Gender or sex

[govuk-design-system]

Use this issue to discuss the gender or sex pattern in the GOV.UK Design System.

What

When and how to ask about people's gender and sex.

[Zeno001]

Unfortunately, I can't access the draft guidance so these are my comments on the current guidance - I hope they are useful.

There is certainly confusion in the minds of the general public about the meaning of the terms 'sex' and 'gender' (they generally think they mean the same thing) but the Equality Act 2010 ("the Act") gives 'sex' as one of the protected characteristics. Gender is not a protected characteristic nor is the term even defined.

The Act at s.11 states, inter alia:

"Sex

In relation to the protected characteristic of sex—

(a)a reference to a person who has a particular protected characteristic is a reference to a man or to a woman;"

The Act at s.212 states, inter alia:

"“man” means a male of any age;"

and

"“woman” means a female of any age."

You say:

"If you do need to ask, use ‘sex’ when you need biological data (for example, if you’re providing a medical service). In all other cases, use ‘gender’."

I think the default should be to ask for 'sex' and it's not at all clear under what circumstances anyone would be asked for their gender. If it is information for an equality impact assessment that is being sought, then it has to be 'sex' that is asked for.

However, where the sex of the individual does need to be known, I would suggest the form advocated by NHS England be used. Their document, Accessible Information Standard Review: Diversity Monitoring Questions, asks 'What is your sex?', giving the options 'Male', 'Female', 'Intersex' and 'Prefer not to say'. This covers all possible cases.

You say:

"If you have to ask about gender, you should:

list the fields in alphabetical order do research to test that this works for your users"

Again, it's not clear when this information would be required - it's certainly nothing to do with any Public Sector Equality Duty - but there are many different views on what labels to give genders and even asking for it and providing a list is fraught with difficulties. I think there would need to be a clear identified specific purpose for collecting and lawfully processing this personal information to be compliant with the GDPR, etc.

You say:

"Don’t use titles to guess gender You shouldn’t guess someone’s gender based on a title because:

some titles aren’t gendered (for example Dr, Rev, Major)"

I think this conflates sex with gender and I suggest that it is sex that shouldn't be assumed from a name or title.

"titles can be changed by deed poll to one that’s different from a person’s gender or sex"

I don't understand this: names can be changed by Deed Poll (in England and Wales but not in Scotland and I'm not sure about Northern Ireland), but surely titles can't be changed?

I would also suggest you might like to give guidance on titles: there is an endless list of possibilities but please consider not making any title a required field - I prefer not to use one at all and I don't think there are any situations where it is necessary. It is convenient to sometimes refer to people as, say, Mrs Smith, but I suggest that if a title is not given it should be easy enough to use the individual's name.

[chrisduffield]

I have to agree with Zeno001, why would you ask for gender? It is useless for any official statistical purposes which would be looking at the protected characteristics under the equality act.

Also why have a link to draft guidance if clicking on it doesn't allow you to read it? If you are going to ask for opinions then allow them to be informed opinions.

[Zeno001]

Thanks, Chris. A primary rule of any data gathering is ensuring you understand why it is being collected. Article 5(1)(b) of the GDPR requires that personal data shall be:

"collected for specified, explicit and legitimate purposes and not further processed in a manner that is incompatible with those purposes"

Since there doesn't seem to be any 'official' or legal definition of the term 'gender', it would seem to be difficult to construct a case for gathering personal data on it. Unless and until someone can come up with a cogent definition of the term and a legitimate use for the data, it seems to me all the Government should be asking for is the sex of the person. Even then, there are probably limited occasions where this is relevant.

[chrisduffield]

Also can you give informed consent if the definition of gender is ambiguous?

[Zeno001]

That's a very good point, Chris.

Are the authors here able to provide any information on any of this?

[alexyingchen]

I echo what people have mentioned already in terms of conflating sex with gender and questioning why you need to ask someone's gender.

Adding to this conversation from the perspective of a transmasculine nonbinary person:

If you are asking for sex (say for medical/health surveys) it makes sense to me to have these 4 options:

• Intersex
• Assigned female at birth
• Assigned male at birth
• Self-describe

The reason I say assigned is because these are judgments that a doctor made. They may not be entirely accurate. There may also be some overlap between someone being female or intersex.

