An estimated 9 million Americans -- or nearly 4 percent of the total population -- say they identify as lesbian, gay, bisexual or transgender, according to a new report released this week from the Williams Institute, a think-tank devoted to LGBT research at UCLA.
Bisexuals make up slightly more than half that group, 1.8 percent of the total U.S. population, and they are substantially more likely to be women than men.
The report is the most up-to-date assessment of that population and produced a lower population percentage than the 10 percent number that advocacy groups have used in the past, which was based on Alfred Kinsey studies from 1948.
The new data comes on the heels of another recent report published by the Institute of Medicine for the National Institutes of Health emphasizing the need for more federally funded research on LGBT health problems.
"Sexual orientation is complex, but measurable," said Gary J. Gates, chief researcher and a Williams Distinguished Scholar. "Hopefully, this will begin to prompt some dialogue on what it means when we say LGBT."
Other key findings were that an estimated 19 million Americans, or 8.2 percent of the population, said they have engaged in same-sex behavior, and 25.6 million, or 11 percent, acknowledged some same-sex attraction.
Gay advocacy groups are hailing the report as a critical first step to inform public policy, research and federal funding. They say the information is crucial in identifying health and economic disparities, discrimination, domestic partnership benefits and the impact of same-sex marriage.
The report was based on a collection of previous surveys in the United States and around the world.
"All were surveys based on population and we tried to take precautions to minimize biases," said Gates.
"Many people still have the 10 percent number in their head," he said of Kinsey, who published his studies in sexual behavior at Indiana University in the 1940s and 1950s.
"Kinsey didn't use a population-based sample, so regardless of the point-estimate, it came from a self-selected group of men," said Gates. "For example, prisoners represented a higher proportion of men in his sample than they would in the population, perhaps biasing the sample toward more same-sex sexual behavior."
Gates said the lower number did not surprise researchers, but it might be new to the general population.
Gay advocates latched on to the Kinsey number in the late 1960s and 1970s, according to Gates.
"That 10 percent emerges as much a decision of politics as a decision of science, which is not to say Kinsey was not a scholar and scientist," Gates said.
But, he added, it was "kind of a smart number -- big enough to kind of matter and not so large that it was threatening to a population that was uncomfortable with it.
"You can argue it was a brilliant strategy and it seemed to resonate with people," he said. "One of the things the study shows, perhaps, is a little bit of maturity of the movement. We can now use more traditional ways to assess the population."
The 3.5 percent of those who identify as LGB may or may not include those who are "closeted," according to Gates.
"We actually did commission data within the survey and asked about to what degree they were closeted," Gates said, "and 13 percent who identified as LGB had never told anyone about it."
The most interesting results, according to Gates, were the data on bisexuals, which made up about half of the overall LGB population. Of those who identified as such, 25 percent said they were closeted.
"Many are still quite discreet about their sexual orientation," he said. "If you ask people in the bisexual population, they will tell you it's a different kind of a stigma they experience. Some don't feel completely at home in the LG community because they often think of them as being way too gay and hold them with a level of suspicion. And in general, they are subject to the general stigma of same-sex behavior."
Gates also cautioned against the results that 11 percent of all Americans had same-sex attractions.
"That was only from one survey and it was restricted to 18- to 44-year-olds," he said.
Respondents had to rate their attraction from exclusive opposite-sex attraction (5) in graduating degrees of same-sex feelings (4 to 1).
"Some in category 4 may be people like me who avoid 1s and 5s in surveys," Gates said.
The U.S. Census does not count how many Americans identify as LGBT, although it will release later this year its count of same-sex spouses.
Richard Socarides, president of the gay advocacy group Equality Matters, said the Williams report is a good tool, but the government should be tracking LGBT Americans in the census.
When he served as special assistant to President Bill Clinton on LGBT policy issue, Socarides said, he urged the White House to "count gay people."
"They still won't do it, and now they are counting same-sex couples who live together, but that's not counting," he said. "Just as the census does a survey of race and ethnicity, they ought to do sexual orientation."
He praised the Williams Institute but said the latest study is "at best an educated guess."
"They take a bunch of surveys that are out there and made some judgment based on the available information," said Socarides.
What the results do show, Socarides said, is "absolute fluidity" in sexual orientation, "especially for young people who don't have the kind of stereotypes that adults do."
The impact of studies like these will be far reaching, according to Socarides.
"It's important first, because the country is in the middle of making important policy decisions about the rights of gay people and, in that context, it's important to know how many Americans are affected," he said. "Another reason for the government to know is funding issues tied to health and education.
"Now that sexual orientation has essentially come out of the closet and people are not so fearful," he added, "you'll see a greater willingness to count gay people and we really need reliable information."