In the closet or not?
A new figure about the number of gay people in the UK has been greeted with scepticism by some who say it doesn't add up, so what is the problem?
Are you gay? Very probably not, if a recent figure from the Office for National Statistics (ONS) is to be believed. It says gay, lesbian and bisexual people make up just 1.5% of the British population.
Such a small percentage made it a big story, with the figure making headlines in many national newspapers. Some were surprised and questioned it. Gay groups in particular greeted it with scepticism.
The figure, say critics, is far too low. After all, when civil partnerships were introduced the government said homosexuals made up between 5% and 7% of the UK population.
Another much-used statistic - this one attributed to Alfred Kinsey, the father of sexual behaviour analysis - puts the figure at 10%.
So why the discrepancy? Some critics point a finger at the methodology used by the ONS. In several newspapers it was reported that the survey had been conducted on the doorstep, with the ONS turning up without prior notice. Critics argue people are reluctant to talk about their sexuality with a stranger on the doorstep.
But the ONS says this wasn't the case. The 450,000 people involved were chosen through random sampling and were then sent a letter asking if they wanted to take part. If they did a date was set for someone to visit or interview them over the phone.
Each member of the household was interviewed separately using question cards and asked a series of questions on a whole range of issues. The ONS had informed them about the nature of the questions beforehand. Not everyone was asked about sexual preference, says the ONS. Some participants were too young, for example. In total 247,623 people were asked and 238,206 gave a valid response.
The question about sexual perception and the way it was asked was, of course, very important. The ONS worked for three years in an effort to get it right. It asked: "Which of the options on this card best describes how you think of yourself?" The options were heterosexual, straight, gay, lesbian, bisexual and other.
The way people were able to respond was also important and explained to all participants in advance. Each person was asked the same question, but shown a different question card with a unique number next to each preference. It meant anyone trying to overhear the answer could not tell which category the respondent had selected.
The ONS says this method reduced the number of refusals and "don't knows". The results show nearly 20% of those who say they are gay, lesbian or bisexual are married and currently living with someone of the opposite sex.
"We're quite confident that the estimate we have is an accurate estimate of self-perceived sexual identity," says Stephen Hicks from ONS.
So why was last government estimate put at 5% to 7%? The ONS says one reason for the difference is that the latest figure is for sexual perception, not behaviour - two different things.
"Someone may engage in sexual behaviour with someone of the same sex but still not perceive themselves as gay," says a ONS spokeswoman.
It's also hard to judge the reliability of the previous figure, says Mr Hicks. This is because they are based on a number of studies that asked different questions, used different methods and different samples in different countries. It's a weakness the original report acknowledged.
The other widely-quoted study came from the infamous Professor Alfred Kinsey. The 10% figure is derived from his two seminal works: Sexual Behaviour in the Human Male and Sexual Behaviour in the Human Female.
He interviewed several thousand men and women in the United States between 1938 and 1952 and found that 13% of the men and 7% of the women were "almost exclusively" homosexual.
But Kinsey's methods have also been questioned. Professor Julia Ericksen from Temple University in the US has studied his work.
"He didn't believe in survey sampling and he was quite antagonistic towards it," she says.
Instead he collected samples in order to record as much variety in sexual behaviour as possible. His interviewees were mainly married college students and members of social clubs.
When he felt he needed to meet more working class people he went looking for them in prisons, believing, says Professor Ericksen, that "most working class people spent some time in prison."
He also used to hang around outside public toilets at Penn Station in New York, a place where gay men were known to meet. He would intercept men who remained in the toilet for an unusually long time and persuade them to be interviewed.
Professor Ericksen says his sample technique cannot be used to give an indication of the behaviour of the population as a whole.
So does this mean the ONS figure is a better estimate of the lesbian, gay and bisexual population than previous attempts? Possibly. If you compare the ONS work with studies that asked similar questions, its findings are not unusual, says Professor Ericksen.
But, the question itself might be too narrow, she also argues. The ONS asked about sexual identity. Researchers say there are two other elements to sexual orientation: sexual behaviour and sexual attraction.
Professor Ericksen says studies in the US have shown that when you ask about all three in the same survey, sexual identity gets the lowest response. For example, there are men who have had, or have desired, sex with other men, who do not consider themselves to be gay, she says.
"If you really want to get a good grasp at same gender behaviour, you really have to ask all those questions and they didn't do that," she says. "It's not surprising that they got a low figure."