Census 2021 gender identity estimates for England and Wales, additional guidance on uncertainty and appropriate use

• There are patterns in the gender identity estimates from Census 2021 that suggest some respondents may not have interpreted the gender identity question as intended, most notably those with lower levels of English language proficiency.

• Some people may have unintentionally given an answer that they had a gender identity different from their sex registered at birth (we refer to people in this category as trans in this article); this has increased uncertainty in the gender identity estimates, compared with other Census 2021 estimates.

• Gender identity estimates from Census 2021 should not be used as precise estimates to support service delivery, but can be used to provide insights; users should refer to the specific advice in this article before using the Census 2021 gender identity estimates to come to conclusions about the trans population.

• The Census 2021 estimate of the trans population of England and Wales aged 16 years and over (0.54%) gives an indication of the size of the population; this is broadly consistent with other sources, but caution is needed for estimates of gender identity subcategories.

• Our analysis in this article concludes that Census 2021 gender identity estimates broken down by age, sexual orientation, housing, health, disability, and unpaid care can be used to provide insights on the relationship between gender identity and these variables, but as a result of patterns seen in our data, we are less confident in the reliability of breakdowns by other variables, partly because of how these variables correlate with English language proficiency.

Census 2021 gender identity estimates can provide a broad indication of the size of the trans population in England and Wales and insights into some characteristics. This article provides more information on appropriate use of the estimates, drawing on new and previously published Office for National Statistics (ONS) research. This article includes comparisons with estimates for Scotland published on 26 March 2025 in their Flexible table builder.

The proportion of those aged 16 years and over who said their gender identity was different from their sex registered at birth (referred to as trans in this article) was 0.54% in England and Wales. This is around 1 in 200 people. This is broadly consistent with the best comparable sources within the UK and internationally, including Scotland (0.44%), as described in our ONS letter to the Office for Statistics Regulation (OSR) on Census 2021 gender identity estimates. Some differences in estimates are expected, reflecting different questions asked of different populations at different points in time.

However, patterns in our data suggest that some respondents may not have interpreted the Census 2021 gender identity question as intended. This means they may have unintentionally responded as trans, most notably those with lower levels of English language proficiency.

There is uncertainty related to all census statistics including, for example, people not responding, which stands at 6.00% for this voluntary question. However, the potential misinterpretation of the gender identity question adds extra uncertainty. This is particularly important here because the group we are interested in (the trans population) is small. This means that a very small percentage of inaccurate responses can have a relatively large impact on the estimates of trans people.

To support interpretation and use of the Census 2021 gender identity estimates, we have split them into three categories of confidence: high confidence, confidence with caveats, and limited confidence.

Category 1: High confidence

Census 2021 gender identity estimates can be used to provide:

• a broad indication of the overall size of the population in England and Wales who had a gender identity different from their sex registered at birth (trans)

• insights into the characteristics of the trans population in terms of age and sexual orientation

• insights into housing, health, disability, and the provision of unpaid care among trans people, compared with those with a gender identity the same as their sex registered at birth

Category 2: Confidence with caveats

The Census 2021 gender identity estimates in this category can be used to provide broad insights, but they have a high level of uncertainty, likely partly related to English language proficiency levels. Users, therefore, may want to consider excluding people whose main language is not English (English or Welsh in Wales) from their analysis of these topics (please see our accompanying dataset). They should also refer to the specific advice in this article before using these estimates. Estimates in this category include:

• geographical comparisons between local authorities, regions, and at country level for England and Wales

• country of birth

• ethnic group

• religion

• economic activity status

Category 3: Limited confidence

The Census 2021 gender identity estimates should not be used to provide:

• estimates of the sizes of the population who identified as trans man, trans woman, non-binary, and all other gender identities

• insights into the relationship between being trans and having lower levels of English language proficiency

• insights into the qualification levels of trans people, compared with people who have a gender identity the same as their sex registered at birth (see Section 9: Education and employment for more detail)

Previous research on the quality of Census 2021 gender identity data

Census 2021 was the first census in England and Wales to include a question on gender identity. The question was asked as:

You can also view how this question looked on the Census 2021 paper form.

