1 in 67 English/Welsh Muslims are transgender

In a word. No.

You are conflating issues.

I don't know if someone else has raised this question because I can't read 8 pages right now...

Does this not mean that the data for sexual violence convictions, and the very large numbers, is even worse than we thought?

Because it was extrapolated from the census how many people said they were trans and combined with crime statistics. So if 1 in 67 muslims accidently said they were trans, does it not mean the % of convicted trans sex offenders is way higher than thought?

Can someone do some maths with this?

Agree data on the number of trans people is a useful thing to have.

Interested by this from ONS:
We developed and tested various different formulations of the gender identity question, ultimately using the version that was shown to give the most accurate results.

Is this because they only tested it in the relevant population and didn't consider the risk of false positives from the much larger non-trans population overwhelming the number of true positives? Or is it an admission that lots of trans people in the test groups would have selected "no" if asked a question with a simpler formulation like "Are you transgender?" ?

Either one is not great.

It's also worth considering what is the smallest population size that can be accurately measured in the census, given that there will be a base rate of errors?

I think the key point from that article is when they highlight that it's arisen because the group of interest (trans people) is such a small group. People with poor English skills may make errors in other questions too (they did at least have the option of skipping the gender identity over which I'm sure many did) But it doesn't make a huge difference to the overall figures reported because the actual numbers are much larger. So it's partly a difficulty with wording a question on such a woolly topic and partly a problem of trying to count a group that's only 0.2% of the population.

I do repeat though that the 0.2% figure (who wrote in the words "trans man" or "trans woman") rather than the 0.5% figure who ticked a box is probably reasonably robust. Having actual data on the number of people who are trans is useful

EXCELLENT!

Biggs said the authors “never thought about how a Bangladeshi grandmother or a Hungarian plumber will think about this question”, adding: “I’d be disappointed if a master’s student at Oxford made that error.”

We know this is the heart of the problem.

Alice Sullivan, a professor of sociology from University College London who has also expressed concerns about the question, said: “It’s such a shame because it’s such a missed opportunity. This should have been the gold standard. It’s so hard to understand how the ONS could fail ‘questionnaire design 101’ in this way. It is particularly important when asking about a small minority group. It’s mind-blowing to me.”

Nice to see TWO university professors on the case from different universities expressing such horror at how much of a basic error this is.

The ONS maintained that the 2021 census results on its gender identity question were “broadly consistent with NHS data collected in the same year”. “However,” a spokesman added, “while the question on gender identity was tested thoroughly, it is possible that individual responses were affected by different interpretations of the question and we will do more work to understand whether that was an issue.

And here we have the ONS doubling down, insisting the data is actually ok and the question was tested thoroughly and then blaming individual members of the public rather than the fact its disappeared up its own arse. Whilst saying they probably should look at it because they have to.

Brilliant! Didn’t know there was a stats watchdog. Hope this wakes them up to the scale of these data problems

Watchdog to examine ‘implausible’ UK census trans figures

www.thetimes.co.uk/article/cffadd74-d889-11ed-80bc-e358583c5d62?shareToken=f6bf85eeac659d8e129c40853824bb31

This second tweet of one about cordial discussion but bringing in "gendered digital misinformation" doesn't appear to be ironic:

https://twitter.com/nancymmk/status/1645747259525545984?s=46&t=WHoOZ3Kv5G6-FyQuvE0LQ

"To forestall the obvious. I KNOW that all women in the public eye get abused on social media. Gendered digital misinformation and gendered online abuse is everywhere. I experience it daily."

I think it simply gets harder and harder to take Stonewall seriously when they are led by someone so unprofessional and incompetent and they are pushing quasi-religious bullshit which isn't backed by robust quality data.

I don't think young women help matters.

Yesterday we watched Loose women (first time in a year or so) and Hearsay/Coronation Street Kim Marsh was on, her make-up was pretty standard, where as her daughter wore clothes and make-up you would see on a drag Queen.

It’s as we’ve always said: a men’s rights and men’s sexual access movement. Just dressed up in 21 century clothes

The literacy trust do some useful work in relation to early years education.

But that also, kind of, illustrates the same
point. That’s the photogenic end of literacy issues.

The ridiculous rates of illiteracy in the prison population, for example, is a very big issue for society. But not one you can make a nice glossy poster about.

I think part of the stonewall problem is that they’re not interested in representativeness or participation. They’re really too focused on the exceptional outliers (the most vulnerable, the most marginalised, etc).

It means that all those pretty mainstream and ordinary gay and lesbian people get forgotten. They’re not interesting enough to merit attention. Who cares about the ways in which homophobia continues to create barriers in their lives, especially for lesbians when there’s a more special
group to shout loudly around?

The carnivalesque-as-default presentation of LGB-and-the-list-of-many-letters is not really that helpful to the majority of people the category claims to include. Life is not a pride parade and, frankly, most LGB people are not reasonably represented by sparkly rainbows, fetish gear or whatever - that’s true even if every so often they might dress up and join a pride parade or whatever.

