An analysis of the child survival hypothesis in Jordan.

Data from the 1976 Jordan Fertility Survey are used to examine the association between infant mortality and fertility behavior. After controlling for socioeconomic and demographic variables, a strong positive association was found between infant mortality and fertility behavior; women who experience infant deaths tend to make up for the loss of children. However, in terms of surviving children (surviving to age one year), the replacement effect is far from complete. This study also shows that reporting or recording deaths as having occurred after infancy, when in fact they have occurred before age 12 months, tends to result in an underestimate of the mortality-fertility association. However, the bias is statistically significant (at the 0.05 level) only when 50 percent or more of infant deaths are so misreported.