Invisible Women: Data Bias in a World Designed for Men

By Caroline Criado Perez New York, NY: Abrams Inc., 2019. 432 pages. $27 hardcover, ISBN: 978-1-4197-2907-2

Invisible Women is an extensively-researched new monograph by Caroline Criado Perez on gender data gaps, far-reaching gaps in data collection practices across domains, industries, and geographies that render the experience of women invisible and subsumed in the “male default.” A feminist activist and journalist, Criado Perez sets out an ambitious program for the book, which she describes as “an exposé of how the gender data gap harms women when life proceeds, more or less as normal…Invisible Women is a call for change…It’s time for a change in perspective. It’s time for women to be seen.” (p. 25). This timely and compelling book offers an impressive array of empirical cases which demonstrate the prevalence of gender data gaps – and the harm they make possible.

Organized into six themes, the book’s chapters deal with an extensive range of women’s daily experiences and the impact of gender data gaps in each. These themes cover women’s experiences: in public spaces; in the workplace; through the design of tools and artifacts; through medical encounters; in political representation; and in the aftermath of crises and natural disasters. The overall breadth of Criado Perez’s program offers sweeping coverage rare for a single project – her book shows gender data gaps are pervasive, relentless, and dangerous. Written with a public audience in mind, this book offers value to a range of information studies scholars – generalists will receive a comprehensive and systematic introduction to empirical issues of gender and data bias; specialists in the area of data studies will find new insights from the book’s wide-reaching and thoroughly-referenced range of empirical topics. Curious how gender data gaps impact the design of Unicode emojis? The planning of snowplowing routes and schedules? The design and testing of safety gear for military and law enforcement? The regulation of chronic chemical exposure in workplaces? Bias in the outputs of natural language processing models in automated language translation? This short list is just a taste of the topics included in this book’s remarkable range.

As Criado Perez lays out each of these themes in the book, readers are transported across a range of domains and geographies, covering not only a number of scholarly perspectives (from disciplines such as urban planning, architecture, medicine, computer science, and public policy, to name a few), but also a variety of points-of-view. Written in a fresh and energetic style, this book offers accessible prose that weaves together scholarly publications, public statistics, and perspectives from domain experts and researchers, alongside vignettes and personal accounts of issues faced by everyday women (often through the recounting of journalistic pieces or the author’s own experiences as a woman). This writing style can itself be read as a feminist intervention–providing lively insights into data bias and how it can be both examined as a scholarly topic and felt as an everyday, lived-through social problem.

Criado Perez’s overall argument is tripartite, which she lays out as three key themes in the book’s afterword. The first is that women’s bodies are consistently rendered invisible through gender data gaps – such invisibility results in a built world that is the largely ill-fitting, and even deadly for women. The second is that gender data gaps make possible the persistence of sexual violence against women: a failure to adequately measure sexual violence against women means we aren’t taking action to eliminate such violence. The third is the problem in how contemporary data collection practices fail to appropriately account for the unpaid care labor of women – this diminishes the social infrastructure such labor provides and its critical role in the smooth, everyday running of modern societies. Ultimately, Criado Perez’s thesis is that sound and precise data collection practices are the springboard necessary to instigate social change, as she writes: “In order to design interventions that actually help women, first we need the data.” (p. 150).

The book delivers on its goal to demonstrate the pervasiveness of gender data gaps and their consequences on women’s lives, yet its overarching argument (that more data are the answer) remains an open question for critical data scholars. The book offers little conceptual engagement with data as representational forms and the limitations of such forms, which feels like a missed opportunity to introduce the book’s wide public audience to critical studies of information (though, the book does briefly mention the need to take care in survey design, noting how poorly-worded surveys can skew results, for example, in the reporting of women’s labor in- and out-of-the-home). The book also seems to adopt a somewhat essentialist and binary view of women’s existence, perhaps a gloss needed in delivering on its laser-focused mission to convince those who might be skeptical that gender data gaps systemically and ubiquitously erase women from public imagination. The book does well in showing how gender data gaps exist in many corners of everyday life and would nicely add a broad, empirical complement to syllabi or reading shelves that also include conceptually-focused and theoretical pieces on this topic, for example, the works of Geoffrey Bowker, Alain Desrosières, Lisa Gitelman, Anna Lauren Hoffmann, or Safiya Umoja Noble.

Christine T. Wolf, IBM Research–Almaden