Handbook for Estimating Demographic Statistics from Incomplete Data in Developing Countries

Abstract

For the demographer analyzing incomplete data this manual is meant to help in the following areas: 1) the basis for theoretical methods and technics; 2) the calculations used to apply this theory translating it into concrete figures; and 3) the mortality models used. The main aim is to list in one place the basic hypotheses underlying each mathematical formula or model under consideration its use and field of application. Researchers then can concentrate on the practical and operational side of the subject. The handbook is divided into 3 parts: 1) mathematical relations 2) application to observed data and 3) the main model life tables. The manual is not a comprehensive treatise on mathematical demography. It presents those relations which are likely to lead to practical application in the field. Matrices are omitted since current adjustment methods do not use them. Part II expresses the practical external mathematical relationships as adequately as possible on the basis of discontinuous data. Part III is deliberately restricted to mortality models because of the fundamental role of mortality in the demographic process the complexity of the phenomenon which occurs at all ages and the difficulty of making observations and the danger of errors and lacunae. These are descriptive (not causal) and determinist (not stochastic) models. However the factors causing current mortality decline in the Third World are very different from those which prevailed in Europe at the time of industrialization. Therefore the authors attempted to overcome the fact that these tables present a poor picture of observed or foreseeable trends. The main purpose of these models is to improve knowledge of mortality rates in populations where statistics are defective; so to be truly applicable they must take into account the particularities of that mortality. The authors recommend successive approximations to test or adjust them. To make it easier for the newcomer they have tried to set out a comparative synthesis as objective and honest as possible of the large number of conceptually varied models described.