchmarking Mobile App Geofenced Samples: Adjusting for National Coverage and Selection Bias
Author(s): Davia Moyse* and YangYang Deng and Matt Jans and Ronaldo Iachan and Richard (Lee) Harding and Kristie Healey and James Dayton and Scott Worthge and Laura O'Campo
Companies: ICF and ICF Macro, Inc. and ICF and ICF Macro, Inc. and ICF and ICF and ICF and MFour Mobile Research and MFour Mobile Research
Keywords: nonprobability surveys; mobile phones; innovative data collection methods; coverage and selection bias; raking; nonprobability benchmarking

Mobile phone technology and geolocation advances have made it simple to locate survey respondents in locations outside of their home, such as while shopping or passing a store. As this nonprobability (NP) sampling method relies on smartphone ownership and specific store presence, demographic and geographic coverage or selection biases, common to probability (P) surveys of the general population, may be exaggerated. To explore this, we sampled mobile panel members in the U.S. when they entered geofenced areas around grocery, convenience, and home improvement stores, and asked them health-related questions. This paper discusses the demographic and geographic attributes of the NP geofenced respondents relative to population totals and a gold-standard P sample health survey. We compare four raking approaches to account for the potential biases: along basic demographics; along expanded demographics; along demographic and geographic dimensions as independent margins; and along controlled, cross-classified margins of geographic by demographic characteristics. The four methods are evaluated based on the distribution of the raked weights and by benchmarking weighted estimates of survey responses to a comparable P survey.

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

Back to the full JSM 2019 program