Gay people count, so why not count them correctly?
By Gary J. Gates,
Back in the 1960s, pioneering gay activists found an obscure passage from a 1948 book written by prominent sex researcher Alfred Kinsey that read, “10 percent of the males are more or less exclusively homosexual . . . for at least three years between the ages of 16 and 55.” They used that quote to claim that 10 percent of the population was gay, even though Kinsey’s study was not designed to make a population-based estimate.
The motivation behind using the 10 percent figure was less about science and more about politics. In those days, gay activists needed to prove the very existence of a gay community. One in 10 was big enough to “matter.” It certainly mattered to me when, as a young, closeted gay man, I would look around a classroom with 50 people in it and think, “Wow, there are four other people here just like me.”
But the percentage was not so large as to overly threaten a society still extremely uncomfortable with the idea of gay people. The fact that the 1-in-10 figure still gets bandied about is a testament to the brilliance of this political strategy.
Lots of Americans have no idea how many people are gay or lesbian. A 2002Gallup poll suggested that one in six Americans had no estimate, and those who did have an opinion put the figure at a whopping 20 percent.
As a demographer who studies the lesbian, gay, bisexual and transgender community, I’ve been asked how many LGBT people there are more often than I can count. Politics may still play a role in why the answer is critically important, but there certainly is no longer a need to prove that gay people exist. Today, quantifying the population is about documenting how LGBT people live their lives. How many marry? How often do they have children? How many are serving in the military? How often do they experience discrimination?
These facts matter because legislatures, courts and voters across the country are debating how LGBT people should live their lives. All parties deserve to be informed by fresh research, not a six-decade-old study. We should be able to search the standard places where scholars and policy advocates go for information about the health and well-being of Americans — all Americans. Places such as the Census Bureau’s decennial count and American Community Survey, the premier sources of demographic data in this country. Or the National Health Interview Survey, a primary source of information about Americans’ health. Or the Current Population Survey, the preeminent source of information about the nation’s economic well-being. Or the National Crime Victimization Survey, where we get most of our data about experiences of crime.
But searching these sources for information about LGBT people would be largely futile. None ask questions about sexual orientation or gender identity.
I recently reviewed findings from 11 large surveys conducted since 2004, seven in the United States and four internationally. Averaging across the U.S.-based surveys, I found that nearly 9 million Americans (3.8 percent of adults) self-identify as LGBT. That’s equivalent to the population of New Jersey.
An estimated 19 million Americans (8.2 percent) report having engaged in some same-sex sexual behavior, and nearly 26 million (11 percent) report some same-sex sexual attraction. The latter figure is equivalent to the population of Texas.
But as a population scientist, I don’t want to have to comb research for pertinent data to average. I’ve attended dozens of meetings with representatives from federal statistical agencies to ask them why they are not counting the LGBT population. They tell me that they worry about survey respondents refusing to answer such questions or, even worse, terminating the survey. They also wonder exactly what questions to ask.
Should they count only those who explicitly identify themselves using terms such as “lesbian,” “gay” or “bisexual”? Or should they measure sexual behavior? Or sexual attraction? For the transgender population, should they include only those who have explicitly transitioned from one gender to another, or should they consider a broader group of people who express their gender in ways that do not easily conform to traditional notions of male and female?
The Ford Foundation recently funded a five-year study in which scholars considered these important questions. They concluded that concerns about non-response or survey termination are unfounded. Respondents decline to answer questions about their income much more often than they refuse to provide information about their sexual orientation or gender identity.
As to concerns about what we mean by LGBT, there is no precise definition. Instead, the scholars came to a consensus on a set of questions that measure different aspects of sexual orientation, including self-identification, sexual behavior and sexual attraction, and documented best practices for assessing gender identity.
Not every survey must ask questions about every dimension of sexual orientation and gender identity, but rather they should consider what dimensions make sense to measure, given the purpose of the survey.
A survey designed to measure workplace discrimination, for instance, may focus more on those who publicly identify as LGBT, since it can be difficult to make a case that you are being discriminated against because of your sexual orientation or gender identity unless you’ve been open about it. A survey that considers HIV risk factors may want to include questions about sexual behavior, since that is the most likely mechanism for becoming infected. A survey of younger people might want to focus more on sexual attraction, since many young people have not had any sexual experiences and are not yet comfortable with specific labels of their sexual or gender identity.
In fact, the number of people who self-identify as LGBT can and will differ from the number who say they have same-sex sexual experiences and the number who acknowledge same-sex attractions. Differences in these estimates do not mean we can’t measure the LGBT population with accuracy. Rather, they demonstrate that sexual orientation and gender identity are complex concepts worthy of further study.
The Institute of Medicine of the National Academies recently released a comprehensive analysis of the state of research on LGBT health and well-being. Recommendations from the report include LGBT inclusion across a wide range of federal surveys, coupled with increased support for federally funded LGBT research. This was welcome news. Federal statistical agencies can and should heed this call. Such data could provide the building blocks for critical research to understand the lives of LGBT Americans, who have too often been marginalized by researchers, as well as society at large.
I often hear LGBT advocates lament that it seems absurd that they don’t have equal rights in this country, given how large their community is.
As a demographer, I look at it a little differently. I’m amazed at how close we are to equality, given how small the community is.
Both perspectives are certainly valid, but both depend in part on an accurate assessment of the LGBT community — however you define it.
Assumptions about people are flimsy; numbers are solid. The reality of our political system is that you don’t really count unless you are counted. So it’s time to stop believing an old estimate and start making an accurate count.
Gary J. Gates is the Williams distinguished scholar at the Williams Institute at the UCLA School of Law and a co-author of “The Gay and Lesbian Atlas.”