Chapter Three: “What Gets Counted Counts”

^{This chapter is a draft. The final version of Data Feminism will be published by the MIT Press in 2019. Please email Catherine and/or Lauren for permission to cite this manuscript draft.}

“Sign in or create an account to continue.” These may be the most unwelcome words on the internet. For most who encounter them, these words elicit a groan– and the inevitability of yet another password that will soon be forgotten. But for people like Maria Munir, the British college student who famously came out as non-binary to (then) President Barack Obama on live TV, the prospect of creating a new user account is more than an annoyance. “I wince as I'm forced to choose female over male every single time, because that's what my passport says, and... being non-binary is still not legally recognised in the UK,” Munir explains. 

For the estimated 9 to 12 million non-binary people in the world-- that is, people who are not either male or female-- the seemingly simple request to “select gender” can be difficult to answer, if it can be answered at all. Yet when creating an online user account, not to mention applying for a passport, the choice between “male” or “female,” and only “male” or “female,” are almost always the only options. These options (or the lack thereof) have consequences, as Munir clearly states: “If you refuse to register non-binary people like me with birth certificates, and exclude us in everything from creating bank accounts to signing up for mailing lists, you do not have the right to turn around and say that there are not enough of us to warrant change. 

“What gets counted counts,” as feminist geographer Joni Seager has asserted, and Munir is one person who understands that. Without the right categories, the right data can’t be collected. And increasingly, without the right data, there can be no social change. We live in a world in which “data-driven” decisions are prioritized over anecdotal ones, and “evidence”--Fox News notwithstanding--is taken to mean “backed up by numbers and facts.” Now, any self-respecting feminist would be the first to tell you that personal accounts should matter as much as any meta-study, and “evidence” can take a range of qualitative and quantitative forms. To disagree with those statements would undo the work of the many feminist activists and scholars of the 1980s and early 1990s who struggled to get qualitative methods, such as interviews and participant observations, accepted as legitimate evidence in the first place. But there is, undeniably, what feminist demographers Christina Hughes and Rachel Lara Cohen call a “pragmatic politics” of using quantitative methods for feminist aims. If the goal is to work towards justice, then by all means use whatever form of evidence is most convincing. It would be an injustice not to!

That being said, there is a second argument in favor of quantitative methods that has less to do with pragmatism, and more to do with the nature of the problem at hand. So many issues of structural inequality are problems of scale, and can seem anecdotal until they are seen as a whole. For instance, when Natalie Wreyford and Shelley Cobb set out to count the women involved in the film industry in the UK, they encountered a female screenwriter who had never considered the fact that, in the UK, male screenwriters outnumber women at a rate of four to one. “Isn’t that a funny old thing?” she said. “I didn’t even know that because screenwriters never get to meet each other.”  

But it’s far less funny when the subject is a matter of life or death, as in ProPublica’s reporting on the racial divide in maternal mortality in the United States, which we discuss in Bring Back the Bodies. While they interviewed the families of many Black women who had died while giving birth, few were aware that the phenomenon extended beyond their own family. But the racial disparity in maternal health outcomes is indeed a structural problem, and it’s why feminist sociologists like Ann Oakley have long advocated for the use of quantitative methods alongside qualitative ones. Without big data, Oakley explains--although she used the term “quantitative research,” since she was writing in 1999--“it is difficult to distinguish between personal experience and collective oppression.”                          

But before issues like the racial divide in maternal mortality, or the structural racism that underlies it, can be identified through large-scale analyses like the one that ProPublica conducted, the data must exist in the first place. Which brings us back to Maria Munir and the importance of collecting data that reflects the population it claims to represent. On this issue, Facebook of all companies was ahead of the curve when, in 2014, it expanded its gender options from the standard two to over fifty choices, ranging from “genderqueer” to “neither”--a move that was widely praised by a range of gender non-conforming groups. One year later, when the company abandoned its select-from-options model altogether, replacing the “Gender” dropdown menu with a blank text field, the decision was touted as even more progressive. Because Facebook users could input any word or phrase in order to indicate their gender, they were at last unconstrained by the assumptions imposed by any preset choice.

