Son preference and child replacement in Bangladesh: a new look at the child survival hypothesis

Birth history data from women in the 1975-76 Bangladesh Fertility Survey were used to search for intentions to replace dead children. The median intervals between successive births of orders (i) and (i + 1) were not shorter when some siblings of orders below (i) had died. Nor was the median duration between the death of a child and the first posthumous birth shorter when the dead child was a boy or when it was survived by fewer than two brothers. The median intervals were generally shorter when the mother lived in an urban rather than a rural area but this difference was attributable only to the shorter duration of breast-feeding by urban women. These results disputed the notions that the timing of births was deliberately quicker to replace a dead child, that attempts at replacement were sex-selective, or that child replacement intentions were stronger in urban than in rural populations.