In this article we refer to people who said their gender identity was different from their sex registered at birth as trans (see Section 12: Glossary for definitions of terms). More information on the question development is available in our Sex and gender identity question development for Census 2021 report.

We published two releases of census statistics on gender identity in January 2023: our Gender identity, England and Wales: Census 2021 bulletin and our Gender identity: age and sex, England and Wales: Census 2021 article. We released further data tables in April 2023, and Scotland’s Census published their Sexual orientation and trans status or history report in June 2024.

Following these publications, some data users raised concerns about potential misunderstanding of the question and the effect of this on the quality of the data. In particular, the England and Wales Census 2021 estimates showed that respondents with lower levels of English language proficiency were more likely than the general population to say that their gender identity was different to their sex registered at birth, some of whom may have unintentionally given an answer suggesting that they were trans.

We undertook a programme of work to investigate the quality of the data. We published the findings in two articles. The first was our Collecting and processing data on gender identity methodology, published in June 2023. The second was our Quality of Census 2021 gender identity data article, published in November 2023. Additionally, the Office for Statistics Regulation (OSR) launched a review of statistics on gender identity based on data collected as part of Census 2021 for England and Wales in April 2023. An interim report on OSR’s findings was published in October 2023 and the final report was published in September 2024.

In September 2024, Emma Rourke, Deputy National Statistician at ONS, wrote to OSR to request a change to the designation of Census 2021 gender identity statistics from accredited official statistics to official statistics in development. This change in designation better reflects the innovative nature of the estimates, the limitations arising from potential question misinterpretation, and the evolving understanding of measuring gender identity. We also committed to carrying out further work to provide more detailed information on the uncertainty associated with the estimates and guidance on their appropriate use. This article presents the findings from this work.

New research on the quality of Census 2021 gender identity estimates

The findings in this article are based on four strands of research on:

• exploring uncertainty in the data by splitting respondents who reported a gender identity different to their sex registered at birth into those that we can be more confident interpreted the question as intended (Type A), and those we are less confident about (Type B)
• applying a range of hypothetical error rates to Type B respondents, ranging from 0% error to 100% error, to explore whether the patterns in the data by geography, characteristics, and other variables hold in different hypothetical scenarios (for more detail on this, see Section 13: Data sources and quality)
• exploring whether the patterns in the data for the whole population are mirrored in the estimates for those with English (English or Welsh in Wales) as their main language
• comparisons with estimates from Scotland Census 2022, which were not available when we undertook the previous work on quality

It is not possible to use this research to provide an alternative set of estimates. The aim of this article is to provide guidance on the validity of different conclusions that could be drawn from the Census 2021 estimates, and advice on using the data for different purposes.

Splitting respondents into Type A and Type B

Respondents who ticked “No”, indicating that they had a gender identity different from their sex registered at birth, were categorised as Type A or Type B based on the write-in provided for the gender identity question and their response to the sex question. Type A respondents are those who:

• were coded as a trans man or a trans woman and provided a response to the gender identity write-in that was different to their response to the sex question, for example, sex “female” and gender identity write-in “man”

and/or 

• provided a gender identity write-in that clearly indicated that they had a gender identity different from their sex registered at birth, for example, “non-binary”, “trans man” or “gender fluid” 

Type B respondents are those who: 

• did not provide a write-in (Type B1)

or

• were coded as a trans man or a trans woman but provided a write-in that was consistent with their response to the sex question and not on the list used for Type A, for example, sex “male”, gender identity write-in “man” (Type B2) 

A response being categorised as Type B indicates higher uncertainty. It is expected that some of these responses will truly reflect someone identifying as trans. However, we cannot be confident that all responses will be someone identifying as trans.

Comparisons with estimates for Scotland

The 2022 Census in Scotland included a voluntary question for those aged 16 years and over on trans status for the first time. However, the wording of the question was different to the question used to collect data on gender identity in England and Wales.

The question used in Scotland was:

More information on the question development for Scotland is available in Scotland’s 2022 Census Sex and gender identity topic report.

