Some people are gay, should the census get over it?
I love the census. There is something about a survey of everyone in a whole country that is very exciting to a research geek like me. It is such a simple idea, but the sheer scale of it means that it is so very very complicated to pull off. I am quite in awe of any organisation that takes it on. And I’m also excited about getting the results to the recent UK census. The last UK census was taken in 2001 and I started my research career in 2002, so over the last nine years the available census information has been getting less and less relevant. So I’m glad the new one is here! As a researcher it is so helpful to have a complete UK profile because this can be used for context in all sorts of research reports, and forms the basis of many a sampling frame used to ensure our surveys are representative.
When I came to complete my census I was surprised to see no question on sexual orientation in the survey, as I know there is considered to be lack of accurate figures about sexual orientation in the UK. I also know that this has a real impact on the priority and direction of policy and practice relating to the lesbian, gay and bisexual population.
“A question on sexual orientation would help to monitor equality legislation and improve service provision to lesbian, gay and bisexual people” (Stephen Williams MP, in a motion in the House of Commons in 2007, asking for a question about sexual orientation to be included in the 2011 UK census).
If you want to know the proportion of lesbian, gay and bisexual people in the UK, putting a question in the census seems an obvious place to start.
However, as long ago as 2006, the Office of National Statistics were putting out papers explaining that they did not intend to include a question about sexual orientation in the 2011 UK census. They said that:
- Sexuality is multi-faceted and difficult to define in a questionnaire format
- Quality and accuracy of responses is an issue, as people may not respond honestly
- Acceptability and impact on response rates is an issue where including a sexual orientation question might put people off completing the census as a whole
This was backed up by information from the 2006 Statistics Canada Census content consultation which found that “Most participants did not approve of including a sexual orientation question on the Census”.
Since then, there has been a fair amount of lobbying from organisations such as Stonewall who believe it is important to include questions about sexual orientation in the UK census.
But it hasn’t happened.
I have been mulling this issue over since I completed my own census.
For those who don’t know, each household gets one census form to be filled in for everyone who lives there. This was the first time I had completed a census as a ‘household’, as in 1981 and 1991 I was a child and in 2001 I was in a shared student flat.
I had not really been aware of the methodological issues around responding as a ‘family’ before. In practice you either end up with one person answering on behalf of all other members of the household, or all members filling in their own section but the whole household having access to the information.
This isn’t methodologically ideal. If only one person guesses the answers on behalf of their family they might get the answers wrong (through human error or not knowing the correct answer) and therefore the data collected will be incorrect. If everyone gets to see the answers for the whole family, individuals might be unwilling to respond truthfully about certain sensitive questions.
With this in mind, I can see the point that the Office of National Statistics are making about sexual orientation. Given its methodological format, is unlikely the UK census would elicit accurate statistics about sexual orientation as correct responses will not be gathered from those who are not ‘out’ to the people they live with. The data collected is unlikely to be complete, which means it is unlikely to be useful. It is pointless and inethical to collect useless data.
And lets be human too, we need to consider the potential negative wellbeing impact of including the sexual orientation question in a survey of this format. If a member of a household is not out to the people they live with, the inclusion of the question could out people when they are not ready, or cause stress for someone worrying about whether they will be outed when they are not ready.
I can see why they didn’t put it in.
However, even if the data collected is not useful (just now) we will never start collecting useful data on sexual orientation until people start accepting questions about sexuality as a standard. This can’t be done unless the questions start going into high profile surveys (such as the census) and we all get used to seeing them and answering them.
It looks like the Office of National Statistics will be trying alternative ways to gather information on sexual orientation and I can see that the Scottish Household survey (for example) would be a good place to do this. It is done face-to-face and one-to-one, and questions can be completed confidentially (not read out by the interviewer) if required. This encourages more honest response and has less negative impact on the wellbeing of the respondent. I’m a great believer in sampling. I know you don’t need to speak to everyone to get an accurate result using an appropriate methodology. It isn’t a census though.
It would have been great for gay rights if the Office of National Statistics had led the way with this issue by doing everything they could to get it right in their flagship survey. They would have needed to amend the methodology a bit (perhaps allowing a privacy setting for online responses), and/or accompany the survey with pro-active census-related support offered to the lesbian, gay and bisexual community via the third sector. Slightly less convenient, but so what? It would have been the right thing to do in an accepting society rather than sweeping it under the carpet for another ten years.