Authors: Stuart Sweeney*, University Of California Santa Barbara, Jessica Marter-Kenyon, University of Georgia
Topics: Field Methods, Africa, Population Geography
Keywords: Sampling, Rwanda, Household Survey, Rural
Session Type: Paper
Start / End Time: 3:05 PM / 4:45 PM
Room: 8228, Park Tower Suites, Marriott, Lobby Level
Presentation File: No File Uploaded
The goal of any household survey is to measure household attributes in such a way that they provide an unbiased characterization of households in the target population. The most direct approach is to use simple random sampling based on a complete registry of households in the target population. In rural Africa, household registries are unlikely to exist and simple sampling rules of thumb (like every xth house on a road) will miss rural households that are not located in rural villages or near roads. We review a sampling design that we specifically designed to capture households in villages and in more remote areas. The umbrella project centered around analysis of a villagization policy, so including non-village households was critical. We develop a spatial sampling approach using a spatial inhibition process combined with aggregate data on certain areal characteristics of Rwandan districts and sectors. After reviewing the sampling algorithm, we describe the resampling approach used to design inverse probability of selection weights. We then evaluate bias and accuracy of the sample with respect to other large-scale survey information (census and Demographic and Health Survey). The paper closes with a description of lessons learned during the implementation phase of the survey.