In addition to using different questions, the questionnaire for Scotland included a definition of the term trans. The questionnaire for England and Wales did not include a definition of gender identity. When asking people to provide a write-in, the questionnaire in Scotland said: “please describe your trans status (for example, non-binary, trans man, trans woman)”. The questionnaire in England and Wales said: “write in gender identity” and did not include any examples. The trans status question in Scotland was asked immediately after the sex question, while the gender identity question in England and Wales was asked later in the census form. The estimates are from censuses taken of different populations at different points in time.

Although there were differences in the questions asked, the Scotland estimates provide a useful comparator for confirming how confident we can be about findings from the data for England and Wales.

The Census 2021 estimate is that 0.54% of people in England and Wales aged 16 years or over had a gender identity different from their sex registered at birth. The majority (93.46%) of people responded that they had a gender identity the same as their sex registered at birth and 6.00% of people did not answer the question. The 2022 Census in Scotland found that 0.44% of people aged 16 years or over were trans or had a trans history.

Just over two-thirds (68%) of trans respondents in England and Wales were categorised as Type B. This means that there is a higher level of uncertainty in the Census 2021 estimate, in addition to the usual uncertainties that come with survey data on a sensitive topic. This means that the Census 2021 estimate should not be used as a precise estimate. It can still be used to give an indication of the size of the trans population in England and Wales, given that it is broadly consistent with other sources. This was confirmed by the Office for Statistics Regulation (OSR) in their final report, where they concluded that “an estimate of around 1 in 200 triangulates with other sources and is not likely to be materially misleading”.

The Census 2021 estimates provide a breakdown of the trans population into subcategories, based on the write-ins that respondents provided.

In England and Wales, almost half (45%) of people who identified as trans did not provide a write-in response (Type B1). Therefore, the gender identity estimates by trans subcategories can be used to provide insights on how people chose to respond to the gender question. However, they are unlikely to accurately reflect the true populations who identified as trans woman, trans man, non-binary, and all other gender identities.

In Scotland, the percentage of people who identified as trans but did not specify their trans status was 16%. This may be related to the fact that Scotland presented response examples.

The rotated bar chart in Figure 1 shows that while 0.06% of respondents to the census (aged 16 years and over) identified as non-binary in England and Wales, 0.20% of respondents did so in Scotland. This suggests that the Census 2021 estimate for non-binary people for England and Wales could be an underestimate of the true proportion of the population who identified as non-binary. It also likely reflects that non-binary was included as an example of a trans status in Scotland.

The census estimates for trans men and trans women are higher for England and Wales than Scotland (0.10%, compared with 0.07% for both trans men and trans women, respectively). However, there is a high level of uncertainty in the estimates for trans men and trans women in England and Wales. Almost two-thirds (63%) of these respondents have been categorised as Type B2 because they provided a write-in that was not clearly trans, such as “male”, and that was consistent with their response to the sex question. The Census 2021 gender identity estimates should not be used to provide estimates of the sizes of the population who identified as each of trans man, trans woman, non-binary, and all other gender identities because of this uncertainty.

We linked a subsection of Census 2021 data to birth registrations data, as approved by the National Statistician’s Data Ethics Advisory Committee. We used this linked data to investigate whether some respondents recorded that their gender identity was not the same as their sex registered at birth, but then provided a write-in that was consistent with the sex recorded on their birth registration. The analysis confirmed some respondents did this and were incorrectly recorded as trans. However, because of limitations with the data and linkage, we could not draw conclusions about the scale or effect of this.

In Scotland, the published data on write-in responses show that the majority of people coded as trans men and trans women provided a clear trans write-in. This means that there is a lower level of uncertainty, compared with the estimates for England and Wales. As with non-binary, trans man and trans woman were included as examples in the write-in prompt on the Scottish questionnaire.

Country

The Census 2021 estimates show that a higher proportion of people in England had a gender identity different from their sex registered at birth, compared with Wales (0.55%, compared with 0.40%), as shown in our Gender identity, England and Wales: Census 2021 bulletin.

When considering estimates for those with English (English or Welsh in Wales) as their main language, and after applying hypothetical error rates, the differences between the England and Wales estimates are reduced. We conclude that the proportion of people with a gender identity different from their sex registered at birth in England is broadly similar to Wales.

