How many LGBT folk are there?
April 8th, 2011
Some time ago I set out to discover to just what extent are gay and bisexual people infected with the Human Immunodeficiency Virus (HIV). But to do that, I had to figure out just how many people are gay or bisexual. But, as I noted at the time, this is not an easy demographic to define:
To know how many gay people there are, one has to explain what one means by “gay person”. And there are several ways to approach this.
Just a few decades ago, self identification would be a useless parameter. Two men could have lived together for decades and been known to friends and family as a devoted couple without either being willing to be publicly identified as “gay” or “homosexual”. Even today, epidemiologists are careful to use terms such as “men who have sex with men (MSM)” so as to include for health purposes those who do not use LGBT identifications.
But sexual behavior is also not a good indicator. Some gay or bisexual persons may identify publicly as gay and yet for personal reasons choose not to be sexually active. A self-identified gay Christian, for example, may elect celibacy or may defer sexual activity until united in the bonds of a committed relationship but they are no less “gay” than the man with a life partner or the woman living for her next sexual conquest.
Perhaps the best definition would be those who are exclusively or primarily attracted to the same sex. But this definition might also include ex-gays and others who would object to being so identified. And for purposes of discussions of “gay community”, it’s hardly fair to include those who have no communion with other gay persons.
Others, thinking in terms of community, might include as “queer” all persons who do not identify at heterosexual. But to my way of thinking this is far too broad for our purposes, including asexual persons, anti-sexual persons, and those who choose to avoid labels, regardless of sexual attractions.
But, relying on what was, at that time, the best info available: an abstract of the 2002 National Survey of Family Growth put out by the Centers for Disease Control, I came up with an answer. And as the Survey allowed me to look at more than one possible definition – and as I got to the same result either way – I could state with a measure of confidence the following:
I think it fair to state that at least 2.3% of men are gay and at least 4.1% are “gay or bisexual”.
I think it fair to state that at least 1.4% of women are gay and at least 4.1% are “gay or bisexual”.
Although this is based on assumptions, by applying the above percentages it is reasonable to state that there are at least 5.3 million gay or bisexual men and at least 5.5 million gay or bisexual women above the age of 15 living in the United States, for a total GLB Community of at least 10.8 million people.
I have generally stuck with my limited analysis waiting for additional or better information. But now Gary Gate of UCLA’s Williams Institute has undertaken the task of assessing the available data and coming up with an answer to that elusive question.
Gates is perhaps the foremost demographer studying the LGBT community. The information that has now become recognized about where gay families are (virtually everywhere) and what census data could tell us about them came primarily from Gates and the Williams Institute. When his Gay and Lesbian Atlas came out in 2004, it tossed over the apple cart of presumptions about same-sex families existing only in the cities.
So Gates’ findings have been given a level of credibility that has not been questioned. He is a specialist talking about his specialty. And, to my initial smug pleasure, Gates arrived at estimations that are not far from my own.
- An estimated 3.5% of adults in the United States identify as lesbian, gay, or bisexual and an estimated 0.3% of adults are transgender.
- This implies that there are approximately 9 million LGBT Americans, a figure roughly equivalent to the population of New Jersey.
- Among adults who identify as LGB, bisexuals comprise a slight majority (1.8% compared to 1.7% who identify as lesbian or gay).
But while I am impressed by his scholarly contributions and defer to his understanding of demographics, I am shocked at some of his premises, methodologies and conclusions. I think that as he further continues his work that he should challenge some presumptions both in analysis and in presentation.
Perhaps the weakest possible method for reconciling variances between study results would be to average them. Not all studies have equal likelihood of accuracy, equal methodology, equal statistical significance, or equal credibility. Outliers ought not have the same input, and a survey of ten people does not carry the same weight as a survey of ten thousand.
Additionally, the purpose of a survey weighs both on questions asked and the conclusions drawn. And let’s not pretend that a survey of Californians says much about Americans as a whole. Which leads me to question the inclusion of both the National Epidemiological Survey on Alcohol and Related Conditions and the California Health Interview Survey.
So I am troubled by a methodology described as:
However, combining information from the population-based surveys considered in this brief offers a mechanism to produce credible estimates for the size of the LGBT community. Specifically, estimates for sexual orientation identity will be derived by averaging results from the five US surveys identified in Figure 1.
