Resumen:
The article illustrates the value of geocoding demographic surveys and conducting spatial analyses to understand the service supply environment and contraceptive behavior. Three Costa Rican data sets are geocoded and pooled in a Geographic
Information System (GIS): a demographic survey, an inventory of family planning facilities, and a census. Displaying survey's results on maps enhances the understanding of the spatial configuration of family planning services and users' behavior. Trend surface analyses improves survey's estimates for small areas and pin-points spatial differences. Cartographic-based measures of accessibility and of contextual characteristics have an edge in objectivity, comparability and flexibility. Multilevel models on contraceptive use and method choice suggest mixed effects of density of services and diffusion effects from neighbors. A model for the choice of family planning outlet arrives at a classic gravity formulation in which larger and closer clinics are more likely to be chosen. Demographic surveys should consider geocoding their sampling units as a routine procedure.