Invisible Women: Data Bias in a World Designed for Men

A brief review of Invisible Women (2019) by Caroline Criado Perez

I learned about Invisible Women by Caroline Criado Perez while listening to the podcast 99% Invisible, an excellent podcast all about design. Roman Mars, the founder and host of the podcast, interviewed Perez about her new book. The episode focused on the example from a Swedish town where research shows that snow plowing patterns are gender-biased against women. Seems hard to believe, right? Give this episode a listen to have your mind blown.

Apart from being an excellent attention-grabber, the snow plowing case study captures the prevalent theme of the book: many (if not most or all) of our designs, policies, language conventions, historical records, and more are biased against women. Some are intentional (keeping women out of positions of political power), some are not (snow plowing). Regardless of the intention, Perez compiles a tremendous amount of research to demonstrate that these biases exist in every sphere of life, that there is data to prove it, and that in the age of big data we are furthering this gap into a gender data gap.

Perez begins by talking about the “default man”, the median male that pops into so many people’s minds when they picture “someone”, “a scientist”, “a president”; the 50th percentile dude who fills the rosters of clinical studies; the person car safety ratings are based on. The default man pervades our literature, our language, our designs, and our thoughts. It terrifies me that I have this in my head too, and now actively attempt to filter my thoughts to change it. We’ll see if it takes.

Building upon this pervasiveness, Perez goes through study upon study demonstrating gender bias in the realms of daily life, the workplace, design, at the doctor’s office, public life, and when disasters strike. Gender bias is exemplified by hiring processes where those in charge tend to hire more men despite equally qualified candidates of all genders. A less obvious example is found in the process of getting tenure. The prime years for the tenure-track overlap with the age range when many women have children. In a seemingly objective set of criteria, this overlap hinders qualified women from gaining tenure (though the criteria are gender biased in other ways as well).

The warning Perez gives us is that in our age of information and big data, this gender bias is building into a gender data gap. If we (read: men in positions of power) ignore problems impacting half the population, we won’t measure them, and we won’t have data to base decisions on. There is a tendency to assume that algorithms and data are inherently unbiased, but that is manifestly untrue (check out Cathy O’Neil’s Weapons of Math Destruction to learn more about that). If we inject bias into what data we collect and we don’t collect sex disaggregated data, we can’t make good decisions for how to size smartphones, what policies to pass, and how disease may manifest differently in women. The data gender gap is real, and it’s growing.

Women have been made invisible in history, policy, and the economy (women perform the majority of unpaid work, particularly unpaid care work). This is obviously not news to many, but I am thankful to have learned far more about this than I otherwise would have. I suppose I can only really speak for myself, but we need to change how we speak, how we think, how we vote, and how we make decisions for the sake of literally half our world.

tl;dr: Invisible Women is excellently written and researched. Everyone should check it out from the library and read it, particularly those who work in design, policy, and data. See if you have any biases and do what you can to address them.