Yvonne van Dongen
Veteran NZ journo incredulous gender ideology escaped the lab. Won’t rest until reality makes a comeback.
If nothing else, being a TERF helps sort ideology from reality in all sorts of spheres.
As soon as I saw the word transphobic in last week’s story on sexist and racist abuse in politics courtesy of a report by an outfit called Areto Labs, I knew we’d strayed into the realm of ideology.
Predictably, their analysis of 150,000 social media comments between August and November showed that a third of comments to political candidates were abusive and that women got what they called “the sharper end”. The co-founder of Areto Labs, Jacqueline Comer, concluded that women are paying more of the price of standing for office.
Cue laments by female Green and Labour MPs about the impact on women’s mental health and the risk this might lead to fewer women in the House, not more, if this went on. As it stands there will be 57 per cent male MPs, 43 per cent women in parliament this time, much the same ratio as the previous parliament.
RNZ and the New Zealand Herald stories highlighted the fact that women received more “gender-related” abuse than men and that included “sexist, transphobic and homophobic comments, gender micro-aggressions and body shaming”.
Given that the transphobic remarks are likely to have included comments by sex realists pointing out that there’s no such thing as being born in the wrong body and please support the Definition of Sex Bill, we can discard that criticism. Such language also alerts the reader to the likelihood that those preparing the report are progressive ideologues, aka not serious people.
Let’s have a look at the company doing the report and also the data. Areto Labs is a Canadian women-founded artificial intelligence (AI) content moderation company with New Zealand ties via its co-founder Jacqueline Comer. For this study candidates self-selected to make use of the moderation platform. Twenty-four women and nine men chose to do that, from five political parties.
That means that, right from the get-go, the study is skewed. The accounts are not a random or representative sample of New Zealand candidates. Given that the data came from an Areto tool that offered “free protection primarily to women and gender-diverse candidates,” it’s obvious that only people who accept this language would use it. In other words, other progressive ideologues. The participants disproportionately feature women and “gender-diverse” (whatever they are) participants at roughly 70 to 75 per cent of the sample.
So candidates who signed up for this study are likely to be those individuals who are more concerned about or already experiencing abuse, and that abuse might just include frustration at their unwillingness to accept reality, i.e., only two sexes and you can’t change sex.
Only nine men took part in the study so that severely limits any comparison on the basis of sex.
Also Areto is a commercial vendor selling the moderation product used to generate the data. The report effectively showcased demand for their tool. This is not independent research.
The data is, in fact, no more than a dump from a vendor’s dashboard. It is not rigorous social science. Findings cannot be generalised to all New Zealand candidates, all social media activity around the election or non-users of the tool. They do not capture offline threats or the full ecosystem of abuse.
The above was all my own work. I then asked AI for its assessment and it listed other weaknesses such as:
“no public detailed methodology paper, validation metrics (precision/recall of the AL classifiers, especially for “extreme” micro-aggressions or New Zealand specific language/context), inter-rater reliability or error rates. Thresholds for “abusive” or “extreme” were set by each candidate (or the tool’s default) so they are not uniform. One comment can trigger multiple flags. Sentiment and category definitions rely on proprietary AI who perform and on political New Zealand English (including passive-aggressive or local idioms) is not independently audited here.”
To be honest, I didn’t understand much of that hence the direct quote. AI also criticised the fact that there were no statistical controls or causal claims.
“Results are purely descriptive. There is no adjustment for profile size, posting frequency, political party/alignment, ethnicity visibility or engagement volume, Difference between men and women could partly reflect these factors rather than gender alone. No confidence intervals, significance tests or robustness checks are mentioned.”
At best I suppose you could conclude that the study flags an issue worth investigating later but its limitations constrain how seriously you can take it.
This bothers me since their conclusions broadcast by the media, that women experience more severe abuse, affirm common tropes, which may or may not be true.
Sometimes I think we are often fighting yesterday’s battles. For instance, trans activists frame their campaign as a trans rights issue, echoing the gay rights movement. We know that it’s nothing of the sort. It uses this framing to steal women’s rights but people who don’t look too closely assume it is a battle for basic human rights, just as gay rights was.
Without proof, the assumption that female political candidates suffer more than male candidates is just that – an assumption. Some of it might even be a strange sort of reverse sexism. When ZB host Mike Hosking grilled the then Prime Minister Jacinda Ardern on air, his voice dripping with contempt, I was incensed and understood her decision not to continue being interviewed by him.
But I’ve since heard numerous hosts treat PM Chris Luxon with similar contempt, and although I’m not a fan, I’m also not as triggered.
I wonder now if I felt more upset by Hosking’s treatment of Ardern because she’s a woman. It may be that I believe women should be treated with more respect while men in public life deserve what they get. Perhaps we are all more protective of women. Or perhaps women whose politics we agree with. The over-reaction to NZF candidate and Platform host Michael Laws’ comments about a female Opportunity Party candidate could come from the same place. It’s okay to make jokes at men’s expense, but not women’s.
In 2017 when sculptor Sam Mahon made a huge work made from horse dung showing politician Nick Smith with his pants down, defecating into a glass of water, to highlight his views on the National government’s freshwater standards, I thought it was funny. Would I have thought it funny if he’d featured a woman? Probably not, because women are especially vulnerable in the bathroom, hence the fight to maintain single sex spaces. Nevertheless the work was offensive.
What about the 2023 billsticker flyers around Wellington depicting Christopher Luxon, ACT leader David Seymour, and NZ First leader Winston Peters superimposed inside a drawing of male genitalia (with Luxon forming the head and Seymour and Peters in the scrotum)? The Wellington City Council ordered the posters removed, and the woman responsible was later formally counselled following 40 public complaints. But imagine if something similar had featured women. Would we think it funny or offensive? Acceptable or not?
Many assumptions do not bear close examination. It is often said that professions heavily populated by women suffer because women are generally paid less than men. It is assumed that employers are able to exploit women in a way they can’t exploit men. But the legal profession now has more women than men. Has anyone noticed legal services now cost less? Same with the medical profession and vets. More women than men in both, but, in my experience, definitely not cheaper.
That’s why the Areto Labs research is so disappointing. It may be that women in politics suffer more savage social media abuse than men but crummy self-reported surveys like this do not prove a thing.
This article was originally published on the author’s Substack.