Dear Mr Humpherson
The ONS has released the final report on its investigation of the quality of the 2021 census’s gender-identity data together with a statement saying it has “confidence in the gender identity estimates at national level”. The report offers the following conclusions in support of this.
- On collecting and processing the data: “This strand of our research … provided no evidence that the design of the question or the statistical processing of the collected data had an adverse effect on the quality of the published statistics”.
- On respondent error and misunderstanding: “There are patterns in the data that are consistent with some respondents not interpreting the question as we had intended… However, for the reasons outlined in this report, we cannot say whether the census estimates are likely to be an overestimate or an underestimate of the true value, given other sources of uncertainty, not least the potential impact of question non-response. Therefore, the overall impact on the data of any misinterpretation of the question cannot be determined.”
We find that the ONS’s investigation has been inadequate and its conclusions are not supported for five reasons.
- The investigation does not address the core issue concerning consistency or not between an individual’s answers to the sex question and gender-identity question.
- Implausible explanations are accepted ahead of more likely interpretations of misunderstanding for answering NO to the question about whether they match.
- Given the small numbers of transgender people and the likely level of misunderstanding the statement that the ONS has “confidence in the gender identity estimates at national level” is not supported.
- The classification of people by the terms “trans man” and “trans woman” by the ONS is not in line with the Census (England and Wales) Order 2020.
- The design of the question relies on the whole population making a declaration based on the idea of gender identity in order to estimate a minority that identify as transgender.
Furthermore the focus in the ONS statement on the numbers being plausible at the national level implies (but does not make explicit) that it is not confident they are plausible at the subnational level.
Attached to this letter are further details on each of these concerns.
Our conclusion is that data on gender identity (including the sex of the people identified as transgender) is not fit for purpose. It was driven by the adoption of concepts and questions promoted by lobby groups that seek to replace sex with gender identity.
The lack of clarity and certainty about the meaning and interpretation of the sex question and the gender identity question also has knock-on effects for interpreting how the sexual orientation question, the sex and the gender identity question interact.
This investigation by the ONS is inadequate and undermines confidence in national statistics. It lays the groundwork for further erosion of clarity on sex, and the wider adoption of a gender identity question, and use of associated data, that has been demonstrated to be unreliable.
The ONS has proved itself unwilling to accept clear indications that the gender-identity question produced unreliable answers, and has adopted a “self-identified” approach to the definition of sex even after the High Court judgment in 2021. We therefore have no confidence that the ONS will adequately address the need for clarity and accuracy about sex and transgender identity in its development of the harmonised questions on sex and gender identity.
We call on the Office for Statistics Regulation to take regulatory action in order to secure public confidence in national statistics, and to prevent the faulty question being replicated.
- The question should be officially retired, with an apology and an explanation to discourage others from using the same wording.
- A warning should be put on the data and it should not be designated as national statistics. The OSR must determine whether the national headline figure and the sub-national figures should have national statistics designation.
- The ONS definition of “sex” for the 2021 census should be corrected to reflect the guidance given to respondents following the FPFW challenge, and the question of whether actual sex should be routinely and clearly collected for population demographics should be reviewed.
The ONS should also:
- publish the free text answers broken down by sex
- publish details on the number of people in the telephone survey who did not confirm their answers to the gender-identity question
- analyse the pattern of households which they classify as having more than one trans member (which may indicate a misunderstanding of the question)
- investigate whether it is possible to cross-check sex and self-identified gender for this section of the population with other data sources such as administrative data from the NHS and the birth and gender reassignment registers.
We would like to know whether the response from the ONS was seen or agreed by the OSR before the ONS sent it and whether it has been discussed by the UKSA board. Is the ONS report the view of the National Statistician? Has it been seen by the Methodological Assurance Review Panel or any other methodological group?
We are publishing this letter and we request that it and any response are also published on the UKSA website and linked to in a way that makes it findable to users of the gender-identity data.
Yours sincerely
Maya Forstater
Executive Director
Helen Joyce
Director of Advocacy
Michael Biggs
Advisory group member