In India, where sons have traditionally been strongly preferred to daughters, low-parity women with no sons have a significantly elevated risk of having their marriage dissolve.1 Using data from the first two National Family Health Surveys, researchers found that the odds of marital dissolution are about 40% higher among women whose only child is a girl than among those who have only a son; for mothers of two children, those with daughters have 70% higher odds of being divorced or separated than those with sons. Moreover, the association between having only a daughter and marital dissolution holds for women in most demographic, social and geographic subgroups.

The surveys, conducted in 1992-1993 and 1998-1999 among nationally representative samples of ever-married women of reproductive age, collected information on marital dissolution (divorce, separation and desertion) and many theoretically important covariates. To assess the relationship between marital dissolution and the sex composition of a woman's children, the analysts examined data on nonwidowed women aged 25 or older who had been married only once and had had at least one child.

Most of the 116,498 women included in the analyses were Hindu (81%), lived in rural areas (67%) and were not members of scheduled castes (75%); half lived in the country's northern region, and the rest were about evenly divided between residents of the East and South. Ten percent had been born in the 1940s, 34% in the 1950s, 45% in the 1960s and 12% in the 1970s. On average, women had been 17.5 years old when they married; they had had a mean of 3.7 years of formal education, and their husbands had had 6.3. One in 10 women had had one child, one in four had had two and the remainder had had three or more.

Overall, 2% of women were divorced or separated. The proportion varied from 1% to 3% according to women's background characteristics, but it ranged more widely by parity. Some 6-7% of women with one child were no longer married, compared with 2-3% of those with two children and 1% of mothers of three or more. The data suggest that the risk of marital dissolution was higher for women with daughters only than for those with at least one son; the researchers used logistic regression to examine this relationship, controlling for women's background characteristics.

In the multivariate analysis, women with one daughter and no sons had significantly higher odds of being divorced or separated than those with only a son (odds ratio, 1.4). Non-Hindu women had an increased likelihood of marital dissolution (2.1), as did residents of the South and the East (2.7 and 2.0, respectively). The odds of marital dissolution fell by about 5% for each year older a woman was when she married and for each year of her or her husband's education. Analyses including interaction terms revealed that the association between having only a daughter and divorce or separation did not vary significantly by a woman's education, religion, caste or residence (region or urban vs. rural). However, the association was weaker for women born in the 1960s than for those born in the 1940s; the researchers note that although this finding could indicate an erosion of the association over time, it also could reflect differences by age or marital duration.

Associations between most of the control variables and the risk of marital dissolution for women at higher parities were similar to those for women with one child. In addition, among mothers of two or more children, urban women and members of a scheduled caste or tribe were at increased risk of marital dissolution. When all of these variables were taken into account, women with two daughters and no sons had a significantly higher likelihood of being divorced or separated than those with two sons and no daughters (odds ratio, 1.7); the risk of marital dissolution for women with one child of each sex was not significantly different from the risk for women with two sons.

Among mothers of three children, women with only sons had the same risk of marital dissolution as those with only daughters or children of each sex; however, when the reference group was women with two sons and one daughter, women with three daughters had about a 60% increase in odds of divorce or separation. No significant associations were found at parity four or higher.

Given the long-standing preference for sons in Indian society, the researchers were not surprised to learn that "having at least one son is associated with a statistically significant and substantively important reduction in the risk of marital disruption among Indian women at lower parities." They were, however, struck by "the pervasiveness of the divorce-inhibiting effect of having sons across subgroups of Indian women"; this finding leads them to conclude that the relationship between marital dissolution and the sex composition of children is unlikely to change in the near future, despite economic development, increased urbanization and women's educational advances in India.

—D. Hollander

1. Bose S and South SJ, Sex composition of children and marital disruption in India, Journal of Marriage and Family, 2003, 65(4):996-1006.