In April we submitted Freedom of Information requests to the 3 census authorities (ONS, NRS and NISRA), asking what analysis they had undertaken to estimate the impact of framing the sex q. in terms of self-declared gender identity on data reliability at the sub-population level.

The response from ONS and NRS indicated that neither had undertaken any such analysis. We have now received a response from NISRA, which likewise indicates that the impact on data reliability at the subpopulation level has not been examined. https://www.whatdotheyknow.com/request/660666/response/1604226/attach/html/2/NISRA%20FOI%20DoF%202020%200185%20Response.DOCX.docx.html

As this blog details, there may be significant differences in how people respond to the sex q. across different groups: in a Swedish population study, 6.3% of 22-29 yr olds wanted to be wanted to be treated as the opposite sex, compared to a 2.8% average. https://murrayblackburnmackenzie.org/2020/06/04/a-negligible-effect-the-untested-assumptions-behind-plans-for-the-2021-census/

It appears however, that census planning across the UK has proceeded based on the untested assumptions that the trans population is small, and that a sex question framed in terms of self-declared gender identity will not affect data reliability across different groups.