A Comparison of Geosampling and Random Walk Methods for Household Sample Selection in Uttar Pradesh, India. (25th September 2021)

• Record Type: Journal Article
• Authors: Lau, Charles Q; Unangst, Jennifer; Eckman, Stephanie; Bhatt, Pramod; Evans, Jonathan; McRoy, Martha; Starr, Kelsey; Thalji, Lisa; Allpress, Justine; Jones, Kasey; Cajka, Jamie; Bunker, John David
• Is Part Of: International journal of public opinion research. Volume 33:Number 4(2021)
• Page Start: 1062
• Page End: 1073
• Publication Date: 2021-09-25
• Subjects: sampling -- household surveys -- random walk -- bias -- variance; Public opinion -- Addresses, essays, lectures -- Periodicals; 303.38
• DOI: 10.1093/ijpor/edab020
• Languages: English
• ISSNs: 0954-2892

Abstract

Abstract: Our research evaluates an innovative sampling technique for household surveys called "geosampling" which leverages recent advances in geographic information systems, computer vision algorithms, and satellite imagery. We compare geosampling to the random walk method. We conducted two surveys in Uttar Pradesh, India: one using geosampling (1, 026 completes) and another using random walk (939 completes). We compare the two sampling techniques along three dimensions: (a) performance indicators—response rates and contact attempts; (b) sample composition; and (c) components of variance. We help researchers understand the survey contexts for which geosampling and random walk are best suited.