Urban and rural fertility in Bangladesh: a causal approach.

Fertility models are constructed from the 1989 Bangladesh Fertility Survey (BFS) employing path analysis. These models are developed and interpreted for urban and rural situations. As a proxy for fertility, the number of children ever born is used, and age, religion, age at marriage, parental childhood residence, and education are considered as explanatory variables. The contribution that these variables give to explaining the 1989 Bangladeshi fertility is compared to the explanatory variables that Ahmed (1981) found suitable for Bangladeshi fertility in 1975. We find that in 1989, compared to 1975, childhood background and education of the mother and age at marriage exert a greater influence on urban fertility, and religion no longer has a significant effect. In the rural case, the effect of religion on fertility has increased since 1975, as has education and age at marriage.