But research by Rena Bivens, a scholar of social media, has revealed that, below the surface, Facebook continues to resolve users’ genders into one of either male or female. Evidently, this decision was made so that Facebook could allow its primary clients-- advertisers-- to more easily market to one gender or the other. Put another way, even if you can choose the gender that you show to your Facebook friends, you can’t change the gender that Facebook’s advertisers ultimately see. And this discrepancy leads right back to the body issues we discussed in Chapter One: it’s corporations like Facebook, and not individuals like Maria Munir, who control the terms of data collection--even if it’s folks like Munir, who have personally (and often painfully) run up against the limits of our current classification systems, who are best positioned to improve them.

Feminists have also spent a lot of time thinking about classification systems, as it turns out, since the criteria by which people are divided into the categories of “male” and “female” is exactly that: a classification system. And while the gender binary is one of the most universal classification systems in the world today, it is no less constructed than the Facebook advertising platform or, say, the Golden Gate Bridge. The Golden Gate Bridge is a physical structure; Facebook Ads is a virtual structure; and the gender binary is a conceptual one. But each of these structures was created by people: people living in a particular place, at a particular time, and who were influenced--as we all are-- by the world around them.

So this starts to get at the meaning behind the phrase, “gender is a social construct.” Our current ideas about the gender binary can be traced to a place (Europe) and a time (the Enlightenment) when new theories about democracy and what philosophers called “natural rights” began to emerge. Before then, there was definitely a gender hierarchy, with men on the top and women on the bottom. (Thanks, Aristotle!) But there wasn’t a binary distinction between those two genders. In fact, according to the historian of gender, Thomas Laqueur, most people believed that women were just inferior men, with penises located inside instead of outside of their bodies, and that-- for reals!-- could descend at any time in life.

For the gender binary to emerge, it would take figures like Thomas Jefferson declaring that all men were created equal, and entire countries (like the U.S.) founded on that principle, before those same figures began to worry what, exactly, they had declared-- and, even more worrisome, to whom it actually applied. All sorts of systems for classifying people date to that era-- not only gender but also, crucially, race. Before the eighteenth century, Western societies understood “race” as a concept tied to religious affiliation, geographic origin, or some combination of both. Although it’s hard to believe, race had nothing to do with skin color until the rise of the transatlantic slave trade, in the seventeenth century. Even then, race was still a hazy concept. It would take the so-called “scientific racism” of the mid-eighteenth century for race to begin to be defined in terms of black and white.

Ever heard of Carl Linnaeus? Think back to middle school, when you likely learned about the binomial classification system that he is credited with creating. Well, Linnaeus’s revolutionary system didn’t just include the category of homo sapiens; it also, lamentably--but as historians would tell you, unsurprisingly--included five subcategories of humans separated by race. (One of these five was set aside for mythological humans who didn’t exist in real life, in case you’re still ready to get behind his science). But Linnaeus's classification system wasn’t even the worst of the lot. Over the course of the eighteenth century, increasingly racist systems of classification began to emerge. These were systems that were designed to exclude, and in instances as far-ranging as the maternal health outcomes we’ve already discussed, to Google search results for “black girls” vs. “white girls,” as information studies scholar Safiya Noble has shown, we can detect the effects of those racist systems every day. 

A simple solution would be to say, “Fine, then. Let’s just not classify anything, or certainly anyone!” But the flaw in that plan is that data must also be classified in some way in order to be put to use. Data, after all, is information made tractable, to borrow a term from computer science (and from another essay that Lauren wrote with a colleague in information studies, Miriam Posner). “What distinguishes data from other forms of information is that it can be processed by a computer, or by computer-like operations,” we write there. And in order to enable those operations, which range from counting to sorting, and from modeling to visualizing, the data must be placed into some kind of category--if not always into a conceptual category like “gender,” then at the least into a computational category like “integer” (a type of number) or “string” (a sequence of letters or words).

