Statistics are political

Today is the New Zealand Census. And if that comes as a surprise to you, your surprise is not a surprise to me. If that makes sense.

I’m obviously a political nerd, and have heard almost nothing about the census. Officially, at least. We’ve had the one letter to our household with the access code (which was … not particularly well written), and there was a Stats NZ stall at the Newtown Festival on the weekend, but besides that, I’ve only heard from people having trouble with it.

People who haven’t received access codes, and have been on the phone to Stats NZ multiple times trying to get one – which can’t be authorised by a first-tier support person, it has to be bumped up to a supervisor, and which can only be sent by snail mail, not email, or maybe it can be sent by email, but it takes a week? Nobody seems to know.

(Let’s take a pause here to remember that New Zealand Post, under the guidance of awesome progressive statesman Sir Michael Cullen, shut half its processing centres and halved the number of deliveries throughout the country, which is why it takes so long to get mail from Wellington to … Wellington.)

People who will not be able to fill out the census because Statistics NZ, in its pursuit of cutting costs and pushing everyone online, has completely screwed up the process for people who are blind.

I suppose I have the option of being out of New Zealand on 6 March. Maybe a quick trip to Aussie is in my future, so I can boycott the Census without breaking the law. It’s a sad situation where I must contemplate being a refugee from my own country to make a critical point about the need for Government to own its responsibilities.

And then there’s the perennial issue, the issue Stats NZ have been aware of since well before the previous census: its complete erasure of trans, genderfluid and any other people who don’t fit neatly into two boxes marked “male” and “female”.

It feels like Stats NZ has not only dropped the ball on sex, gender, orientation and identity, it’s kicked the ball as hard as it could and shrugged as it went over a fence. Its own justification for the lack of questions demonstrates that it’s incredibly easy to explain the difference between biological sex, gender identity and sexual orientation. But they won’t. The very scientific and unbiased conclusion they’ve come to is “well, it’s too hard and there aren’t many of them anyway, so who cares?”

I care. I care about marginalized groups being represented and included in the official records of our society. I care about the huge amount of work done by gender diverse people and groups who have tried so damn hard to help Statistics NZ get this right, and who have been ignored for years because some platonic notion of statistical propriety trumps their existence and welfare.

I care that it took four years for our world-leading source-of-much-nationalist-smugness health system to find one surgeon who can provide gender confirmation surgery and wonder if maybe that would be different if we didn’t erase trans people from our premiere official population count.

I care about the message that is sent when the Minister for Statistics blames people’s “silly answers” in test surveys for the lack of decent gender and sexuality questions – yet “Pastafarian” – a parody specifically created to mock our attitudes towards organised religion – is not only permitted but gets its own autocomplete suggestion on the online form.

Screenshot from Twitter.

I also care about the assumptions our society makes about sex and gender and childbearing. At the last census I recall, but cannot confirm (damn Google) that Stats NZ stated they “impute” people’s sex based on their answers to a range of questions – so if you ticked male but said you’d given birth to two children, they would “correct” your sex to female. This is obviously completely unproblematic and hunky-dory as long as it makes the data ~clean~.

At this stage, there’s an outstanding OIA request on asking if Stats NZ continues this practice. It will clearly not be answered before the census closes today.

It simply irks me that for all the head-patting and condescension the LGBTQI+ community gets from Stats NZ, and the hand-wringing over ~reliable data~, the census is a shambles anyway. They’ve even re-jigged the meshblocks so apparently it’s going to be hell for anyone doing long-term research to match this census to previous ones (edit: though I’ve also heard they recalculate previous census data to match the new meshblocks). A couple of additional questions to shine light on an underrepresented and marginalized community whose health, legal and social needs are often ignored and diminished was hardly going to ruin everything for the data nerds.

This is something we should all care about. Not just because of the particular concerns around gender, sex, accessibility, and a government department’s ability to get the basic logistics of its primary job right.

Statistics are political. Data are political. They do not exist in a vacuum, because they are shaped by human perceptions and decisions from day one. Decisions about who should count, or what should count – religion but not political ideology; (assumed female fertility) but not sexual orientation – are political because they have political impacts.

A minor but illustrative point: the census includes motorcycles and scooters under “Other”, not “Motor vehicle”, for commuting options. What about a scooter makes it not a motor vehicle? How does this data reflect assumptions about how people commute, or should commute, or want to commute? How does this reinforce our preconceptions around policy to reduce congestion or address climate change?

I’m sure there’s a reason. I’m sure there’s academic papers and statistical standards and longstanding taxonomic principles in play. And every single one of those exists within a simple context: our society has been car-obsessed for a century with all the consequences for urban design, social behaviours and infrastructure spending.

It’s very nice to think you’re able to sit above the world everyone else lives in and observe it like a wise man atop a mountain, but it’s bullshit.

(This is also why groups (*cough*TOP*cough*) who try to claim neutrality or objectivity, because their policies are “evidence-based”, are either lying or kidding themselves (*cough*TOP*cough*). The only thing more dangerous than having biased data is having biased data and insisting it is not biased. If we willfully ignore the role of unconscious bias and attitudes in shaping the data we collect, decisions justified by that data will only harm people.)

It does make it wonderfully easy to ignore the existence of the queer community and then say “oh no, we made that decision on entirely statistical grounds” though.

The problem is, census planning takes a hell of a long time. The work putting together the 2023 census is likely already well underway, and making the kind of significant shift that’s clearly required – producing modern, relevant information with modern, relevant and inclusive processes – may simply be beyond the capabilities of a department which couldn’t even take decisive action moving its staff out of a lethally unsafe building.

6 Replies to “Statistics are political”

  1. The previous National government raised mean spirited penny pinching to a fetish and actively repressed the collating of information that may lead to inconvenient facts being made public. Having canceled one census, is it a surprise they deliberately underfunded this one so we all we are going to get is a half-arsed, badly promoted, badly funded online survey of the digitally literate middle class?

  2. I did notice that they now had questions on whether your dwelling had mould or damp: couldn’t recall those being in before. Also noticed the lack of choice in the gender question. But one important concern is how they’re going to ensure that most people fill the forms in, now that no-one comes to visit. It’s also less likely that those without easy, or any, digital access, will bother ensuring they get a paper form.

  3. “The only thing more dangerous than having biased data is having biased data and insisting it is not biased. If we willfully ignore the role of unconscious bias and attitudes in shaping the data we collect, decisions justified by that data will only harm people.”

    YES! YES!

  4. The “post a code” thing is also going to underestimate and miscount homeless people in a really nasty way. And a few of them will inevitably end up getting fined for not being at their last known address.

    In Australia with the plebeshite we also had problems with whoever collected the post filling in the forms “on behalf” of people whose politics they disagreed with. I suspect that will happen less with the census, but I suspect a few people will have their answers “corrected” by their parents.

What do you reckon?

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s