I cannot tell from Gate’s description of his methodology whether he gave weight based on sample size or took other steps to smooth the results. Nor is it clear what sort of margin of error can be relied upon in Gate’s calculations.
But considering that we are talking about small numbers to begin with, I am concerned that “averaging results” can lead to results that are statistically meaningless.
Adopting the lowest possible estimate
I recognize and appreciate the value of estimating conservatively. This is the hallmark of a good demographer and demonstrates professionalism over advocacy. However, when dealing with a population which you know for certain is underestimated, it is not accurate to set the lowest estimate as though it were the the truest estimate.
And even if one is estimating those gay and lesbian people who so identify (as opposed to other measures), it is a given that public surveys under-count your population. There are virtually no people who will identify on a survey as being gay when they are not. But there are a not-insignificant number who do identify as gay – to themselves, their friends, their family – but who will not disclose their orientation in a survey.
So language that says “are” rather than “at least” falsely implies that the lowest number is accurate.
Male and Female Sexuality
Gates does note that the gay/bi breakout is not mirrored between gay men and lesbians. But I don’t think that he adequately emphasizes, or perhaps even recognizes, the extent and complexity of this difference.
Based on my experiences writing here, reviewing the literature, and discussing the issue with others who approach sexuality from different perspectives, I’ve come to conclude that male and female sexuality is very different. Men and women, and especially gay men and women, experience attraction differently, respond to it differently, and have differing incidences of fluidity in sexual desire.
So it is not only a simplistic presentation to ‘average’ the results of male and female sexual identity, but it presents a false picture of both. So while “Women are substantially more likely than men to identify as bisexual” is informative, the following is a useless statistic:
Among adults who identify as LGB, bisexuals comprise a slight majority (1.8% compared to 1.7% who identify as lesbian or gay).
A better set of bullet points – less likely to be distorted by dishonest or lazy media – would be to report what the demographics tell us about each group (to the extent that they do). For example, Gates would have been both more informative and more accurate to say:
- At least 3.5% of adults in the United States, around 9 million people, identify as lesbian, gay, or bisexual.
- Of men, at least 2.2% identify as gay and an additional 1.4% as bisexual. Of women, at least 1.1% identify as lesbian and an additional 2.2% as bisexual.
And even that is not fully informative. Those of us in the community are well aware that identifying as bisexual rather than gay is often as much a matter of caution as it is a reflection of internal perception.
For some reason Gates found it necessary to discuss difference between the sexual identity, sexual behavior, and sexual attraction of Americans by looking to surveys in Canada, the UK, and Australia. I think it fairly obvious that distinctions between attraction and identity are driven to a large part by culture. And it is also fairly obvious that Australian attitudes do not well explain the identity of Americans.
While his estimates of gay men and gay women were sloppy, his estimates of transgender Americans are, at best, fanciful.
Gates achieves his estimate by averaging two surveys, one of which was at 0.1% and one of which was at 0.5%. This is nonsense. Frankly there isn’t enough data to make any credible estimate, and Gates should know better.
When you are averaging two numbers, one of which is five times the other, and both of which are pretty much in your margin of error, anything you come up with is little more than a guess.
Gates should simply have noted that there is an additional demographic, likely less than one half of one percent of the population, of persons who identify as being transgender and left it at that. Even “likely less than one half of one percent” is aggressive, considering the scarcity of raw data.
To put a numerical value (0.3%, or 700,000) is to distort reality and misapply his profession. This does a service to no one.
I was initially pleased by Gates’ efforts. Finally someone credible was attempting to answer that oft-guessed-at but seldom analyzed question as to just how big the LGBT community might be. And I also felt validated that his answers so closely mirrored the conclusions of my limited efforts. Points for me!
But upon closer inspection, this report is notable as much by its limitations and failings as it is by its effort to provide an answer.
I don’t disagree with Gates’ conclusions, necessarily. They are in the ballpark of reasonable, however shoddily he got there. And they serve a purpose in that they can – I hope finally – put an end to Kinsey’s 10% figures.
But this report should be treated as nothing more than a stepping stone. An interim effort on which future work can build. And it is on this that Gary Gates and I agree:
Understanding the size of the LGBT population is a critical first step to informing a host of public policy and research topics. The surveys highlighted in this report demonstrate the viability of sexual orientation and gender identity questions on large national population-based surveys. Adding these questions to more national, state, and local data sources is critical to developing research that enables a better understanding of the understudied LGBT community.