Campus & Community

Whites more likely than blacks to die soon after spouse’s death

2 min read

Study of 410,272 elderly American couples shows little ‘widowhood effect’ among blacks

A longitudinal study of 410,272 elderly American couples indicates that the “widowhood effect” – the increased probability of death among new widows and widowers – is large and enduring among white couples but undetectable among black couples, suggesting that blacks may somehow manage to extend marriage’s well-documented health benefits into widowhood.

The results, by Harvard sociologists Felix Elwert and Nicholas A. Christakis, are published in the February 2006 issue of American Sociological Review.

“The health effects of a spouse’s death differ radically between blacks and whites,” says Elwert, a doctoral student in sociology. “We found strong evidence of the widowhood effect among white couples: Men were 18 percent more likely to die shortly after their wives’ deaths, and women were 16 percent more likely to die shortly after their husbands’ deaths. By contrast, the estimated effect of a black spouse’s death on the mortality of his or her surviving spouse is essentially zero.”

Upon marrying, blacks and whites appear to receive the same health benefits, which previous research has attributed to factors such as emotional support, economic well-being, caretaking when ill, enhanced social support and kinship, and the promoting of healthy behaviors and discouraging of taking risks. Elwert and Christakis suggest such benefits may be longer-lasting for blacks, persisting even after a spouse’s death.

Citing prior research, the investigators identify several possible reasons for this enduring marriage benefit among blacks. Almost twice as likely to live with relatives and far more active in religious organizations, elderly blacks tend to have stronger and more extensive social networks than elderly whites. Black couples are also less likely than whites to adhere to a rigidly gendered division of labor, which may reduce mutual dependence.