Gender Differences in Child Health: Evidence from the Demographic and Health Surveys

Abstract

The study of data on sex differences in mortality and morbidity obtained from Demographic and Health Surveys (DHS) indicated the weaknesses in using cross-national aggregate data for examination of gender differences. The data did confirm a continuing higher mortality among girls in developing countries. The standard for measuring excess female mortality was based on societies where discrimination against girls appeared to be small. A life table method using locally weighted least squares techniques was used to calculate excess mortality. At high levels of under-5 male mortality (a probability of dying by age 5 of 300/1000 live births) the female advantage was 10% among those aged under 5 years 15% in infancy and 3% between ages 1 and 5 years. At low levels of under-5 male mortality (100/1000 live births) the female advantage increased to 18% among those aged under 5 years 21% in infancy and 10% between ages 1 and 5 years. Data quality issues included the potentially wide confidence intervals and small sample sizes. A comparison of DHS data on female disadvantage with other data for 35 countries excluding China and India showed general agreement except for Trinidad and Tobago Indonesia Sri Lanka Thailand Jordan and Pakistan. Female mortality was higher than male mortality in 21 countries. The ratio of male-to-female infant mortality had a median difference of 0.17 between the observed and standard discrimination ratios in the Middle Eastern crescent. Discrimination was more evident in mortality among girls aged 1-4 years and the overall median disadvantage was 0.11. For under-5 mortality the disadvantage was 0.06 and highest in the Middle East. About 13% of the variance in female mortality disadvantage was accounted for by the coefficient for immunization. About 13% of the female mortality disadvantage was explained by female disadvantage in treatment for diarrhea. Female mortality disadvantage was unrelated to sickness rates or stunting or wasting but a positive association was found with relative lack of immunization among girls and a negative association was found with diarrhea treatment rates. Female enrollment rates in primary school were lower than male rates. Multivariate analysis found that variance in female mortality disadvantage was most strongly explained by primary school enrollment and immunization; education immunization and diarrhea treatment individually accounted for little of the variance.