It’s been argued that classification systems are essential to any working infrastructure-- and not only to computational infrastructures or even conceptual ones, but also to physical infrastructures like the checkout line at the grocery store. Think about how angry you get when you’re stuck in the express line behind someone with more than fifteen items. Or, if that’s not something that gets you going, just think of the system you use to sort your clothes for the wash. It’s not that we should reject these classification systems out of hand, or even that we could if we wanted to. (We’re pretty sure that no one wants all of their socks to turn pink). It’s just that we rarely question how classification systems are constructed, or ask why they might have been thought up in the first place. In fact-- and this is a point also made by the influential information theorists Geoffrey Bowker and Susan Leigh Star-- we tend not to even think to ask these questions until our systems break.

Classification systems can break for any number of reasons. They can break when an object-- or, more profoundly, a person-- can’t be placed in the appropriate category. They can break when that object or person doesn’t want to be placed in an appropriate category. And they can break when that object or person shouldn’t even be placed in a category to begin with. In each of these cases, it’s important to ask whether it’s the categories that are broken, or whether-- and this is a key feminist move-- it’s the system of classification itself. Whether it’s the gender binary, or the patriarchy, or-- to get a little heady-- the distinction between nature and culture, or reason and emotion, or public and private, or body and world, decades of feminist thinking would tell us to question why these distinctions might have come about; what social, cultural, or political values they reflect; and, crucially, whether they should exist in the first place.   

But let’s spend some time with an actual person who has done this kind of thinking: one Michael Hicks, an eight-year-old Cub Scout from New Jersey. Why has this kid started to question the broken systems of classification that surround him? Well, Mikey, as he’s more commonly known, shares his first and last name with someone who has been placed on a terrorist watch list by the U.S. federal government. As a result, Mikey is subjected to the highest level of airport security screening each time that he travels. “A terrorist can blow his underwear up and they don’t catch him. But my 8-year-old can’t walk through security without being frisked,” his mother lamented to Lizette Alvarez, a reporter for The New York Times, who covered the issue in 2010.

Of course in some ways, Mikey is lucky. He is white, so he does not run the risk of racial profiling—unlike, for example, the many Black women who receive TSA pat-downs due to the perceived threat of their natural hair. Moreover, Mikey’s name is not Muslim-sounding, so he does not need to worry about religious or ethnic profiling either--unlike, for another example, people named Muhammad who are pulled over by the police due to the perceived threat of a Muslim first name. But Mikey the Cub Scout still helps to expose the brokenness of the categories that structure the TSA’s terrorist classification system; the combination of first and last name is simply insufficient to classify someone as a terrorist or not.

Or, consider another person with a history of bad experiences at the (literal) hands of the TSA: Sasha Costanza-Chock. Costanza-Chock is non-binary, like Maria Munir. They are also a design professor at MIT, so they have a lot of experience not only living with, but also thinking through broken classification systems. In a recent essay, they describe how the seemingly simple system employed by the operators of those hand-in-the-air millimeter-wave-scanning machines is in fact quite complex-- and also fundamentally flawed.

No one but a gender non-conforming person would know that, before you step into a scanning machine, the TSA agent operating the machine looks you up and down, decides whether you are male or female, and then pushes a button to select the appropriate gender on the scanner’s touch-screen interface. That decision loads the algorithmic profile for either male bodies or female ones, against which your measurements are compared. If your body’s measurements diverge from the statistical norm of that gender’s body-- whether the discrepancy is because you’re concealing a deadly weapon, or because the TSA agent just made the wrong choice-- you trigger a “risk alert,” and are subjected to the same full-body pat-down as a potential terrorist. So here it’s not that the scanning machines rely upon an insufficient number of categories, as in the case of Mikey the Cub Scout; or even that they employ the wrong ones, as Mikey’s mom would likely say. It’s that the the TSA scanners shouldn’t rely on the category of gender to classify air-travelers to begin with. 

