Skip to main content
Part of complete coverage on

Census Bureau revises down same-sex couples figures

By the CNN Wire Staff
updated 6:18 PM EDT, Tue September 27, 2011
Jodie Vandermark-Martinez slides a ring on her partner, Jessica Vandermark-Martinez, during their wedding ceremony in Iowa.
Jodie Vandermark-Martinez slides a ring on her partner, Jessica Vandermark-Martinez, during their wedding ceremony in Iowa.
STORY HIGHLIGHTS
  • NEW: Census Bureau provides "preferred estimates" that lower same-sex couples figures
  • NEW: Obstacles to an accurate count include names used by both genders
  • NEW: The revised estimates identify about 131,000 married same-sex couples

Washington (CNN) -- Counting the number of American households with same-sex couples proved difficult, the U.S. Census Bureau said Tuesday in presenting revised estimates for its 2010 figures.

According to the 2010 survey, the total number of households with same-sex couples exceeds 900,000.

However, revisions due to errors in the sampling process led to a "preferred estimate" of 646,464, the Census Bureau announced.

The revised figures still show a huge increase in overall numbers of households with same-sex couples over findings in the previous census a decade earlier, said Martin O'Connell, who heads the Census Bureau's fertility and family statistics branch.

Overall, the households with same-sex couples are far below 1% of total U.S. households, the Census Bureau said.

The revised figures are much closer to those compiled by the American Community Survey in its 2010 survey.

According to the Census Bureau, the revised estimate of married same-sex households is 131,729, well below the initial count of 349,377. The revised estimate of same-sex households with unmarried partners was 514,735, much closer to the original census count of 552,620.

The breakdown in the preferred estimate is 64,223 married male couples and 67,506 married female couples.

O'Connell explained how the initial census figures were fine-tuned to try to eliminate errors in the sampling process.

The errors involved various aspects of the process.

For example, a questionnaire sent by mail was easier to use and produced a more accurate response than a form presented by house-to-house surveyors, O'Connell said. In particular, the door-to-door form caused some participants to check the wrong box in identifying their gender, he said.

In addition, the Census Bureau tried to screen out possible mistakes in gender identification involving participants with first names such as Taylor or Morgan, which traditionally involved males but now also are common among females.

While the 2010 figures reflect increases in same-sex households, they don't provide a detailed analysis of the evolving impact of legalized same-sex marriage in some states, O'Connell said.

"We know where they're living now, but we don't know when they got married or if they got married in that particular state," he said.

ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT