Anderson, Barbara A. (2004). Considerations and recommendations in undertaking population projections for South Africa intended to assess the impact and consequences of HIV / AIDS. PSC Research Report No. 04-561
This paper discusses the use of population projections for South Africa in estimating the future demographic impact of HIV/AIDS. These issues raised apply to projections generally and to projections of the impact of HIV/AIDS in other countries. A focus of this paper is examination of the assumptions underlying projections and assessment of the robustness of the projections to violations of these assumptions. It is prudent to set ranges of assumptions for a wide variety of factors that influence the results of a population projection. It would be an unmanageable task to do projections that used all possible combinations of assumptions. It is suggested that for each factor considered, preliminary analysis be undertaken to determine whether the reasonable range of assumptions for that factor is likely to make a substantial difference in the outcome of a projection, when other factors are held constant. If variations in a given factor make little difference, then the work can proceed with a single value assumed for that factor. Thus, preliminary analysis is actually a sensitivity analysis that should lead to final projections being carried out under a manageable number of combinations of assumptions.
PSC Research Report No. 04-561
Anderson, Barbara A.