Prof Alice Sullivan on the data disaster that is gender identity

Human error of this type does tend to be quite consistent over decades.

It would seem that the TRAs in NHS England have been pushing back against reality again

https://www.telegraph.co.uk/news/2025/04/01/trans-activists-urge-doctors-fight-nhs-data-overhaul/

However, is anyone else starting to get irritated by this need to establish impeccable left wing credentials before anyone can be taken seriously on this topic?

Who is doing that?

As the article is in the Telegraph in their world would be a mark against her.

Or maybe they were surprised as by and large in terms of MSM it is the right wing / Tories that recognise and support a pro sex matters position.

Which reminds me, that it is odd that the article in getting a quote from Sex Matters (glad that they know who they are), didn't add the "disclosure" (which is on their web site) that "Professor Sullivan is a member of the Sex Matters advisory group."

However, is anyone else starting to get irritated by this need to establish impeccable left wing credentials before anyone can be taken seriously on this topic?

Agreed, it's incredibly tiresome!

But I feel like it's unfortunately still needed because as Sullivan notes in her comments in the Telegraph piece, so many lefties see it as a tribal left/right issue, and as we can see from the ONS debacle, even though the left/centre left was out of power for many years it still has a considerable amount of cultural power in public institutions and academia. So speaking to the left and making it permissible to say heretical things about gender has a purpose.

It's bloody slow going though and I'm longing for the day when all the brandishing of leftie credentials is unnecessary.

I too forgive them the wallpaper/kitchen/cookery guff in my sheer joy at a data scientist talking sense on this issue. Well done Prof AS, excellent job.

With regard to male / female cancer data is it possible that, as well as patient error (deliberate or otherwise), we might have clinician error? Being tired and ticking the wrong box is not unlikely here and there, surely.

Every time I've been pregnant, healthcare professionals have inaccurately recorded my ethnicity. I would never tick the box I've been given. This is on several occasions, all predating modern notions of self-identifying. Now, in my case, it doesn't matter hugely - the chance of particular illnesses is fairly minor. But there are different odds of various illnesses, and I am keen that this is judged correctly.

Thanks for the link to the article, OP. I'm not a data scientist, but striving for clear accurate data seems a no brainer to me!

I think the point of the Sullivan report is not so much inaccurate self-reporting (so, people who are biologically male selecting “female” as their sex marker) as it is what questions institutions are actually asking. The report shows that the question has gradually shifted away from “sex” through “gender” towards “gender identity.” I agree, you can’t completely control how individuals respond to questions, but you can - and should, for good, accurate, quantitative research - be very, very careful about the questions that are asked.

Thanks for the article. It’s a nice day so I didn’t mind the extraneous detail about the wallpaper and recipe book.

I'm aware of a woman with urinary/bladder issues who, back in the 80s, self-diagnosed herself as having prostate cancer because her symptoms were similar to those experienced by her husband who had been formally diagnosed. So if people were self-reporting, I can see how these anomalies could occur.

For anyone who didn't see it the first time round, this is Alice Sullivan talking about the important of questionnaires and reliable data in Swansea in 2023. It's really interesting and she was very funny about the people organising the census:

The number of these patients each year is in single digits over all such cancers combined. This means for patients treated using the same approaches, there are too few to publish reliable statistical analyses. The alternative would be to widen the period of diagnoses to get sufficient numbers but that would be like analysing, say, breast cancer care as a single entity over the time before and after tamoxifen was introduced. This is why there are no publications. The size of the US population is much larger but they do not feel able to publish either, presumably for similar reasons.

I do and I am not talking about male breast cancer in the above. The registration record is a mixture of medically reported and patient reported data. The pathology reports from which the type of cancer is recorded is not self reported. Afaik, things like “sex”, ethnicity are taken from forms patients complete when they commence treatment. People can change their responses to (at least) these questions and researchers have had to develop rules to deal with different responses at different times by the same person. Changes to self reported “sex” are much less common than changes to self reported ethnicity (which are most commonly between the “prefer not to say” and an ethnicity label)

we need to record biological sex in all sorts of contexts to understand what's going in the world around us and make sensible plans

@Gasp0deTheW0nderD0g

Hear hear. I would go further and say all data is potentially good data, and birth sex could be good data even 'out of context'. It might not be obviously relevant, it might be a total surprise that it turns out to correlate with some other, apparently unrelated, measurable, but that's all the more reason to collect it.

Sorry, it did load, but slowly.

Link not loaded.

I think this letter, explaining how the Great 2021 Census Gender Cock Up happened should be compulsory reading for every single doctor, teacher, health professional, MP and anyone else who thinks the much vaunted 0.5% of the adult population identifying as trans is a given and verifiable fact.

It simply isn’t true, it is abundantly clear that in areas where many of the population have poor written English skills as English is an additional language the question was misinterpreted and the data received is compromised beyond redemption, and while this isn’t mentioned I would venture to suggest that for many of the population who do have English as a first language but whose reading skills are limited the same problem of misunderstanding almost certainly arose since the question was so poorly worded.

Yes, thank you, this is very interesting. I found their website, https://digital.nhs.uk/ndrs. They automatically record cancer diagnoses, treatment and outcomes, as well as rare diseases.
You can opt out, but I've had cancer and had no idea this record was being made.

For now, her focus is on the forthcoming second part of her review which looks at barriers to research on sex and gender, primarily in universities.

Any research that correctly sexes its subjects is transphobic (see: Phoenix v OU). Presumably 'pure' medical research gets a free pass, or things are getting really dangerous.

DSD's can create an increased risk of cancer, which might account for some of the confusion in those figures.

Very interesting @PollyNomial could you link to some more info about this?

You sound as if you may have professional expertise here, which I haven't, but firstly breast cancer can occur in males. Secondly, if it's mostly self-report we know that people's knowledge of what's going on in their bodies, which body parts they have, and so on, is abysmal, not to mention high rates of illiteracy and very poor literacy in the past. So many women (my own grandmother included) believe they urinate out of their vagina, for example. So I could believe they pick the wrong thing to tick on the box, if that's how it works.

Archive: https://archive.fo/SmdF2

If they are mistakes, they’re at a remarkably consistent level over more than half a century. And one has to remember that most health data is self reported which is why “sex” has pretty much always been (at best) “augmented sex”. An example where this isn’t the case is genetic testing data but this is relatively recent and most patients don’t have these.

This is absolutely shocking

This review is providing an academic gloss on what is a political call to strip trans people of our hard-won rights to privacy, dignity, and respect in public spaces.”

TRA have managed to destroy their own dignity & respect by their utterly appalling bullying behaviour