Effects of Family Composition on Mortality Differentials by Sex Among Children in Matlab, Bangladesh

From AcaWiki
Jump to: navigation, search

Citation: Muhuri, Preston, Samuel H. (1991) Effects of Family Composition on Mortality Differentials by Sex Among Children in Matlab, Bangladesh. PDR (RSS)
Internet Archive Scholar (search for fulltext): Effects of Family Composition on Mortality Differentials by Sex Among Children in Matlab, Bangladesh
Tagged: uw-madison (RSS), wisconsin (RSS), sociology (RSS), demography (RSS), prelim (RSS), qual (RSS), WisconsinDemographyPrelimAugust2009 (RSS)

Summary

The principal behaviors that are believed to produce sex differences in child mortality in South Asia are dietary practices and use of health facilities. Sons are generally preferred over daughters because sons have higher earning potential and are expected to assist in times of crisis or when parents are aged and no longer able to support themselves. However, there is evidence that girls too are valued. In this article, Muhuri and Preston examine the effect of sex composition of older siblings on the mortality of male and female children in Bangladesh. Moreover, they explore the association between numbers of children of a particular sex and mortality and explore the differences in association by socioeconomic status. Muhuri and Preston use data from the Matlab Demography Surveillance System and the 1982 Matlab census on over 14000 singleton births (index children) and the children's survival and migration status up to age 5 and the number of surviving older brothers and sisters at the birth of the index child. To assess the relative effects of the variables of interest on male and female mortality, they use logistic regression. They find that because parents cannot choose the sex of their children, the sex composition of children in a family is largely uncorrelated with other household characteristics. However, they find that girls with surviving older sisters (either 1 or 2 plus) have at every age higher death rates than girls without older siblings. It appears that girls' mortality is higher in families, which already have daughters. However, there is little evidence that having 2 or more older sisters is a greater disadvantage than having only one. Muhuri and Preston find a mild hint of excess male mortality in families with older sons. However, the possibility that having more than 1 older brother is worse than having only one is suggested in 6 of 9 comparisons. As these comparisons imply, female mortality exceeds male mortality to the greatest extent among sibships that contain older sisters. Male and female child mortality rates are relatively similar in families in which only older brothers are present, or in which there are no older siblings. However, if only older sisters are present, girls' mortality is more than double that of boys are all ages. Since children of higher birth order are more likely to have older sisters, the death rates of girls rise much more steeply with birth order than do the death rats of boys, which are largely unaffected by the presence of older siblings of either sex. According to the results of multivariate analysis, excess female mortality is highly concentrated among girls with one or more older sisters. Girls born into families with 2 or more sons have mortality that is significantly lower than mortality of boys born into such families. Also, boys born into families with 2 or more daughters have lower mortality. These results are consistent with the notion that boys and girls are not perfect substitutes for one another and that parents have, in a sense, separate targets for the numbers of sons and daughters. The addition of socioeconomic and residential variables has little effect on the family-composition coefficients. However, child mortality is significantly higher among families with very little dwelling space per capita, among families with few consumer items and that receive no remittances, and among families in which the mother had received no schooling. Children born in the treatment area of an important maternal and child health intervention have significantly lower mortality. Birth order has no significant independent influence on child mortality. The coefficient of the interaction between being female and having older sisters had the largest effect on mortality of all variables. The male pattern does not emerge until the index child has at least 2 older brothers and the coefficient is less than half of the female equivalent.