So when we say that what gets counted counts, it’s folks like Costanza-Chock, or Mikey, or Maria Munir, that we’re thinking about. Because broken classification systems like the one that underlies the airport scanner’s risk detection algorithm, or the one that determines which names end up on terrorist watch lists, or simply (simply!) the gender binary, are often the result of larger systems that are themselves broken, but that most people don’t often have the opportunity to see. These invisible systems are what philosopher Michel Foucault would call systems of power. Systems of power don’t simply determine the categories into which individual objects or people are sorted; they over-determine how those groups of objects or people experience the world.

What does it mean for a system to over-determine how people experience the world? Many feminists would point to the example of the patriarchy--a word that describes the combination of legal frameworks, social structures, and cultural values that contribute to the continued male domination of society. But for a more concrete example, we could return to Facebook. It’s not only that anyone who types in a gender that is not “male” or “female” is reduced, in the eyes of advertisers, to the single category of “unknown.” It’s also that, at the level of code, these three remaining categories are further reduced to numerical values: 1, 2, and 3, respectively. So when an app developer requests a list of users sorted by gender for any reason-- whether it’s to sell them useless diet pills, 50% off retail (which no one ever wants); or to offer them a free financial consultation, first come first served (which many people do)-- they receive a list in which male Facebook users are hard-coded to be always first in line.

Now, the software engineers who wrote the word-to-number code were almost certainly not intending to discriminate. They were probably only thinking, “How can we make our gender data easier to sort and manage?” And when it comes to computational data, it’s almost always easier and more efficient to deal with numbers than it is to deal with words. But it’s also not a surprise that in a group of engineers which is a reported 87% male, no one thought to point out (or maybe just that no one felt comfortable saying out loud) that a data classification system in which men are always ranked first might lead to problems for those who ranked second or third-- not to mention those excluded from the list altogether. In fact, if you were to ask a feminist theorist like Judith Butler to weigh in, she’d tell you that the inadvertent and invisible way in which systems of power reproduce themselves is exactly how the gender binary consolidates its force.

It’s not only Facebook that’s to blame. Gender data is almost always collected in the binary categories of male and female, and visually represented by some form of binary as well. This remains true even as a recent Stanford study found that, when given the choice among seven points on a gender spectrum, more than two-thirds of the subjects polled placed themselves somewhere in the middle. It’s also important to remember that there have always been more variations in gender identity than Anglo-Western societies have cared to outwardly acknowledge or collectively remember. These third, fourth and n-th genders go by different names in the different historical and cultural circumstances in which they originate, including female husbands, indigenous berdaches, Hijras, two-spirits, pansy performers, and sworn virgins, along with the category of transgender that we most commonly use today. 

Now, as data analysts and visualization designers, we can’t always control the collection process for the data we use in our research. Like the Facebook engineers, we’re often working with data that we’ve obtained from someplace else. But even in those cases--and, arguably, especially in those cases--it’s important to ask how and why the categories of the dataset we’re using were constructed, and what systems of power they might represent. Because when it comes to classification systems, there’s power up and down, side to side, and everywhere in between. And it’s on us, as data feminists, to ensure that any differentials of power that are encoded in our datasets don’t continue to spread.

Whether we like it or not, we’re all already swayed by these systems of power, as well as by the heuristic techniques that reinforce them. Before you say, “Wait! No one taught me those techniques!” consider that “heuristic techniques” is just a fancy term for the use of mental shortcuts to make judgements--in other words, common sense. The tendency of people to adhere to common sense offers a great evolutionary advantage, in that it’s enabled humanity to survive over many millennia. (What tells you to run away from a bear? Common sense! What tells you not to eat rancid meat? Also common sense (and your gag reflex)). But as the renowned work of cognitive psychologists Daniel Kahnemann and Amos Tversky has showed, this reliance on heuristics eventually leads to an accumulation of cognitive biases--what might be otherwise understood as a snowball of mistaken assumptions that, in a world more challenged by structural inequalities than by grizzly bears, leads to profoundly flawed decision-making, and equally profoundly flawed results.