Region

The rotated bar chart in Figure 2 shows that within England, the London region had the highest proportion of people who reported their gender identity was different from their sex registered at birth (0.91%).

Considering the percentage of people who identified as trans who also stated their main language was English, the pattern across the regions is the same. London still had a higher proportion of people who identified as trans, compared with the other regions of England.

This means that despite London having a higher proportion of trans respondents categorised as Type B, the data supports the general conclusion that London has a higher proportion of people who identified as trans when compared with the other English regions.

Local authority

Looking at the Census 2021 estimates for local authorities, of the 10 local authorities with the largest proportions of people who reported a gender identity different from their sex registered at birth (trans), 8 were in London, with Newham (1.51%) and Brent (1.31%) having the highest proportions. In Scotland, it was found that the four “city” council areas (Dundee City, City of Edinburgh, Glasgow City, and Aberdeen City) had the largest proportions of people reporting as trans.

There is a higher level of uncertainty in London local authority estimates. For example, over 80% of trans respondents in Newham and Brent (and other London local authorities) were categorised as Type B. This is compared with lower levels of uncertainty in Oxford and Norwich, which were also in the top 10 local authorities.

Taking this uncertainty into account, we cannot say with confidence which local authorities have the highest proportions of the population who are trans. However, we can say more broadly that local authorities that are more urban, have large student populations, or have a relatively high proportion of their population aged 16 to 34 years are likely to have a higher proportion of their population who are trans.

Age

The Census 2021 estimates show that younger people were more likely to have a gender identity different to their sex registered at birth, compared with older people. For those aged 16 to 24 years, 1.00% reported being trans. This decreased with age to 0.22% for people aged 75 years and over. This finding holds following our research and was mirrored in Scotland, so we are confident that the Census 2021 estimates can be used to provide broad insights on how gender identity varies by age.

The accompanying dataset shows how the proportion of Type B responses increases as age increases. The estimates should not be used as precise estimates of the size of the trans population in each age group, because of uncertainty in the data.

English language proficiency

The Census 2021 estimates showed that respondents with a main language other than English, particularly those with lower levels of English language proficiency, were more likely than the general population to say that their gender identity was different to their sex registered at birth.

However, there is a high level of uncertainty in the gender identity estimates for people with a main language other than English, as shown in Table 2. Because of this high level of uncertainty, and the relatively large differences in the estimates when compared with those for Scotland, we advise against using the Census 2021 estimates to provide insights about people with lower levels of English language proficiency being more likely to be trans. We think the estimates reflect greater misinterpretation of the question by people whose main language is not English (English or Welsh in Wales).

Country of birth

In the Census 2021 estimates, a higher proportion of people born outside the UK reported being trans, compared with people born in the UK (1.29%, compared with 0.36%). This was mirrored in Scotland where 0.71% of people born outside the UK reported being trans, compared with 0.41% for those born in the UK.

Looking at the Census 2021 estimates in more detail, those born outside the UK who reported being trans had a higher proportion categorised at Type B, compared with those born in the UK (86%, compared with 53%). This is potentially linked to differences in English language proficiency (please see our accompanying dataset). This higher degree of uncertainty means we cannot be fully confident that a higher proportion of those born outside the UK are trans, compared with those born in the UK. It is likely that any difference between groups is not as large as shown in the Census 2021 estimates.

Ethnic group

The rotated bar chart in Figure 3 shows that the proportion of people with a gender identity different from their sex registered at birth was highest for "Other ethnic group" (1.96%) and lowest for "White: British" (0.34%). When we consider respondents with main language English, the rate for "Other ethnic group" becomes the second highest (1.27%, just slightly under the highest at 1.29%) while the rate for "White: British" stays the lowest (0.33%). 

We have conducted further research by ethnic group. We have looked at a range of hypothetical error scenarios, described in Section 13: Data sources and quality, and the results from Census 2022 for Scotland. We have concluded that we can be relatively confident that the proportions of people who identified as trans in the "Mixed or Multiple", "Other ethnic group", "White: Irish" and "White: Gypsy or Irish Traveller, Roma or Any other White background" ethnic groups were higher, compared with the "White: British" ethnic group.

