There are far fewer same-sex couples in South Carolina and across the U.S. than 2010 census data first indicated, the Census Bureau said Tuesday.
The number of same-sex couples in South Carolina who are married is now estimated to be less than one-third the number previously reported.
Revised census estimates say that in 2010 there were 1,596 same-sex married couples in South Carolina -- not 5,300, as was reported earlier. Statewide there were 7,214 same-sex households, with either married or unmarried partners, not 11,532.
The Census Bureau put most of the blame for the errors on census data collectors who filled out forms while interviewing couples and checked off the wrong boxes for "male" or "female" often enough to dramatically skew the statistics.
For example, if a form said Thomas and his spouse, Martha, were both female, they would have been erroneously counted as a same-sex married couple.
Martin O'Connell, chief of the Fertility and Family Statistics Branch at the Census Bureau, said the error rate was small for the total population, but had a large impact on the census numbers for same-sex couples because of the small subset of the population involved.
Same-sex couples in 2010 made up just over half of 1 percent of U.S. households, based on revised estimates.
The revisions are likely to reignite debate about whether census reports accurately reflect the size of the nation's gay and lesbian population.
"Typically same-gender couples would report in lower numbers, simply out of fear," said Warren Redman-Gress, director of Alliance for Full Acceptance, a Charleston-based gay rights advocacy group.
"After the 2000 census, the first time the so-called lavender question was asked, it was estimated that same-sex couples were underreported by 60 percent," he said, citing the Human Rights Campaign and Pew Research Center.
But the Census Bureau has concluded that the same-sex households were greatly overreported in the 2000 census.
The bureau now says that more than 80 percent of reported same-sex married couples in the 2000 census were probably opposite-sex couples.
The revised census estimates were arrived at by attempting to correct the gender errors. The Census Bureau determined which names were very likely male or female and compared them with the gender reporting.
If a form said "Thomas" was female, that would have been changed. If the name were more gender-ambiguous, such as Pat, no change was made, O'Connell said.
Nationwide, the Census Bureau now says there were 646,464 same-sex households in 2010, not 901,997.
The Census Bureau also revised figures from the 2000 census, and says there were 358,930 same-sex households in 2000, not 594,391.
Redman-Gress said that in any case "it seems that same-sex couples make up a significant portion of the population and deserve the same protection as opposite-gender couples."
While the 2010 census data released in August overreported the number of same-sex households nationally by 28 percent, the overreporting was higher in some states, and particularly in states with relatively small populations. South Carolina's same-sex households were overcounted by 37 percent.
O'Connell said the errors were just discovered this year, by comparing 2010 census data with the annual surveys conducted by the Census Bureau.
"It took about two months to figure out what, exactly, was the problem," he said.
Staff writer Adam Parker contributed to this report.Reach David Slade at 937-5552.