New guidance on collecting and reporting data about sex and gender identity in official statistics

[Imicola]

I've not spotted another thread on this, but I've just been made aware that the Office for Statistics Regulation has finally produced new guidance on how producers of official statistics should collect and report data about sex.

I've not read it in full yet...what I have read makes clear arguments about the need to ensure clarity around the use of terms (it doesn't provide standard definitions, but includes the interpretation of terms that they use, which I think seem clear), that sex and gender shouldn't be used interchangeably and that Sex is a protected characteristic.

I'll have a more detailed look later, but the one concern I have is that while it talks about ensuring the right response options are available (e.g. it talks about gender identity options including non-binary or whatever), it doesn't include anything mention of ensuring response options are available to cover those who don't believe in gender ideology/don't have a gender identity. In my view, any data collection which fails to provide response options which accurately reflect this means that factually incorrect data will be collected about some individuals, and this would be in contravention of the GPDR principle of accuracy (and will also undermine the quality of the resulting data).

There is an email address for feedback, so once I've read it all I'll send some in.

Collecting and reporting data about sex and gender identity in official statistics: A guide for official statistics producers – Office for Statistics Regulation (statisticsauthority.gov.uk)

Following

Thank you. This is long overdue and hopefully a sign that trans extremism is starting to retreat with facts, science and the requirement for accurate research and data being rightly prioritised.

They first surveyed people on this about 6 years ago, I remember a post here inviting contributions. Interesting that it took a fuck up of a census in between to realise that ideological language and obfuscation won't bring in accurate data. Because I'm fairly certain lots of people would have pointed that out to them, at the time.

Thanks @Imicola

They've put in a handy ! what's changed section

What’s changed in OSR’s guidance?

This guidance updates and supersedes the OSR previous draft guidance on collecting and reporting data about sex in official statistics, published in 2021.

This guidance was developed based on feedback on our pre-existing guidance, as well as our regulatory work on these issues. We sought feedback on a draft version of this guidance from official statistics producers across a wide range of government departments and the Devolved Administrations. We also met with a number of external stakeholders to gather wider views. We would like to express our gratitude to all the individuals who engaged with us, this process was instrumental in the development of this guidance.

Any future updates to this guidance will be recorded here.

That would be exceedingly handy. If it actually said what had changed.

I can't recall exactly, it's a while since I last looked, but I think the interim guidance basically just said it was complicated and didn't offer any useful guidance. So perhaps the main thing that has changed is that they have actually provided some guidance?!

Wouldn't it just. Almost as if they didn't understand what that section was for!!

Doesn't give me a lot of faith in them having understood what the rest of the document is for.

Alternatively they don’t want to explicitly remind people what the guidance said before because it was rubbish?

Deciding what data to collect is context dependent, and data controllers have to have a good reason for each piece of data.

You can't both oppose the use of gender identity as a marker, while also insisting it is recorded. For serious crime, it makes sense to record the protected characteristics of both suspects and victims. This should include transgender status.

Thanks - posting to remind me to come back and read this very useful link!

I thought this is what we want? A sex category AND a transgender status category, for clarity. I didn't think anyone was opposing a gender identity marker? Only when it's in place of a sex marker of course.

Then you could show legitimate statistics which could be of benefit to the trans community e.g. trans identified individuals on average live to X - if that's significantly lower than the population baseline then you would work out strategies to help trans identified individuals. If the lines are blurred between sex and gender, nobody gets helped!

Doesn’t it make the contentious phrase ‘gender identity’ an almost legal concept?

Yes, I think so long as they are capturing sex clearly, I'm not overly fussed about also including something on gender identity...BUT the options need to be available to allow those who don't agree with gender ideology to select an accurate answer. E.g. if the options are woman, man, non binary, unicorn... what would i pick? Woman is not my gender identity, i don't have one. The options need to recognise this and include something for the rational thinkers, otherwise ther resulting data is inaccurate.

Plus, as discussed on earlier threads, there has been research that finds many communities aren't clear / dont have a shared concept of what gender is.

So if the keep the compulsory question of sex at birth, but also want to give the option of "gender identity" then they will need to include an explanatory note. Otherwise the data will be meaningless.

The recent updated Public Sector Equality Duty guidance told public bodies they shouldn’t collect data on gender or gender identity, so how does that stack up with this?

I thought that was Kemi Baddenoch in her role as a Minister.

Office for Statistics Regulation is supposedly an independent group working on improving statistical collection.

Although I suspect the Government leans on them to include questions in line with their political direction.

As you well know, 'transgender status' is not a protected characteristic.

Yes a clear non-circular definition of gender identity would be required in order to be sure the data collected reflects the definition. I look forward to that never being provided.

Also, https://uksa.statisticsauthority.gov.uk/news/review-of-data-statistics-and-research-on-sex-and-gender/ due by August.

It’s good they have clarified that sex is important.

I hope that, in time, they will also row back from the imposter term “gender identity” which is meaningless and will return to collecting statistics based on the protected characteristic of gender reassignment.

Alternatively, if they want to use the term “gender identity” they should ask about it in the same way they ask about religious beliefs: beliefs based on faith alone, with no rational basis for believing them.

They had this to say:

Data about sex and gender identity should be explained and defined for the purpose of a particular set of statistics, and terms, including gender, should not be used interchangeably or as a substitute for each other.

About time, too.

Alternatively, if they want to use the term “gender identity” they should ask about it in the same way they ask about religious beliefs: beliefs based on faith alone, with no rational basis for believing them.

Yes this was my thinking; if you have stats on biological sex and, say, religion, then you can get an idea of what's going on in that community. Same with those who believe in a gender identity.

If you don't record sex with gender identity, then gender identity becomes worthless anyway. But sex is useful without the need to record gender identity, so is obviously more important.

PPs have pointed out that this guidance doesn't contain any guidance, and does contain the unscientific and ideological term 'gender identity'. I agree with @DadJoke that they should focus on the protected characteristics. Coupled with sex data, this tells a data analyst everything they need to know - 'gender identity' does not need to feature at all.

I think we should be responding to the call for input. TRAs are overrepresented in the tech/data arena and they could be all over this.

I started thinking about what the model questions should be. I'm hoping they're not going to get the sex question wrong (they've had years of experience of dealing with things like DSDs and errors and corrections). So what about the other question?

They should be trying to catch everyone who would answer 'yes' to the question 'do you have the protected characteristic of gender reassignment?' But that question is rather arcane. Would a better question be 'are you transgender?' 99.9% of people would understand that, and it could be footnoted.

The footnotes would have to be illustrative rather than prescriptive. Because not all people with the PC believe in gender identity, or have medical treatment, or have a GRC, or even cross-dress. The only prescriptive question I can think of is 'do you, at least sometimes, wish to be treated legally as if you were of the opposite sex, or were neither male nor female?' (Ludicrous, isn't it?)

You'll pick up a few otherwise statistically normal blue-haired they/thems, but this seems unavoidable.