Think.

  • Mon, 9, Nov, 2020 - 5:00:AM

Data feminism: confronting systemic biases

The pursuit of science is one of the most noble there is.

That is coming from a woman studying Physics, from someone who loves being methodical, sceptical and fair. All the qualities that the study of science drills into you. But the more I foray into the field, the more I see that science and all it represents can be and has been tainted.

Science is not objective, much as we want it to be. “Data is never a raw, truthful input – and it is never neutral,” says Catherine D’Ignazio, author of Data Feminism, a book pushing for equitable science.

Western science has been dominated by patriarchy since its inception. In the era of Hippocrates, wandering womb syndrome was the blanket diagnosis for a woman experiencing ailments of any kind. Supposedly, the uterus detached itself and went walkabout the female body. This lethal misunderstanding gave rise to the idea of female hysteria and its prescription: lobotomies.

Sigmund Freud made a lot of bold claims and contributions to psychoanalysis. But his method was based almost entirely on case-by-case study, without research and control groups. Freud selectively studied boys and made conclusions based on the psychology and development of boys. His theories have largely been written off as hugely gendered and unscientific.

So there’s the past, with its many cases of systemic biases against women, LGBTQ+ communities and people of colour in science. And now there’s the present, which is teeming with the harmful legacies of science under patriarchy.

As a result of gender unequal research, a lot of people don’t know what symptoms of heart attacks look like in women. A study found that on average women wait 37 minutes longer to contact medical services than men do, because we aren’t aware that heart attack symptoms often present themselves differently for women.

An unexpected platform of education and awareness has come from Tiktok. As with heart disease, endometriosis, ADD and ADHD, autism in women is underdiagnosed and under researched. Autistic women have taken to social media to share their experiences with autism, being diagnosed and not fitting the ‘mold’ of autism.

Data feminism and data intersectionality is becoming increasingly relevant as societies move towards using AI systems in decision-making processes. The idea is that human error can be mitigated by training unbiased computer programs to do the job. But the errored humans we don’t want prosecuting in the courtrooms, giving bail or determining its prices, and removing children from risky homes are the same ones who built the AI.

As D’Ignazio said, data is not the truth. Data is the input. If our inputs are flawed, the outputs will be flawed, too. I like to think of it as a mirror. If people are worried about AI systems spiralling into racism, sexism and overthrowing their mortal overlords, why don’t we take a look at who they learnt from?

Predictive policing systems, like PredPol in the US, disproportionately target people of colour. Child protection systems disproportionately flag poor homes as higher risk. Studies into AI revealed that systems grouped “woman” and “female” with humanities and arts occupations and “man” and “male” with maths and engineering occupations. Just like we do.

Hopefully these case studies have given you some more tangible reasons behind the push for more women in STEM. It has certainly become somewhat of a buzzword, but it’s not merely achieving gender equality for equality’s sake. It’s still a matter of life and death, of healthcare, of families separated or families reunited.

D’Ignazio recommends university data science programmes including multiple mandatory ethics courses. Diversifying science with more women, LGBTQ+ communities and people of colour will make it more likely that data sets are more inclusive and representative. They can see the issues that are invisible to a white, male gaze. 

And it’s not enough to include marginalised groups in science; we should place science in the hands of marginalised groups. “We need to put communities who will be impacted by the information systems into the process of making them.”

See, I really do love the pursuit of science.

But I’ve since discovered the pursuit of intersectional science, and it’s much, much better.

TAGGED IN

  • Women in STEM /
  • Women in Tech /
  • Systemic Oppression /
  • data /
  • Data Science /
Support Villainesse

Comments ( 0 )

Be the first to have your say login or register to post a comment

You might also love

Aimee
Lew

Regular Contributor All Articles