However, we are less confident that the "Black, Black British, Black Welsh, Caribbean or African" and "Asian, Asian British and Asian Welsh" ethnic groups had a higher proportion of people who identified as trans, compared with the "White: British" ethnic group. This is because these groups had a higher level of uncertainty, with 92% and 91% categorised as Type B, respectively.

Religion

The rotated bar chart in Figure 4 shows that in the Census 2021 estimates, the proportion of people with a gender identity different to their sex registered at birth was highest for those who said that their religion was "Any other religion" (2.89%). There is a low level of uncertainty for this group (26% categorised as Type B) and this finding was mirrored in Scotland, so we are confident that it is valid. 

The Census 2021 gender identity estimates for those who identified as Hindu, Muslim and Sikh are also higher than for the general population. There is a high level of uncertainty in these estimates, with the proportion of trans respondents categorised as Type B above 90% for these groups. This is potentially related to English language proficiency levels for people who identified with these religions (please see our accompanying dataset). Looking into this further and using a range of hypothetical error rates, we conclude that we cannot be confident that a higher proportion of those who identified as Hindu, Muslim and Sikh are trans compared with the general population, so the Census 2021 gender identity estimates for these three religions should not be used.

In both England and Wales, and Scotland, the proportion of people reporting as trans was lowest for those who identified as Christian. The level of uncertainty is high for Christians (85% categorised as Type B in Census 2021 estimates). However, we can still be confident that the proportion of Christians who are trans is lower than most, if not all, other religions, from looking across our strands of research described in Section 3: Background and outline of new research. This is likely, at least in part, to reflect the older age profile of people who identified as Christian, as described in our Religion by age and sex, England and Wales: Census 2021 article.

Sexual orientation

The Census 2021 estimates show that a higher proportion of people who said they were lesbian, gay, bisexual, or another sexual orientation reported being trans than those who said they were straight or heterosexual (5.78%, compared with 0.32%). These findings held when applying the hypothetical error rates and we are confident that this finding is valid.

The Census 2021 estimates show that trans people are more likely to live in flats and less likely to live in detached or semi-detached houses than people with a gender identity the same as their sex registered at birth. They are more likely to live in over-crowded households or households with the ideal number of bedrooms, and are less likely to live in under-occupied households. They are more likely to rent rather than own their property, compared with people with a gender identity the same as their sex registered at birth.

The Census 2021 findings on housing match those for Scotland (except those for overcrowding and under-occupancy, for which the data for Scotland are not currently available) and hold across our research. For example, the findings still hold when considering the younger age distribution of trans people. We are therefore confident that the Census 2021 insights about housing for the trans population are valid.

Highest qualification

The Census 2021 estimates show that trans people were more likely to report having no qualifications and were less likely to report having Level 1, 2, 3, 4, or above qualifications, compared with people who did not report being trans. These findings are the opposite of what was found in Census 2022 for Scotland.

The level of uncertainty in the gender identity estimates for those with no qualifications is high, with 87% categorised as Type B. For those with Level 1, 2, or 3 qualifications, the proportion of respondents categorised as Type B was 55%. For those with a Level 4 qualification or above, it was 60%. Looking at the potential impact of this uncertainty on the estimates, and given the difference in findings to those in Scotland, we cannot be confident in the findings about gender identity and highest qualification presented in the Census 2021 estimates for England and Wales. We would therefore advise against using these estimates.

Economic activity status

The Census 2021 estimates show that trans people are less likely to be in employment and are more likely to be unemployed, compared with people with a gender identity the same as their sex registered at birth. They are also more likely to be a student, be long-term sick or disabled, be looking after home or family, or to have reported being economically inactive for another reason. Reflecting their younger age profile, they are less likely to be retired.

Looking across the strands of research described in Section 3: Background and outline of new research, we are confident that all of these findings are accurate insights about the economic activity status of trans population. However, we have concluded that the rates of trans people being more likely to be looking after home or family are similar among trans people and those with a gender identity the same as their sex registered at birth.

Health and disability

Despite having a younger age profile, the Census 2021 estimates show that trans people were more likely to report being in bad or very bad health, compared with people with a gender identity the same as their sex registered at birth. They were also more likely to report having a disability. These findings hold across all strands of research described in Section 3: Background and outline of new research and were mirrored in Scotland. We can therefore be confident that the Census 2021 estimates can be used to provide insights into the health and disability status of the trans population.

