NEWS11 March 2020

Caroline Criado-Perez: ‘There is an urgent need for sex-disaggregated data’

AI Healthcare Impact 2020 Media News Public Sector Technology Travel UK

UK – Diversity is essential for research, as biased data has serious implications for how resources are allocated and products and services are designed, said Caroline Criado-Perez at the Impact 2020 conference.

Caroline C-P impact 2020_crop

Discussing the research she undertook when writing her book, Invisible Women: Exposing Data Bias in a World Designed for Men, Criado-Perez highlighted the extent to which data does not reflect women’s experiences and called on researchers and those collecting data to ensure it is sex-disaggregated, i.e. specific to women.

Data reflects those who collect it, she said, explaining why she doesn’t think describing data as “the new oil” or like water or blood are helpful. “My issue with these analogies as they all present data as this tangible substance with inherent properties that exists in the ether ready for us to extract, but data us more than just numbers – data is us and we are data.”

She shared two examples of data sources that had not sex-disaggregated the data – labour force statistics, which historically undercount the number of women in paid employment – and public transport data.

“The data was skewed to the detriment of women, and this is not a coincidence – most datasets share this. The vast amount of data we collect has mainly been collected on men. Pretty much everything you encounter in the world – the offices we work in, public transport, medical treatment, the phones we use and the apps on them, have been designed to work for men – so just don’t work that well for women.”

This is “not a conspiracy”, argued Criado-Perez, but instead it is “the result of a way of thinking we all share”.

Dating back to Aristotle’s assertion of women as a mutilated version of men, historically women have been viewed and represented as deviating from the male standard. The idea of the default male is dangerous, noted Criado-Perez, because it leads to large gaps in knowledge and understanding, with implications for everything from how life-saving medicines are developed to how public transport networks are designed.

“There is an urgent need to collect sex-disaggregated data – not doing so is not a deliberate failure, but without any sort of knowledge of how gender might affect [the subject of the research].”

However, when women take part in the data collection process, evidence has shown that data is less biased. “When women are involved in research, the research that is produced is more likely to have sex-disaggregated data collection,” said Criado-Perez. “That doesn’t mean it’s perfect – because women are also subject to these biases – but they’re less likely to forget that women exist than men.”

She also flagged research on how people tend to perceive gender-neutral terms, including words like participant, user and researcher – as male, regardless of our gender.

Lack of inclusion of diverse experiences and perspectives results in major blind spots for researchers, designers and developers. For example, Criado-Perez talked about Apple’s health tracking app that couldn’t be used to track periods. “It’s not credible to me that Apple deliberately set out to exclude women – they just forgot periods happen, because they didn’t have enough people who have periods on their design team,” she said.

Criado-Perez, whose activism led to the inclusion of Jane Austen as the first woman on a Bank of England note, said she was “furious and upset” when writing the book. “I was terrified I wouldn’t be listened to.”

However, since it was published, she has felt vindicated by the response of individuals and organisations contacting her to say that they have changed their approach to research and running businesses. “I don't think this is about bad people doing bad things. This is about people in bad systems and most people have no idea this is going on. When consciousness is raised, it’s shocking, and people want to fix it.

“For me, it’s about how to frame the issues in such a way that people don’t just feel defensive or hopeless. Data is not neutral, but it’s not personal. It’s so clear that this is about the way we’re counting, collecting and analysing data, and those things can be changed.”