Join our community of smart investors

Further Reading: Open your eyes to data biases and gaps

Data failings have rendered women’s needs largely invisible in everything from medical care to urban planning
November 26, 2020
  • Data biases and gaps means women's needs are not well considered
  • Male-biased policies can have serious consequences for women
  • Companies and investors can learn from this

Data matters. Whether in its simplest form or as information extracted from millions of transactions, data is a valuable tool. It enables governments, individuals and investors to make decisions, and businesses to gain a competitive edge.

But to serve those purposes, data must be good quality and free from gaps and biases. You might suppose that these would, in this day and age, be meticulously rooted out, but that is not the case. In her book Invisible Women, Caroline Criado Perez builds a thick wall of evidence proving that not only do data gaps and biases exist, they spawn calamitous outcomes.

Prompted by her observations that much of the world seems to work better for men than it does for women – think toilet queues (non-existent for men, long for women), pension pots (bigger for men, smaller for women), or voice-activated technology (responsive to men, deaf to women), Ms Criado Perez set about exposing the long-standing and systematic ignoring of women in data.

It's an exclusion that has happened more by accident than design and it stems from the acceptance of men as the norm, and women as a sub-set, or perhaps sub-standard variant. Men have been used as a universal model for centuries, typically forming the majority – and often the entire sample – of representatives in studies. For example, a quarter of drug manufacturers do not use women in drug trials despite evidence that most drug dosages need to be altered for women. If women are included, they often form only a small fraction of the sample. A 2016 paper found that females made up 19 per cent of participants in US studies for new HIV treatments, and 11 per cent of studies into finding a cure. Elsewhere, trials of a new cardiac therapy had such low female participation the data failed to reveal that it did not work for women. Women, says the exasperated author, are not simply “smaller men”.

These data gaps are compounded by the lack of female representation at the policy-making stage. Women going to the theatre know they will face a long queue for the loo and the cause of this is a failure to consider women, according to Ms Criado Perez. Traditionally the same floorspace is given to men’s and women’s toilets. But when cubicles in the men’s are included, their capacity is double that of women’s. Gender-based data would have revealed that women are more likely to have children with them, more likely to be elderly (there are more older women than men) and therefore slower. Plus due to anatomical differences women take longer to use the facilities. So equal floor space does not create an equal outcome. 

Toilet queues are annoying, but some data biases are life-threatening. Crash-test dummies have almost exclusively been based on the male body. Proper data is not normally gathered on crash impacts on passengers, who are more often women. These gaps in data driving car safety measures mean that as a woman you are almost 50 per cent more likely than a man to be seriously injured in a crash.

In urban planning, the fact that women are often scared in public spaces at night has not been factored into planning. Modern data has revealed that twice as many women are afraid when they walk through multi-storey car parks as men, and far greater numbers of women worry about walking home from a bus stop or station. So women spend money (on taxis) and time (taking a different route) to avoid these situations. “But all too often,” says Ms Perez, “the blame is put on women for feeling fearful rather than on planners for designing urban spaces and transit environments that make them feel unsafe.”

Default male thinking hurts women’s finances, too. Pension systems are built around the number of years that we work, and earnings. This works well for men, less so for women who take on unpaid caring roles. And when they do rejoin the workforce the rules work against them. Women are more likely to have more than one part-time job, but even if the income from these means they should qualify for auto-enrolment, they can't enrol because their income is from multiple employers. Around 32 per cent of employed women do not earn enough to benefit from auto-enrolment, compared with only 14 per cent of employed men. Meanwhile, in the US, where married couples are encouraged to file joint returns, women lose out because the income of the lowest earner (almost always the female) is added to that of the highest earner, with the result that her smaller income is taxed at the highest rate. 

There are lessons for investors and businesses in Ms Criado Perez’s depressing tales of biased data. You might think you have all the information necessary, but perhaps you just haven’t spotted the holes, or biases in your own thinking. When one excited sports reporter congratulated Andy Murray on being the first person ever to win two Olympic tennis gold medals, the tennis star replied that he wasn’t. “Venus and Serena,” he said, “have won about four each.”

Biases and gaps can sully product development and cost you market share. A new tech product that's perfect for men but doesn’t work for women means lost sales. Much better to follow the example of the VR headset manufacturer who factored in the possibility that the user might be wearing mascara. The reason for this enlightened approach: the founder of the company was female.