I personally believe you should ask for gender in the census to account for diversity. I can't think of many other reasons to ask for it.

If you do ask for gender, I think 2 options are valid and might require some A/B testing with trans folks. Option 1: a list of checkboxes including options like man, woman, nonbinary, genderfluid, agender, two-spirit, etc. Option 1: open text field to self-describe.

I would definitely recommend against using language like male or female when asking for gender. Those feel like medical terms.

[Zeno001]

Sorry, Alex, those four options don't make any sense for a variety of reasons.

Intersex - or Difference of Sex Development - is simply a set of developmental conditions and do represent any separate sex class. Those with a DSD are still considered to be of the male or female sex class.

Sex is not 'assigned' at birth: it is observed and recorded.

Sex cannot be 'self-described' as it can only be either male or female.

Before anyone can be asked to provide their 'gender' there first would have to be a cogent definition of it and I too question why that needs to be known. However, it must not be confused or conflated with a person's sex and it is this that is most commonly required. For example, the Equality Act 2010 defines the protected characteristic of sex where sex is a reference to only female or male. In terms of health, it is sex that is a determinant.

[alexyingchen]

I think you misinterpreted my part of my comment. I was not suggesting sex and gender are the same thing, nor am I suggesting you are required to ask for either in a form. Like I said, if you need to ask for sex, then I have this suggestion. If you don't need to ask, then obviously don't.

Intersex is a sex class and intersex children are often forced to undergo surgeries against their informed consent. Check out the Intersex Justice Project for more http://www.intersexjusticeproject.org/about-us.html and the work of Pidgeon Pagonis http://www.pidgeonismy.name/about-flatiron

Assigned (x) at birth is a very common term in the trans community.

If you feel like self-describe does not make sense in this context I can agree with that.

But I adamantly disagree with everything else you said. Does your team have any trans or intersex people? If not, you are all looking at this page from a very cisgender perspective and I urge you to do more research about trans and intersex bodies.

[Zeno001]

My apologies if I misinterpreted your comment and I fully agree that unless you really need to know someone's sex, then you shouldn't be asking for it. For the census, for example, it is important because the NHS and other bodies need to know the numbers of each sex in the population in different areas, but I agree data gatherers should be carefully considering whether they need to record sex. However, it is important in some circumstances to record sex - and as an example, I'm thinking of where public authorities, etc need to ensure or demonstrate they are not discriminating against one sex or the other, say, in recruitment: if they don't record the data, they will never know.

'Assigning' at birth may be common in some circles but it is inaccurate. For over 99% of the population, sex is obvious and is simply identified and recorded at birth:

"Sex has a biological basis, whereas gender is fundamentally a social expression. Thus, sex is not assigned—chromosomal sex is determined at conception and immutable. A newborn’s phenotypic sex, established in utero, merely becomes apparent after birth, with intersex being a rare exception."[1]

and

"Humans are sexually dimorphic, with rare intersex conditions being anomalous developments of dimorphic sexual classes. It is not possible to change biological sex. There is no agreed scientific basis for someone having the mind of someone from the opposite sex or being born in the wrong body."[2]

The problem with using 'assigned' is that it can lead some to believe that a baby's sex is in some way an arbitrary or capricious choice that was made at that time and/or is mutable.

I happen to have several trans and intersex friends: the trans ones are in no doubt about their sex, regardless of surgery they've had or treatments they are taking, and one particular intersex friend is an advocate for those with DSDs and writes frequently about the unnecessary treatment of babies. You might like this recent interview with her: "Biological sex is not a spectrum: there are only two sexes in humans"

However, I think we have wandered quite a way off the original topic!