The boring, non-sparkly, issues that affect people’s everyday lives do matter. Pushing those boring, ordinary lesbians aside to focus on drag queens is simply not helpful. Sure, it’s easier from a marketing perspective to show images of trans women in sparkly dresses than to try to present lesbians who often just look pretty much like many straight women (especially given the wide variety of ways in which any group of women might choose to style themselves) doing ordinary things. It’s all more eye catching and distinctive. But it’s not representative.

Stonewall should actually give a shit about representing the significant proportion of the population who just are LGB without needing to be spectacular or exceptional in doing so. Instead they’re off trying to ‘save’ the ‘most vulnerable’, redesigning flags to include ever more colours and just missing the whole point.

When is our day of awareness of adult English illiteracy? Name an organisation who advocates this and gets huge amounts of media coverage on the subject?

Here is the national literacy trust (nope I'd never heard of them before either)
https://literacytrust.org.uk/parents-and-families/adult-literacy/

They say here that
16.4% of adults in England, or 7.1 million people, can be described as having 'very poor literacy skills.' They can understand short straightforward texts on familiar topics accurately and independently, and obtain information from everyday sources, but reading information from unfamiliar sources, or on unfamiliar topics, could cause problems. This is also known as being functionally illiterate.

I think this is actually now outdated information as ironically the latest census should be the most recent data set with information along these lines.

That's a lot more people than there are trans identifying...

...yet the highly literate and educated ONS census bods overlooked this gem in favour of the importance of trans visibility.

Like Nancy Kelley that's fucked up and a dereliction of duty. It highlights their unsuitability to do the job and the fact they have been way over promoted in not understanding THE most basic part of the brief for the premise of the census. Be simple enough so everyone can understand enough to give data that's useable and has any value.

Instead it looks like they've produced shit that's worse than useless. It's effectively potentially harmful.

InvisibleDragon and RedtoothBrush

Thank you for your posts very helpful. I can absolutely see that eliding different groups results in the dominant group consolidating their existing power. That’s really depressing. Sex absolutely matters.

I think the very best example of bullshit stats is to say that if you were describing a human to an alien who had never seen a human it would be accurate to say the average person has less than two legs.

We know that this 'average' person is not the majority of people walking around. It is not an accurate representation. You have to look a bit beyond that to understand why is actually being said when you say the average human has 1.8 legs. The reality is that most people have 2 legs. A small number have 1 and a bit legs, some have 1 leg and a very small number have no legs. And that's what society looks like and that's what society should cater for.

We need to understand that most people have 2 legs but some are disadvantaged and have less so should be catered for rather than be invisible to society. This is a more sophisticated level of understanding which better reflects reality and social disadvantages.

Understanding the concept of why accurate data collection matters is hugely important.

Any self proclaimed data geek should know that, when you look at data for important decision making you really should be also having a good look at the quality and reliability of that dataset and looking for possible data bias and flaws in methodology. Nancy Kelley really is scoring a big fat fail on this.

Ben Goldacre and Margaret McCartney are two of the best authors on this subject on a wider level - I'd recommend them together with Perez's Invisible Women. All three are widely acclaimed and respected as being essential reading in terms of understanding data. They are NOT niche.

For Nancy Kelley to be so blithely ignorant and willfully blind as to spot the anomaly as being inconsistent with other known data and her own knowledge of her own community is a catastrophic admission of being unsuitable for the roll she has at the head of a diversity charity where this stuff is the basic bread and butter of enabling change for the better. It really really matters.

Her failure to ask questions is deeply deeply problematic. The male lesbian problem at Stonewall who are trying to influence others isn't a mere blip it's institutionalised sexism and homophobia. It's a feature of Stonewall to lack a professional level of understanding data and other disadvantaged groups. It's built on the ideology and belief that transwomen are the most vulnerable rather than having a good understanding of information and data to present as evidence to create a compelling argument that this is actually the case. (It's clearly not as I say above).

At company X, 20 employees identify as LGBT according to their figures. This is our of 1000 employees. So 2% of the workforce. There's a slight under representation of LGBT employees compared to the general population.

However because there's no breakdown in the data between lesbians, gay men and transwomen you don't see that there's no lesbians and 17 gay men and 3 transwomen. This means there is actually an over representation of both male groups and an under representation of lesbians. In fact it looks possible that the company looks actively a hostile workplace for lesbians. They either don't work there or they don't feel comfortable to be out of the closet. Company X is actually a deeply sexist environment to work in.

But the all male LGBT representative panel can argue that they need to employ more gay men or transwomen because LGBT people are under represented in the company. This is the premise of Invisible Women mentioned above - when you go gender neutral and talk about people rather than males and females, this tends to disadvantage women because they become invisible and in becoming invisible any way of being able to see bias or disadvantage disappears with it. Far from increasing diversity, it actually works to decrease it.

Company X on seeing the data could then invite Stonewall to do the whole diversity programme thus making it more hostile to lesbians (with the whole TWAW and can be lesbians bullshit) because of the premise of under representation of the LGBT community. Except gay men and transwomen are actually already over represented and the problem is actually lesbian visibility and lesbian employment. Which Stonewall is institutional and politically tone deaf to.