Buster Benson, a product manager at the crowd-funding platform Patreon, has made a hobby of classifying these cognitive biases, and with John Manoogian, has visualized them in the chart you see above. If you look at the lower half of the image, you see can see the two quadrants-- “Need to Act Fast” and "Not Enough Meaning" --that include some of the key cognitive biases that come into play when collecting and classifying data.

Now imagine, for a moment, that you are designing a new survey for an analysis of gender and cell phone usage, but you have not yet finished reading this book. Gender is something you are pretty familiar with, you might say to yourself, since you have a gender, and everyone else you know has a gender too. But this is called the overconfidence effect, found on the lower left of the chart in lime green. Still, you go on: in your experience there are two genders, male and female, and everyone else you know would say so, too. (This is called the false-consensus effect, also on the lower left). Men and women should clearly be placed in separate categories, since they are different kinds of people. (This is called essentialism; file under “Not Enough Meaning”). Also, everyone knows that women like talking—stereotyping alert!—so in addition to gender data, how about collecting cell phone minutes data too. (You’ve just committed a fundamental attribution error, in blue on the right).  

Fast forward past the data collection phase to the analysis portion of the project. You note that you were right in your initial assessment of the situation: women did talk on their cell phones more than men. This forms the basis of your subsequent analysis. (This is called confirmation bias). In addition, in your zeal to confirm your essentialist beliefs, you entirely missed an important phenomenon: millennial-aged people of all genders have extremely large social networks. Your expectation bias prevented you from discovering some important insights that might have informed the design of a new product. You receive a negative performance review, and you are fired.

What interrupts this series of bad decisions? Recognizing that common sense is often sexist, racist, and harmful for entire groups of people--especially those groups, like women, who find themselves at the bottom end of a hierarchical classification system; or like non-binary folks, who are excluded from the system altogether.

As should now be clear, a feminist critique of classification systems is not limited to data about women, or to the category of gender alone. This point can’t be overstated, as it forms the basis for the theories of intersectional feminism that inspire this book. Feminist scholars Brittney Cooper and Margaret Rhee address this issue directly in their call to use feminist thinking to “hack” the binary logic that simultaneously underlies the racism experienced by Black people in the United States, and erases the other forms of racism experienced by Latinx, Asian American, and Indigenous groups. “Binary racial discourses elide our struggles for justice,” they state clearly, and we agree. By hacking the binary distinctions that erase the experiences of certain groups, as well as the systems of power that position those groups against each other, we can work towards a more just and equitable future.

Even though the stakes of this project are high, it’s possible for anyone, including you, our readers, to contribute. One of the best visualizations of the concept of intersectionality that we’ve found, for instance, comes from a series of posts on anonymously-authored WordPress blog. “Intersectionality, Illustrated” offers a series of visualizations that employ color gradients to represent the multiple axes of privilege (or the lack thereof) that a person might encounter in the world. At the center of each visualization is a solid circle, which represents that person’s goals and dreams for their life. Colorful lenses spiral out from the center, each representing an aspect of that person’s identity: ethnicity, age, sexual orientation, and so on. In this visualization, opacity is employed to show whether a particular identity trait contributes to an enhanced capacity to achieve one’s personal goals, or a diminished one. A directional gradient underscores how that trait alternately supports the person’s goals, or distances them from them. In this way, the viewer begins to literally see how an intersection of privileged positions-- a term used to describe the advantages offered only to particular groups, such as those that come along with being white, male, able-bodied, or college-educated--can lead to an array of colorful options for the future. An intersection of disadvantaged positions, on the other hand, such as being gay, or transgender, or disabled, or poor, reduces– and, at times, eliminates altogether– that person’s ability to pursue a particular life path. It’s a simple visualization, which relies only upon the creative use of color, opacity, gradient, and form, and yet it illustrates a powerful point: that one’s identity, and therefore one’s privilege, is determined by multiple factors that all intersect.

