Search

The purpose of this paper is to analyze the differentials and detect factors affecting infant mortality on the basis of data obtained from the fertility surveys from those countries participating in the World Fertility Survey. In particular, this includes the surveys carried out in Colombia, Peru, Costa Rica, Panama, and the Dominican Republic. 3 types of explanatory variables may be considered from the information available: 1) context variables related to the mother's environment; 2) socioeconomic variables based on the educational and economic characteristics of the mother and her last husband; and 3) biological factors (from each woman's pregnancy history) such as mother's age at birth of the child, order of birth, interbirth interval, etc. The countries, whether high or low mortality, present great differences in child mortality in most of the variables considered. In Panama and Costa Rica there are population sectors with infant mortality rates of around 100/1000 live births, whereas in Peru these are over 150/1000 (children from mothers without education, low agricultural strata, etc.). Besides presenting the differentials, a methodological test is made through the application to Costa Rica and Peru of the Proportional Hazards Model which permits analysis of the effects of variables when acting simultaneously upon mortality in early childhood. The variables which show the highest disparity in mortality level are: natural region among the context variables, education of mother among the socioeconomic variables, and interbirth interval and maternal age at birth of their children among the biological ones.

"A procedure is described to obtain an age distribution of emigrant surviving children from the reports of such children obtained from women by a census or survey. A simple form of the procedure, requiring a minimum of calculation, is shown to work almost as well as the more elaborate form, and is recommended for most uses. "These new procedures, explicitly introducing an age model of migration, yield age distributions which are substantially different from, and probably more realistic than those obtained through the use of model fertility and mortality schedules alone.... The age distributions of migrant children [in Colombia] as obtained in this paper are compared with those obtained by Somoza.... The procedure described here produces a total almost 10 percent higher because it shifts the age distribution of the migrant children towards higher ages with lower proportions of surviving mothers." (summary in ENG), (excerpt)