---

1 Byng R, Bewley S, Clifford D, et al. Gender-questioning children deserve better science. The Lancet 2018;392:2435. doi:10.1016/S0140-6736(18)32223-2

2 Bewley S, Clifford D, McCartney M, et al. Gender incongruence in children, adolescents, and adults. Br J Gen Pract 2019;69:170–1. doi:10.3399/bjgp19X701909

[alexyingchen]

I think we are still pretty on topic. I didn't ask how many trans friends you have, I asked how many are employed to work on this design system. I understand that I'm not going to change your opinion, so I'm leaving this comment for documentation and future reference, i.e. if this task is ever taken out the backlog.

The underlying reason why doctors perform harmful surgeries on intersex children is because they believe human sex is binary. If intersex is a legitimate separate class, they would not be mutilating genitalia or forcing hormone treatment because there is no motivation to do so. To simultaneously denounce these procedures and uphold binary sex is contradictory.

And post-surgery, if society views these folks as male or female, we are making that categorization after an act of violence, without their consent.

That interview was very self-contradictory. The one thing I learned is this problem space requires vigorous user research, because the intersex people I know would choose "intersex" instead of "male" or "female" but it sounds like the interviewee might choose "female," "intersex female," or "female with intersex characteristics."

Turning off notifications for this thread.

[Zeno001]

You asked me about my team I don't have a team: my friends are my 'team'.

Thanks for the discussion.

[helengaskell]

Has anyone completed any research on this pattern? My project is exploring this area as there is currently a mandatory M/F database field in some legacy systems which we may be required to use. I am in touch with GDS but would like to know if any other GOV.UK projects have used this pattern and if so, how.

Edited to add: We will be conducting user research on our proposed use of this pattern with trans/intersex/non-binary users and will share the results here if/when possible.

[chrisduffield]

Why are you only looking to research using trans/intersex/non-binary people? This will probably lead to a very different result than looking at the wider population you are targeting.

As I stated before there is no legal obligation for gender but there is a legal obligation to ensure that the service does not discriminate based on sex.

From: Helen Gaskell [email protected] Sent: 15 January 2020 13:58 To: alphagov/govuk-design-system-backlog [email protected] Cc: Chris Duffield [email protected]; Comment [email protected] Subject: Re: [alphagov/govuk-design-system-backlog] Gender or sex (#69)

Has anyone completed any research on this pattern? My project is exploring this area as there is currently a mandatory M/F database field in some legacy systems which we may be required to use. I am in touch with GDS but would like to know if any other GOV.UK projects have used this pattern and if so, how.

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/alphagov/govuk-design-system-backlog/issues/69?email_source=notifications&email_token=AEDIZAT6UQNTGHKBR6BCMY3Q54I7HA5CNFSM4ELSE6M2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEJAMQFI#issuecomment-574670869, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AEDIZATR6CUYDHSPZLO6J4DQ54I7HANCNFSM4ELSE6MQ.

[helengaskell]

@chrisduffield I am not. Our prototype has been extensively tested across a wide range of users (over 160) and we have not asked research participants about their sex/gender to date. However I believe asking this question will have a disproportionally negative emotional effect on trans, intersex and non-binary users. Our research will either confirm or contradict this hypothesis.

One of the many reasons we are conducting this research is that at least one person from every group of research participants has asked us why we are asking this question, and many have followed up with questions about how this will affect people who are trans, intersex or non-binary. I understand that the majority of people are cis-gender, but there is a very wide understanding of this challenge amongst our users.

[helengaskell]

I would also like to add that the service I am working on will not discriminate on the basis of gender or sex. This data may end up being collected because legacy systems have mandatory data fields, but we are still exploring this - including whether those mandatory fields can be removed, thereby removing the need for the question altogether. The sex or gender of service users is not and will not be technologically or manually connected to decision-making whatsoever.

I am commenting here to see how other services have dealt with this situation and to see what research has been done.

[Zeno001]

Can you say more about your research, Helen? What's the demographic and area?

[helengaskell]

I'm here looking for research. Our service is largely for non-UK nationals (often speakers of English as a second or third language) so any research on this pattern specifically dedicated to that user base would be useful. My team does of course have a user researcher and I will share any research she conducts that is relevant to this pattern.

