Abstract
This paper is a critical assessment of techniques used for the estimation of cause of death structure in Africa in the light of contemporary theories of mortality. The technique mostly used for the indirect estimation of cause of death structure is based on the mortality experience in Western countries. With the marked differences in the epidemiologic transition experienced in Western countries and those being experienced in African countries, it is not very likely that an estimation of cause of death structure based on data from the former can accurately estimate cause of death structure in the latter. This argument is supported in the paper through a comparison of observed estimates of cause of death structure for some African populations with indirect estimates. It is proposed that a better approach is to use a residual method whereby one proceeds from nearly true estimates of cause of death rates to obtain estimates for unknown causes via the known general mortality.