Social Network Methods for Estimating Adult Mortality: Evidence from Brazil

Dennis Feehan, Princeton University
Matthew J. Salganik, Princeton University

Measurements of adult mortality are a vital part of understanding the health and well-being of populations everywhere. In countries that lack high-quality death registration data, including most of Sub-Saharan Africa and Asia, adult death rates must be estimated using alternative strategies. This study aims to enlarge and improve our arsenal of methods for doing so. We introduce the network method, a new, survey-based technique that can be used to estimate adult death rates from respondents' reports about their social networks. We test this method using a nationally-representative survey of 25,000 Brazilians that was conducted in 2012. We will be able to compare the estimates from the network method to high-quality death rates obtained for Brazil's 27 state capitals from the vital registration system and the 2010 Census. Furthermore, we compare the performance of the network method to the sibling survival method, a widely used alternative.