[helengaskell]

So far one user in every five has questioned why we are asking for this information and raised an objection. Many non-UK passports have a third option and our user researcher has come across this being an issue several times when a paper-based alternative (i.e. legacy form which includes this info) is used. So research done to date would suggest that where this question is asked, we should provide information about why we are asking it.

[Zeno001]

It's clear that knowing whether a person is female or male is important for medical and some other reasons, but it's difficult to see what relevance it has for other areas. I suspect many ask for it without thinking but, like all other personal data, there needs to be a lawful basis for asking for it. Passports (and driving licences) are not a reliable source of information for a person's sex.

[caionishi]

I'd like to let you my 5 cents here (as a transgender UX who works on healthcare for 5 years). I'm outside the UK scenario - but have in mind that I was looking to solve this from a global perspective.

1. To solve the problem with the general understanding of "sex" and "gender", just extend the label a little bit more. In this case, "Biological sex" or "Sex at birth", and "Gender identity". This complementation seems simple, but incredibly inclusive and a little more "self-"explanatory than just giving "Sex" or "Gender" as an option when you are not able to provide a tooltip (or another resource) on interface to give an explanation about them or why are being asked.

2. So, as suggested by the current guide available, by default you should only see "gender" being asked. This means the service/whatever would use this info to apply the correct pronouns and "service" experience. In a clinical context, you have to ask for both: a) sex will orientate about clinical treatment/calculations/etc, b) gender will orientate of how this person must be treated.

3. Just to register (for future readers), that there's no reason to ask about Sexual orientation.

4. The key point is also about which options would be given to answer the "sex" and "gender" fields. Have in mind that you have to simplify as much as possible, as well as preserving the inclusion and diversity for all. So, for "Sex", as mentioned, Male/Female/Intersex refers to the possibilities you have at birth. For "Gender", it's where you should have 4 options: Male (cis and trans), Female (cis and trans), Other, and, Prefer not to say. The point here is that it doesn't matter if you're cis or trans, both may have the same gender. "Other" instead of "non-binary" turns it broader once "non-binary" is not exactly applicable to all (e.g. fluid) - and here we'll have a great range of possibilities, that from a form-point-of-view, are respectfully included under "Other". And "Prefer not to say" is needed to respect the user in case of him/her/they are questioning or similar - just keep in mind it's a different status than "Other".

5. When using a "Gender" field, instead of following the approach suggested by the guide to "avoid the use of pronouns", just ask which one is the preferred one by the user. As simple as that. From a development point of view, there's no logic to be applied between the sex/gender/pronoun that will correctly correlate them and support all possible scenarios. Diversity really means diversity :)

[Zeno001]

Sex is immutable[1] so there is no need for "Biological sex" or "Sex at birth"; sex on it's own is a sufficient descriptor - the modifiers provide no additional detail or information.

One main and important reason for recording a person's sex is for health/medical reasons. Another is to monitor possible discrimination In terms of the Equality Act 2010;[2] one of the protected characteristics is sex and s.11 of the Act states, inter alia:

Sex In relation to the protected characteristic of sex— (a) a reference to a person who has a particular protected characteristic is a reference to a man or to a woman;

where (at s.212):

"man” means a male of any age; “woman” means a female of any age.

Intersex - or rather those with a Difference of Sex Development (DSD) - are not a separate sex class: DSDs are simply anomalous developments of dimorphic sex classes: they are either still make or female and not some intermediate or other sex class.

This is entirely separate from any notion of 'gender' - this is a term not defined in any UK legislation.

However, if someone wants to record information about someone's 'gender' they would need to adequately describe what was meant by the term and be sure they had a lawful basis for processing that personal data under GDPR. It should also not be confused or conflated with sex.

In terms of sexual orientation, this is also a protected characteristic under the Act so monitoring it is essential for equality purposes.

---

1 Bewley S, Clifford D, McCartney M, et al. Gender incongruence in children, adolescents, and adults. Br J Gen Pract 2019;69:170–1. doi:10.3399/bjgp19X701909

2 Equality Act 2010. https://www.legislation.gov.uk/ukpga/2010/15/contents.

[caionishi]

Thanks, @Zeno001 . Not sure if you got I understood and agreed with what you said (now more extensively), just gave a UX perspective (what it matters from an experience point of view, not legal - and where we have the issue and reason for creating a guideline around it).

