Reproductive preferences and subsequent behaviour in an rural area of Bangladesh

Abstract

This thesis reports on a study of married women in Matlab, a rural area of Bangladesh. On the basis of both cross-sectional (quantitative and qualitative) and longitudinal data from the treatment and comparison areas, the study investigated reproductive preferences and their determinants, intensity of desire for more children and contraceptive use, reliability of preferences in predicting subsequent fertility and change in preferences over time. Desired family size, ideal family size and desire for more children were similar in the treatment and comparison areas; sociocultural differentials in desired family size and desire for more children were also absent. Desired family size declined by about 25 per cent over the period 1975-90 in each area and the decline was more marked for sons than for daughters; the decline was homogeneous across the woman's age and number of living children categories. For those who changed their preferences from wanting more children in 1984 to wanting no more in 1990, the interim birth had no influence on subsequent change. There were no sociocultural differentials in use of contraception except by religion in the treatment area among those who wanted no more children either in 1984 or 1990, while in the comparison area there were differentials by religion in 1984 and by education of woman in 1990. Among women who wanted a child after one year, the sociocultural differential in use of contraception existed in the treatment area in 1984 but disappeared in 1990 while in the comparison area the differentials existed by possession of items in 1984 and by education of woman in 1990. Among those who wanted a child after one year, use of contraception was lower for those who had no son than for those who had one or more sons in each year in the treatment area, but only in 1990 in the comparison area. The abortion ratio was higher for those who wanted no more children than for those who wanted more, and higher in the comparison area than in the treatment area. At the aggregate level intended fertility and actual fertility were closer in the treatment area than in the comparison area. However, accuracy at the aggregate level depends to a large extent on many counterbalancing inconsistencies on the part of individuals. In both the areas those who wanted no more children had fewer subsequent births than those who wanted more, while those who wanted no more children had fewer births if their husbands also wanted no more than if the husbands wanted more. In the comparison area subsequent births occurred homogeneously across sociocultural subgroups, but in the treatment area there were fewer births among the better-off, among users of contraception and among Hindus. For women who wanted the next child 'soon' the inconsistency levels were similar in the two areas, but for those who wanted the next child 'not so early' the inconsistency level was higher in the comparison area than in the treatment area; however, sociocultural differentials in inconsistency were largely absent.