this percentage among the lower two educational groups. The very different trends in early childbearing by school levels across regions underscores how the social context—even when measured as crudely as world region—modifies the effects of individual-level characteristics (such as schooling) on the transition to adulthood. In particular, not only is it likely that these crude schooling categories represent different things in different places but also that factors other than schooling are differentially important in different settings.
In Panel B of Table 8-4 we can see that for all regions of the developing world, whatever the direction of the overall change in the prevalence of early childbearing, the trends for women from both rural and urban areas were in a consistent direction. However, in regions experiencing declines, the absolute size of the declines were greater in urban than rural areas in sub-Saharan Africa and Southern Asia, leading to a widening gap between urban and rural areas in the prevalence of early parenthood in these regions. Early parenthood, however, remains a much more common phenomenon in rural areas. Urban-rural differentials in the percentage having a child before age 18 are particularly dramatic in Western and Middle Africa and in South-central and South-eastern Asia, where there are 14 percentage point differentials between rates of early childbearing in urban and rural areas. Unfortunately, we are not able to present differentials by household wealth, because household wealth is measured at the time of the survey and the births have occurred at various points in the past.
As previously discussed in Chapter 4, the evidence on the age pattern of maternal mortality is conflicting. While some evidence suggests that first parenthood below age 18 or 20 may carry more health risks than first parenthood at slightly older ages, other data suggest that the risks, if they exist at all, are more likely to be concentrated at younger ages (e.g., ages 15-16). Despite these uncertainties in the evidence, popular concern about childbearing below age 18 tends to focus on maternal health concerns (e.g., Save the Children, 2004; United Nations, 2004).
While more evidence is needed on whether adolescent mothers experience poorer outcomes than do older mothers, it is critical to note, particularly for health policy and programs, that maternal morbidity and mortality in the developing world for women at any age, including adolescents, are greatly influenced by poverty, poor nutrition, and limited access to medical services for problems of pregnancy and delivery. The data presented in Table 8-4 indicate that a higher percentage of rural and less well educated