It’s been a busy old week for trans politics in Scotland. The Court of Session ruled against the challenge to the official guidance given for the sex question on the national census, If you are transgender the answer you give can be different from what is on your birth certificate, in part it seems by interpreting the word can as meaning only that it is physically possible to do so.

Meanwhile, the court of appeal ruled that the Scottish government had exceeded its powers in single-handedly redefining sex, for the purposes of its gender representation act, which said that the boards of public bodies had to be at least 50% women or men in skirts, which seems meaningless, given that it’s Scotland we’re talking about.

The results were also published of a major study, undertaken by Savanta: ComRes on behalf of the BBC, into attitudes towards aspects of the trans debate in Scotland.

This was a sizeable study, with 2,038 online responses gathered. These were random respondents, from Savanta’s panel, and not a self-selecting sample, so we should expect them to be reasonably representative.

As required by British Polling Council rules, the full set of data tables have been published and can be downloaded here

It’s routine for published tables to exclude the initial questions, which determine basic demographics, which means that the first nugget of information has to be wheedled out indirectly. Because the 2nd question is only asked to ‘cisgender respondents’ we know that 7 of the people who took part identified as being trans. If you don’t have your calculator handy, that means that from a reasonably robust sample of the population, 0.34% of respondents were trans.

If you’re interested, 2 of the 7 were aged 16-24, a further 2 aged 25-34, and 1 each in the 35-44 and 45-54 brackets and 1 aged 65 or over, 6 of the 7 were white, 3 were from white-collar households and 4 from either unskilled or economically inactive households, and none of them voted Lib Dem in the 2021 Holyrood election. Make of that what you will.

Here’s where it would have been interesting to know about those demographic questions, because throughout the data tables the male and female breakdowns add up to 2,031. This either means that respondents were asked separately if they were trans or not and those who answered ‘Yes’ weren’t counted as either male or female, or, more likely, a standard market research gender question was asked, with the usual options of male, female, non-binary/other, and only those who selected the 3rd option were counted as trans.

This isn’t a huge error but it’s annoying that more care wasn’t taken. I imagine any trans respondents were annoyed to be asked the 2nd question, from which only non-binary (not cis) people were excluded, asking them if they knew any trans people.

On the plus side, the first published question tackles one of the main issues any survey on this topic faces, and asks the respondent what they think transgender means. The options range from self-identification, They express their gender differently to the sex they were registered as at birth, through to the more old-school meaning of transsexual, They have had gender reassignment surgery to permanently transition from their sex registered at birth. There are two intermediate options; they’ve had a medical diagnosis of gender dysphoria or they’ve obtained a gender recognition certificate, which unsurprisingly proved popular with no-one. There were also options to say you didn’t know (11% of respondents) or you’d prefer not to say (just 27 people were too bashful to answer).

I think the two sides of the gender debate largely agree that you are trans if you say you are (while, of course, violently disagreeing about what that means), so it’s fair to say that the self-identification answer is the correct one, and was the one chosen by 51% of respondents.

This means that if you read any survey about this issue that doesn’t define terms then half of the respondents won’t correctly understand the questions. That’s important.

This survey does, at the start of the next question, clarify that transgender is ‘a general term for people whose gender identity is different from the sex registered at birth‘, so we at least know for future questions that respondents were steered along the right path. This, however, makes it annoying that those of us reading the data are denied some important information before a key question, question 4, which determines the level of support for making the process of obtaining a gender recognition certificate easier.

Fifty-seven percent of respondents support making the process easier, with just 20% opposing such a move (18% have no opinion and 5% don’t know), but the question opens with ‘Given this[sic] information on the previous pages’…without telling us what the information on the previous pages was. This is, in my opinion, a more serious oversight than the mix-up with the gender question. BPC rules demand that questions are asked in a neutral fashion and must be published in full, but we are here effectively being denied the context of the question.

The reason this is interesting is that the follow-up questions ask about specific reforms of the process. A plurality of respondents (44% vs 37%) oppose cutting the time applicants must prove they have lived in their acquired gender from 2 years to 6 months, removing the requirement for a medical diagnoses enjoys only marginal support (40% vs 38%), and reducing the legal minimum age for application from 18 to 16 is opposed by the majority of respondents (53% vs 31%). The only place where support is overwhelming is in favour of making it a criminal offence to make a false declaration when applying for a gender recognition certificate (60% support, 16% oppose). While this new criminal offence is the key, indeed the only, safeguard for the Scottish government’s planned reform of the GRC process, it’s hardly the gender critical position to oppose this.

Hence we find ourselves in the strange position where vastly more people support making the GRC process easier than oppose doing so, but nobody is actually strongly in favour of any of the particular methods of making it easier that have been proposed. This is generally the point where those of us who work with survey data all the time go for a smoke break and shake our fists at the sky, yelling, “What do you mean, you bastards???”

With my word-count in mind, this is a decent length questionnaire and, so far, we’ve made it to question 5, so I’m not going to go through step-by-step, but I did want to touch on a final issue. It’s long been the assertion of the trans-rights side of the argument that the majority of women support them, and looking down the Male/Female column throughout these data tables you’ll consistently see more women than men supporting the trans-rights side of the argument.

However, this is a reasonably well put-together survey, asking reasonable questions, and allows us to dig a little deeper. Question 7a, for example, asks whether the respondent agrees that it’s important for legislation to continue to provide for single-sex spaces, such as hospital wards or changing rooms. Here 64% of women agree, compared to only 59% of men. Disagreement is just 8% amongst women. Unfortunately, we can’t break this down further by both gender and age, but we can see that even in the youngest age group (16-24) agreement still wins hands down (47% vs 15%)

The follow-up question, on whether transwomen should be able to use women’s toilets seems to contradict this, with the plurality of women (45%) saying that they should be able to, compared to only 20% who say they should not.

The issue with this question is that it offers only 4 options. Either transwomen are allowed to use female toilets, they are not, you have no opinion, or you don’t know. There is no nuance and once you, not unreasonably, remove the no opinions and the don’t knows you have a super-majority of women (70%) in favour of TW being allowed to use female toilets. Expect to see that number a lot in future arguments.

How fortunate, then, that at question 12, a little more detail is obtained. Question 12 asks when transwomen should be able to use single sex spaces, such as toilets and changing rooms. Here the options are either by simply identifying as a woman, by going through a process to legally change their sex, by legally changing their sex and having sex-change surgery, not at all, or don’t know.

Here the numbers are much more finely balanced. Both overall and looking at female respondents only, a minority (28% and 36%, respectively) say that transwomen who have not surgically changed sex should be allowed to enter female spaces. Excluding the don’t knows splits female respondents almost evenly, with 43% saying that uncut transwomen should be allowed in and another 43% saying only after surgery, and 14% saying, no, not at all.

I suspect we’re going to see very carefully selected portions of this poll shared by both sides for quite some time to come, and it is fascinating. It would be great to see an anonymised data set released, to allow some proper drilling down into the data, but I hope at least it shows others polling in the same field that there’s better ways than asking pointless questions, like “Do you agree that transwomen are women?”

Just annoying that a poll that got so much right when asking about gender managed to mess up the question on gender. Let’s hope others do better.