Census Numbers Differ, But Divorce Consequences the Same

Because of the negligence of many states in submitting vital statistics to the National Center for Health Statistics (NCHS), national reports of the number of marriages and divorces since the late 1990s have been based on incomplete state records. In an effort to obtain a more accurate accounting, the Census Bureau recently began including marital history items to its American Community Survey (ACS), which polls about 2 million households annually.

The new ACS data, based on monthly surveys conducted throughout 2008 and 2009, suggest a higher number of marriages and divorces relative to NCHS vital-statistics reports, a plausible outcome given the limitations of current NCHS data. At the same time, ACS data claim that a greater number of men marry than do women, while a greater number of women divorce than do men. For example, ACS data collected throughout 2009 found that 2.29 million men married and that 1.10 million men divorced, but that 2.21 million women married and 1.22 million women divorced, in the previous twelve months. For comparison purposes, the NCHS claims that 2.08 million marriages and 1.04 million divorces occurred in 2009.

This discrepancy between men and women points to serious problems with the ACS data. How could a large, nationally representative survey find that nearly 78,000 more men marry than women? Or that nearly 121,000 more women divorced than men? The differences were even greater using data collected in 2008, as the data claims that about 100,000 more men than women married, and that 131,000 more women than men divorced, during the previous twelve months.

The Census Bureau’s explanation is not at all credible, attributing the discrepancy to “gender differences in marriage.” Women, an August 2011 report claims, tend to live longer and tend to marry older men, while men tend to remarry more. All true, yet each marriage and divorce in America still represents two parties—a man and woman—who would ordinarily report the same event during the same polling year. Moreover, the report—parroting an earlier unpublished paper by the same authors—defends the ACS survey data as not only comparable to, but also more reliable than, data compiled by the National Vital Statistics System, which has historically represented actual counts of marriage certificates and divorce decrees issued by the states.

This journal is not persuaded that the ACS survey data can replace vital-statistics records; the solution is not a new survey but a greater commitment on the part of the NCHS and the states to ensure robust record collecting, keeping, and reporting of what are, after all, vital statistics.

Nonetheless, the 2011 ACS report deserves praise for acknowledging the immediate impact of divorce on economic well-being. Women who divorced, the report found, “were more likely to receive public assistance,” “reported less income,” and were more likely “to be in poverty,” relative to men who divorced in the previous year. Likewise, the children whose parents divorced last year “were more likely to be in a household below the poverty level,” “were more likely to be living in a rented home,” and “were more likely to be living with their parents’ unmarried partners,” relative to other children.

The report also notes that more than 1.1 million children suffered from their parents’ divorce in the previous year, a finding that ought to wake up Congress. Any other federal report finding that an additional million children are at risk every year of some disadvantage related to their health, education, or welfare would immediately prompt congressional hearings. Congress should do no less when it comes to negative outcomes of parental divorce on American children.

(Diana B. Elliot and Tavia Simmons, “Marital Events of Americans: 2009,” American Community Survey Reports, ACS-13, U.S. Census Bureau, August 2011; and Diana B. Elliot, Tavia Simmons, and Jamie M. Lewis, “Evaluation of the Marital Event Items on the ACS,” unpublished paper, .)