Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
NOTES ON The DATA As already noted, the data used are from the World Fertility Surveys for Colombia and Costa Rica, which obtained reproductive histories, including dates of birth and death of all children, for nationally representative samples of women in the reproductive ages. (These are described in detail in the First Country Reports, and the quality of the data have been explored in other WAS publications, c.f. the Scientific Reports.) Detailed information on breastfeeding and contraceptive use are, however, available only for the last closed and the open interval, and information on duration of contraceptive use was collected only in the two countries studied here. Despite being selected because of data availability, Colombia and Costa Rica are appropriate for a comparative study of the impact of breastfeeding and contraceptive use on reproduction because they are countries with relatively similar levels of fertility, but quite different patterns of breastfeeding and contraceptive use. In both countries, the WAS was taken in 1976; in both cases, 93 percent of last and next-to-last birth intervals are concentrated in the period since 1960. Yet these intervals are not representative of all birth intervals begun in this period; in fact, the farther from the survey date (i.e., the closer to 1960) an interval began, the less likely it is to be the last or even next-to-last in 1976. The situation is problematic because the estima- tion procedures used here require that we assume the childbearing process to have remained relatively unchanged over the entire period; a stationarity assumption that may not be met by the data because the period between 1960 and 1976 was one of considerable social, economic, and demographic change for both countries. In order to 4
understand potential sources of bin, we will examine Bone of these changes rather closely. Figure 1 sham the downward trends for the per iod 1960-76 in fertility (total fertility rate) and infant mortality (loo). The level of fertility and the pace of decline are similar in both countries, but infant mortality is somewhat rawer in Costa Rica than in Colom- bia. Figure 2 displays the age profile of fertility for both countries for To time points. The pattern of decline has been similar in both counts ies, resulting in a concentration of childbear ing at younger ages, that is perhaps more marked in Costa Rica than in Columbia. There has also been considerable social and economic change. While both Colombia and Costa Rica still have large age icultural sectors (with 30 and 22 percent, r espectively, of the gross domestic product being der ived from agriculture), urbanization and industrialization have both increased since 1960. The o~id-1960s also saw the development of family planning program in both countries. Although Colombia and Costa Rica are both about equally urbanized, per capits income and literacy are considerably higher in Costa Rica, as is contracep~ tive prevalence. In contrast, breastfeeding is more colon and of longer duration in Colombia than in Costa Rica. Thus, similar levels of fertility are obtained by quite different behavioral mechanisms . in boy counts ies, family planning clinics have tended to be concentrated in c ities, and contraceptives have been Are available to the urban population (see Sanin, 1976; Gomez and Bermudez, 1974) . These social and demographic changes af feet the representativeness of the last two birth intervals. As already noted, births occurring farther away from the survey date and associated with the last two intervals are less and less representative of all births for that period. In particular, births occurring far from the survey date will be associated with longer intervals Short intervals from earlier periods will be correspond dingly underrepresented. These long intervals will tend to be associated with births to women who are subfecund, older women, higher-parity women, and women using Contras caption. This last group is of same concern since these women must have started using contraception at a time when it was relatively unusual to do so. Therefore, they may be an atypically more ~modern. group for their period, and the ir ch ildbear ing exper fence may be closer to that of women who gave birth in the more recent period. In
6 TF~ 10. S.O IqO .10' .05 1. Colombla ___.Coste Eke No 1965 1970 19t5 . 1 96S 1970 197 S FIGURE 1 Total Fer tility Rates and Infant Mortality Rates, 1960-78 S our ces : National Research Council (1982); tabulations from the Costa Rica study by the Panel on Latin America, Committee on Population and Demography, National Academy of Sciences.
AS9R .4 .2 .1 ASS R .2 .1 //, //, / . ,- //,' / ,, . ! / / 7 Colombia 4W, //, _ _ _ 1961-2 1965-6 t970-1 1975-6 _ _ " "am\ `. "N '\,'N \ ~ 15-19 2~24 =-29 30-34 35-89 40-44 45-49 Costs Rice . 1960 .4 .3 l? ~ '- - -'`"-"2 ~ _ "_ _ 1966 1970 _._ 1975 ~ i: 15-19 20-24 25-29 30-34 35~39 40-44 45~49 FIGURE 2 Age-Specific Fertility Rates, 1960-75 Sources: National Research Council (1982); tabulations from the Costa Rica study by the Panel on Latin America, Committee on Population and Demography, National Academy of Sciences.
8 addition, modern coitus-independent contraceptive were, for the most part, not available until Me 1970s, which means that these longer intervals had to be achieved through the effective use of more traditional methods, and thus may bias upwards our estimates of the impact of the use of these methods on fertility. An attempt to assess these biases is made by progressively excluding births from earlier periods, repeating analyses and compar ing results . Thus, all last and next-to~las t births are analyzed first, followed by only those occurs ing since 1960, since 1965, and since 1970 .