Unpaid care

The Census 2021 estimates on unpaid care show that provision of unpaid care is similar among trans people and people with a gender identity the same as their sex registered at birth. Looking across all strands of research described in Section 3: Background and outline of new research, we are confident that this finding is valid.

Gender identity

Gender identity refers to a person's sense of their own gender, whether male, female, or another category, such as non-binary. This may or may not be the same as their sex registered at birth. The question on gender identity was new for Census 2021. It was voluntary and was only asked of people aged 16 years and over.

More information on the classifications used can be found on our Gender identity variable: Census 2021 page.

Sex

This is the sex recorded by the person completing the census. The options were "Female" and "Male".

Trans

The term "trans" is used in this article to describe anyone who stated in Census 2021 that their gender identity was different to their sex registered at birth. This includes people who did not give a specific gender identity and those who identified as a trans man, trans woman, non-binary, or with another minority gender identity. 

Trans man

A trans man is someone who was registered female at birth, but now identifies as a man.

Trans woman

A trans woman is someone who was registered male at birth, but now identifies as a woman.

Non-binary

Someone who is non-binary does not identify with the binary categories of man and woman. In the Census 2021 estimates, the category includes people who identified with the specific term "non-binary" or variants of this. However, those who used other terms to describe an identity that was neither specifically man nor woman have been classed in "All other gender identities".

Type A

Type A respondents are those who:

• were coded as a trans man or a trans woman, and provided a response to the gender identity write-in that was different to their response to the sex question, for example, sex female and gender identity write-in "man"

• provided a gender identity write-in that was an unambiguously trans response, for example, "non-binary", "trans man", or "gender fluid"

Type B1

Type B1 respondents are those who ticked "No", indicating that their gender identity is different to their sex registered at birth, but did not provide a write-in.

Type B2

Type B2 respondents are those who were coded as a trans man or a trans woman, but provided a write-in that was not a clear trans response and was consistent with their response to the sex question, for example, sex male and gender identity write-in "man".

Age

A person's age on Census Day. This was 21 March 2021 in England and Wales.

Country of birth

The country in which a person was born. For people not born in one of the four parts of the UK, there was an option to select "elsewhere". People who selected "elsewhere" were asked to write in the current name for their country of birth.

Coverage adjustment process

The coverage adjustment process is what census data goes through to account for individuals and households that were missed during the initial count. By incorporating statistical methods to correct for coverage errors, the final population estimates more accurately reflect the true population. This adjustment ensures robust population outputs for smaller geographical areas and detailed demographic breakdowns.

Disability

People who assessed their day-to-day activities as limited by long-term physical or mental health conditions or illnesses are considered disabled. This definition of a disabled person meets the harmonised standard for measuring disability and is in line with the Equality Act (2010).

Dwelling type

Classifies dwellings by their type of accommodation. This could be:

• a whole house or bungalow

• a flat, maisonette, or apartment

• a temporary or mobile structure, such as a caravan

Economic activity status

People aged 16 years and over are economically active if, between 15 March and 21 March 2021, they were:

• in employment (an employee or self-employed)

• unemployed, but looking for work and could start within two weeks

• unemployed, but waiting to start a job that had been offered and accepted

It is a measure of whether a person was an active participant in the labour market during this period. Economically inactive people are those aged 16 years and over who did not have a job between 15 March to 21 March 2021, and had not looked for work between 22 February to 21 March 2021 or could not start work within two weeks.

English language proficiency

How well people whose main language is not English (English or Welsh in Wales) speak English.

Ethnic group

The ethnic group that the person completing the census feels they belong to. This could be based on their culture, family background, identity, or physical appearance.

Respondents could choose 1 out of 19 tick-box response categories, including write-in response options.

There were two stages to the ethnic group question. The respondent identified first through one of the following high-level options before selecting a tick-box:

• "Asian, Asian British or Asian Welsh"

• "Black, Black British, Black Welsh, Caribbean or African"

• "Mixed or Multiple"

• "White"

• "Other ethnic group"

Some of the tick-boxes required further identification through a write-in response.

