Adult Mortality from Sibling Survival Data : Does the Corrected Method Perform Better?

Masquelier, Bruno [UCL]

Due to the lack of complete registration of deaths in most countries of sub-Saharan Africa, adult mortality is still measured through unconventional techniques. Estimates based on sibling survival have long been deemed implausibly low, but they have received increasing acceptance in recent years. They can provide valuable counterpoints to model-based estimates, which are typically derived from childhood mortality and standard age patterns of mortality. This more optimistic view in the literature is partly due to the work of Gakidou and King (2006). The weighting scheme they suggest to correct for selection biases in sibling histories has been applied to DHS surveys, and it yields much higher estimates than previous calculations based on the same data. After reviewing the main features of this procedure, this paper offers a methodological critique of its application to DHS data. Microsimulations are used to demonstrate that the "Corrected Sibling Survival" method may substantially overestimate mortality rates, especially among males.