Accounting for Spatial Anonymization in DHS Household Surveys
The household surveys conducted by the Demographic and Health Surveys (DHS) Program are a key source of data for estimating demographic and health indicators at a fine spatial scale in low- and middle-income countries. The surveys provide jittered coordinates for visited clusters, where coordinates are randomly displaced according to a known jittering distribution. We develop a geostatistical model that accounts for the positional uncertainty when analysing DHS surveys, and provide a fast implementation using Template Model Builder. The approach works for both Gaussian and non-Gaussian likelihoods. A simulation study with binomial data shows that the new approach performs equally or better than the common practice of not accounting for jittering, both in terms of more accurate parameter estimates and improved predictive measures. The improvement increases with increased amount of jittering. An analysis of contraceptive use in Kenya shows that the approach is fast and easy to use in practice.
READ FULL TEXT