https://lgballiance.org.uk/facts/
Christ, not the 6% thing again - they were being absolutely honest about what they could actually say from stats collected (yes, I know, revolutionary):
We used Mailchimp to send our newsletter and when we set up our account in 2019 we added some subscriber questions which, as it turned out, provided us with ambiguous data.
We asked people whether they were lesbian, whether they were lesbian/gay or if they preferred not to say. The flaws being that we couldn’t tell whether those who ticked lesbian/gay were men or women and that none of the fields were compulsory – so many people skipped them altogether.
The result was that we had 4,502 newsletter subscribers and 316 ticked the box describing themselves as lesbian. That’s 7% of the total. A further 949 ticked the box lesbian/gay and 1,427 were unspecified or preferred not to say. Based on that data that means that between 316 (7%) and 2,376 (53%) of our subscribers were lesbian.
The 7% figure was used in court because it’s important that evidence is based on provable fact and it is a fact that, at a minimum, 7% of our subscribers were lesbians. However, common sense told us that that number was really much higher.
So loads of people grasped onto that, because they're used to people lying about stats.
In August 2022 we commissioned a survey of our subscribers to help us plan to deliver services and support to LGB people. One of the questions we asked was about sexual orientation. That data showed that 34% are lesbian, 33% are gay men, 12% are bisexual, 20% are heterosexual and 1% preferred not to say. We are satisfied that this data is robust.
So there we go, when they did a better survey, they discovered 79% of their subscribers are LGB.