The inward systematic looking in on the perceived problem rather than actually correctly identifying the actual problem in the first place due to missing important data, leads to the situation becoming worse rather than improving for the under represented group in something of a vicious circle.

This is why sex matters in the details on representation. A sex blind data set has the danger of misrepresenting what's actually happening.

The company has no way of seeing the invisible lesbian issue, much less addressing it without it.

I hope this illustrates the point in a way that's understandable.

The ONS's flawed data on gender in areas of high levels of low literacy will inevitably cause a similar problem to the one at Company X as money is invested in a none existent trans community. Money given to this folly is taken away from the vulnerable groups that are most in need of that cash for actual real world problems instead. There is likely to be a heavier pushing for unisex facilities in the area, when this is culturally highly inappropriate. It's more likely to result in the disappearance of Muslim women from the public sphere back into homes, and this in turn reduces their opportunities to improve their independence and literacy levels. It makes them more vulnerable to things like domestic abuse. All because some jumped up privileged pricks can't comprehend why simple English and simple concept that are universally understood should only be on the census.

It highlights well why the argument that transwomen are the most marginalised in society, simply can't be true. Their need to be identified and recognised is institutionalised into the ONS and political parties whilst the needs and visibility of people who don't speak or read English are completely forgotten about.

I have little faith that this finding will be taken on board by councils or anyone in planning unfortunately, so there will be ramifications for years and other surveys / institutions will continue to overlook the importance of simple English (hello NHS) particularly to the detriment of women.

Previous post was meant to be @Slothtoes !

Fantastic post @RedToothBrush.

@Slo
I think what RTB means about disappearing lesbians is that if combined LGBTQ totals are used a company can technically meet it's diversity quota if say 10% of its employees are gay men. Fantastic, amazing and there is an LGBTQ person on the board of directors as well (another gay man). But if none of those LGBTQ people are lesbians (and certainly not on the board of directors), then lesbians are still totally under represented across the company and in senior positions, whilst the company has met its diversity targets by ... Employing men.

That's all true @RedToothBrush . The same issues arise with ethnicity and lumping all non-white British groups in together for analysis. However, if the size of your group is small, it'll never be possible to have representative numbers of these very small groups. You'd have to have an organisation of at least 1000 for even one trans man and one trans woman to be expected to be included, if the group was representative (alongside about 10 gay men, 12 gay women, 4 bisexual men, 9 bisexual women and 2 people of other sexualities). So being gay is much, much more common than being trans, according to the census, which could be surprising to many people given the degree of media coverage trans issues get.

The group I'm working with at work has a size of 13 so that "correct" proportion of all LGBT groups combined is still less than one whole person!

Please can you say more about the disappearance of lesbians with lgbtq+ statistics? I get it absolutely within lgbtq+ policy.

Do you mean that lesbians are more likely to tick a different sexual orientation or to tick ‘don’t want to say’ than people with other sexual orientations, including on a census, out of mistrust of how stats would be used against them? Sounds like another underserved group where extensive user testing of the census questions and accompanying explainer info should have been focused in advance, in that case?

Isn't it funny how self proclaimed stat nerd Nancy Kelley, hasn't picked up on the amazing disappearing lesbian stats yet...

👏 👏 👏

The statical bollocks of lumping the LGBT together, is another of those problematic figures.

It's used to say 'lgbt groups are under represented'. But there's something very interesting within that data too.

I can't be arsed to go hunting for the BBC's data on its employees, but it was striking to see how OVER represented gay men were at the corporation. Technically to hit their diversity targets the BBC should be looking to reduce the number of gay men it employs. Especially gay men in higher ranked positions. Lesbians were surprising under represented. It begs a few questions.

After talking about it on MN I remember seeing a conversation on MN amongst lesbians who don't like to tell their workplace their sexuality because of the harassment it has led to in the past or because they feel it somehow restricts them in their career due to stereotypes.

Then there is the representation in political parties which has been noticed on FWR. The LGBT Lib Dems are the best example. Last time I looked at the staff on their executive it was pretty revealing. Their lesbian representation was male. In fact the entire exec was male bar one female who identified as non-binary. Then there's Stonewall itself. I remember seeing a photo not too long ago, which made another thread on MN. It was of Stonewall's key personnel. You've got the token Nancy Kelley and then another massive representation of males. This pattern repeats through politics. Gay men are the most over represented group in parliament. Lesbians are under represented. (Actually young women between 20 and 50 are one of the most under represented group - no prizes for guessing why. Women in politics overwhelming tend to be over 50, particularly in local politics.)

You have this amazing disappearing lesbian phenomena hidden by the bullshit stats. By lumping lesbians in with gay men and transwomen, it's possible to both marginalise them whilst taking up their part of the LGBT quota and STILL claim to be under represented.

All of these ways of collecting data are actively working against women and other minority groups because of who is setting the questions and the way they are setting the questions.

In an era of big data this matters more and more.

At this point this thread really deserves Invisible Women by Caroline Criado Perez to be plugged and recommended for anyone who hasn't read it yet. It's essential reading.

FWR is definitely picking up on some many of these type of issues when institutions are turning a blind eye to the problem either by blindness or deliberate design.