Sibling-Linked Data in the Demographic and Health Surveys

This paper highlights an aspect of the enormous and little-exploited potential of the Demographic and Health Surveys, namely the use of data on siblings. Such data can be used to control for family-level unobserved heterogeneity that might confound the relationship of interest and to study correlations in sibling outcomes. These uses are illustrated with examples. The paper ends with a discussion of potential problems associated with the sibling data being derived from retrospective fertility histories of mothers.