Un modelo Fay-Herriot espacial para estimar la tasa de embarazo adolescente en Colombia

Abstract

Achieving estimates at a higher level of disaggregation means that estimation models of small areas acquire great relevance in the analysis of data that come from surveys with smaller scopes. This document is a research exercise whose objective is to ensure that the estimates of the adolescent pregnancy rate in the departments of Colombia produced by the 2015 National Survey of Demography and Health of Colombia can be extrapolated to all municipalities in the country. To this end, an adaptation of the spatial Fay-Herriot model that incorporates information at the municipal level on the pregnancy rate and administrative information on the characteristics of the municipalities of the country will be used. The obtained estimations were compared with those derived from the direct estimations in the same degree of aggregation of each data source, observing improvement in terms of accuracy.