[Zeno001]

Thanks for that, caminishi. I was trying to put a legal and UK perspective on this. The user experience is important but I think there has to be a foundation of what information is required by various laws and use-cases. I don't claim to know all of them so I confined my comment to just health and the Equality Act, which is the starting point for many instances of asking for sex.

Different countries will have different cultures around the word sex: the (sometimes rather reserved) British have for a while used the word gender to mean sex in common language, but the meaning of the term gender nowadays is extremely unclear and it is the term sex that is written into many laws so I think that has to be what is asked - and I think most people know what is being asked when that term is used. The same can't be said for the term 'gender'.

[caionishi]

Well, laws evolve as well terminologies and broader understanding of them, right? Looking for the thread, I'm unsure if you got the point this is not about opinions or current legal definitions of something, but to turn it better until works well enough for who matters (and here it comes researches as the one mentioned by @helengaskell).

Btw, @helengaskell it would be great to exchange some info if you're still working on these topics :)

[Zeno001]

But if the information is being collected by government for health or legal purposes then it needs to be accurate in the first instance. If the law was to evolve then that would be a reason to reconsider the terms, but not before.

[DeanMarkTaylor]

For those following this thread you may be interested in this video: SciShow: There Are More Than Two Human Sexes https://www.youtube.com/watch?v=kT0HJkr1jj4

It details why defining a persons "sex" is not that straightforward.

Reference documents are included in the video description which includes: NHS: Differences in sex development

[Zeno001]

The journal paper I cited earlier (Bewley et al. 2019) states:

Humans are sexually dimorphic, with rare intersex conditions being anomalous developments of dimorphic sexual classes. It is not possible to change biological sex.

ie sex is not a spectrum and it is immutable.

As three evolutionary/developmental biologists explained it:

sexual reproduction in almost all higher species, including humans, proceeds via fusion of one small and one large gamete (anisogamy). “Sex” refers to one of the two reproductive roles in this process. Individuals that have developed anatomies for producing either small or large gametes, regardless of their past, present or future functionality, are referred to as “males” and “females”, respectively. No anisogametic species produces more than two different gamete types. Thus, there are precisely two reproductive roles in anisogametic species. Thus, there are precisely two sexes — sex is binary.

There is no peer-reviewed scientific paper that even supports, let alone proves that sex is anything other than binary.

[thomasforth]

The current guidance suggests that "You should only ask users about gender or sex if you genuinely cannot provide your service without this information". I propose changing to this to "You should only ask users about gender or sex if you cannot provide your service without this information or if knowledge of the user's gender or sex would improve the service".

The prompt for this change is that the Coronavirus Status Checker service, released on 2020-04-05, does not (as of writing) ask the user's gender or sex. This seems to be in keeping with this design guide, but since we know that Covid-19 mortality rate is dependent on sex it lowers the value of the service. This has been pointed out by, for example, Caroline Criado Perez on twitter. I will try and make the same suggestion to NHS England if changes here do not propagate to them.

[Zeno001]

Indeed. The sex of the person is very important in understanding risks and deciding treatment in many medical conditions and not asking for the person's sex is a serious omission in that NHS Covid-19 tool (gender is irrelevant).

In terms of the wording, account has to be taken of GDPR and asking for users' sex should only be done if you have a lawful basis for processing those data.

[vickytnz]

@thomasforth i would be mindful that the example you're using - a medical checker - is very different from most services.

I view the guidance's purpose as giving reason for service teams to 'push back' when asked to ask about sex or gender for statistical reasons. More blurry terms like 'improves the service' opens the opportunities for arguments that getting demographic data improves it (particularly as google analytics becomes less useful).

If the data is not as useful without sex, then the original guidelines stand that users can't really complete the service (give sufficient information to government) without it

Alternatively the field could be optional. That wouldn't go against the guidelines, and would encourage your team to explain why it's useful for users to give this information.