Fig. 1

Proportion of survey participants who self-reported as lesbian, gay, bisexual or ‘other’. This figure shows the results of 22 National Social Surveys that included a question on sexual identity. For each survey, it provides the name, data collection period, proportion (filled square) and 95% confidence interval (horizontal brackets) of people who self-identified as lesbian, gay, bisexual or other. In addition, proportion of missing data as well as the overall survey sample size and response rate are shown. Missing data represent all people who responded ‘don’t know’, ‘prefer not to say’, refused or gave no answer to the sexual orientation question. For the Family Resources Survey, the LGB estimate is unweighted, because the weighted proportions were less precise than the unweighted proportions. For the Count Me In Survey; National Cancer Patient Experience Survey; and 1970 British Cohort Study, the LGB estimates are unweighted, because the surveys sampled the entire target population. For the EU Agency for Fundamental Rights—Violence Against Women Survey, the LGB estimate is for the response category of ‘non-heterosexual’. This survey made no differentiation between lesbian/gay, bisexual and other. The data is for the UK and could not be specified for England. For the British Social Attitudes Survey, the estimate is for the response categories ‘gay’, ‘bisexual’ and ‘can’t choose’. This survey had no category for ‘other’. For the Count Me In Survey, the proportion of ‘no answers’ was very high, while the survey response rate could not be retrieved. It is therefore likely that at least a proportion of people with ‘no answer’ were in fact not eligible to respond or never asked the question. For the First Longitudinal Study of Young People in England survey, we were not able to obtain original data