General health

A person's assessment of the general state of their health from very good to very bad. This assessment is not based on a person's health over any specified period of time. 

Highest qualification

The highest level of qualification is derived from the question asking people to indicate all qualifications held, or their nearest equivalent. This may include foreign qualifications where they were matched to the closest UK equivalent.

Occupancy rating (bedrooms)

Occupancy rating (bedrooms) measures whether a household's accommodation is overcrowded, ideally occupied, or under-occupied. This is calculated by comparing the number of bedrooms the household requires with the number of available bedrooms. The number of bedrooms the household requires is calculated according to the Bedroom Standard, as described on our Occupancy rating for bedrooms variable: Census 2021 page.

Religion

The religion people connect or identify with (their religious affiliation), whether they practise or have belief in it. This question was voluntary.

Sexual orientation

Sexual orientation is an umbrella term covering sexual identity, attraction, and behaviour. For an individual respondent, these may not be the same. For example, someone in an opposite-sex relationship may also experience same-sex attraction, and the other way around. This means the statistics should be interpreted purely as showing how people responded to the question, rather than being about who they are attracted to or their actual relationships.

Tenure

Whether a household owns or rents the accommodation that it occupies.

Unpaid care

An unpaid carer may look after or give help or support to anyone who has long-term physical or mental ill-health conditions, illness, or problems related to old age. This does not include any activities as part of paid employment. This help can be within or outside of the carer's household.

Application of hypothetical error rates

To undertake our analysis, we split Census 2021 England and Wales respondents who reported having a gender identity different from their sex registered at birth into:

• those that we can be more confident interpreted the question as intended (Type A)

• those that we are less confident about (Type B)

We then applied a range of hypothetical error rates to Type B respondents to explore whether the patterns in the data by geography, characteristics, and other variables held in different hypothetical scenarios.

It is important to note that we have limited evidence on what the actual error rate from respondent misunderstanding is likely to have been, or how it might have varied by geography or population characteristics. As a result, we have not published these hypothetical rates. However, as part of our research, we tested different scenarios based on applying error rates ranging from 0% to 100%. We explored two types of scenarios, which included:

• applying the same error rate across subgroups

• applying higher error rates to some subgroups than others, considering factors like the level of English language proficiency among the group, which may have caused the actual error rate to be higher

Table 3 provides an example of applying error rates to the gender identity estimates by age group.

If one subgroup had a consistently higher proportion of trans people than another, across all scenarios, we could be confident that the finding is valid.

If the finding held across most, we could be fairly confident in it. If it only held in some scenarios, there is some evidence to support it, but we are less confident in its validity.

In these latter types of cases, there is an element of subjectivity around what a plausible scenario might look like, so we have considered the results alongside the findings from the other strands of research, described in Section 3: Background and outline of new research, before coming to an overall conclusion.

For the variables discussed in this article, the gender identity estimates, and the proportions of respondents categorised as each of Type A, Type B1 and Type B2, are provided in the accompanying data tables.

Related datasets

There are several related datasets users may find useful to use alongside this article, including:

• our Sexual orientation and gender identity data combining multiple variables, England and Wales: Census 2021 contains flexible datasets where you can choose the area type and coverage of Census 2021 sexual orientation and gender identity data, combined with other variables.

• our Census 2021: Quality of Census 2021 gender identity data dataset provides supporting information for the Quality of Census 2021 gender identity data release, including a table on gender identity write-ins

• Scotland's Census 2022 estimates for Scotland published on 26 March 2025 in their Flexible table builder

The publication of this article concludes our planned research and publications on the quality of the Census 2021 gender identity estimates.

Producers and users of official statistics can refer to the Office for Statistics Regulation's guide to collecting and reporting data about sex and gender identity for more information on how data about sex and gender identity should be collected and reported. Priority work is being undertaken within the Government Statistical Service (GSS) to develop best practice harmonised standards for the topic of sex and gender identity data. This includes further testing with the wording used for the England and Wales Census 2021 and Scotland's Census 2022, within the context of changing social norms and respondent and user needs. More detail on this work can be found in the GSS Harmonisation workplan, published in December 2024.