In addition to the intersection of the various aspects of a person’s identity, each individual aspect can be quite complex. Again, an anonymous person on the internet offers among the most inspiring examples for considering how we might visualize gender, for example, if we weren’t limited to to the male/female split. The creator of the Non-Binary Safe Space Tumblr shows how gender might be visualized as a spectrum, or as a branching tree. They sketch out how non-binary genders might be placed around a circle, in order to emphasize shared sensibilities rather than differences; or plotted on a Cartesian plane, in which “male” and “female” serve as the axes, with infinite points in between. They even wonder about designing a series of interactive sliders, with “female” and “not female,” “male” and “not male,” and “other” and “not other,” serving as the respective poles; or even a 3D cube, with a vector charting a person’s changing course through their evolving sense of self. These are designs that, like “Intersectionality, Illustrated,” come from personal experience, and they offer a powerful point of departure for thinking through new classification systems and visualization schemes.  

 When we went to track down the permissions for the Non Binary Safe Space Tumblr, we discovered that the site had been taken over by spammers. But maybe it’s a sign of the times (along with the inevitable descent into spam) that some of these ideas have already begun to enter major publications. For example, when Amanda Montañez, a designer for Scientific American, was tasked with creating an infographic to accompany an article on the evolving science of sex and gender, she envisioned a spectrum not unlike the one pictured above. But she soon found confirmation of what feminist theorists have been saying for decades (and what we’ve been saying so far in this book): that sex and gender are not exactly the same thing. More than that, what we might think of as the easier concept to explain--the biological category of sex--is just as fluid and complicated as the social category of gender.  

The result, “Beyond XX and XY,” a collaboration between Montañez and the design firm Pitch Interactive, is a complex diagram, which employs a color spectrum to represent the sex spectrum, a vertical axis to represent change over time, and branching arrows to connect to text blocks that provide additional information. Montañez hopes that visualization, with its careful adherence to terminology, and inclusion of only properly categorized data, will help “raise public awareness” about intersex as well as transgender and non-binary people, and “help align policies more closely with scientific reality, and by extension, social justice.” In other words, Montañez made what was already counted count.

Even when working with binary gender data, designers can still make those limited categories count. For example, in March 2018, when the reporters on the Lifestyle Desk of The Telegraph, a British newspaper, were considering how to honor International Women’s Day, they were struck by the significant gender gap in the UK in terms of education, politics, business, and culture. They didn’t have the time or the expertise to collect their own data, and even if they had, there’s no telling as to whether they would have collected non-binary gender data. But they wanted to ensure that they didn’t further reinforce any gender stereotypes. They paid particular attention to color, with the awareness that even as many designers are moving away from using pink for girls and blue for boys, most still adhere to the logic that associates warm colors with women and girls, and cool colors with men and boys. Because the stereotype that women are warmer and more caring, while men are cooler and more aloof, is still firmly entrenched in many cultures, the associated colors are easier to interpret—or so this argument goes. 

This stereotype is, of course, another hierarchy, and the goal of the Telegraph team was to mitigate inequality, not reinforce it, and so they took a different source for inspiration: the “Votes for Women” campaign of early 20th century England, in which purple was employed to represent freedom and dignity, and green to represent hope. When thinking about which of these colors to assign to each gender, they took a design principle as their guide: “Against white, purple registers with far greater contrast and so should attract more attention when putting alongside the green, not by much but just enough to tip the scales. In a lot of the visualizations men largely outnumber women, so it was a fairly simple method of bringing them back into focus,” Fraser Lyness, the Telegraph’s Director of Graphic Journalism told Lisa Charlotte Rost, herself a visualization designer who interviewed Lyness for her blog. Here, one hierarchy, the hierarchy in which colors are perceived by the eye—was employed to challenge another one—the hierarchy of gender. Lyness was right. It was a “fairly simple method” to employ. But when put into practice, it had profound results.

There are all sorts of instances of designers, as well as journalists, artists, activists, and scholars, using data and design to bring issues of gender into view. P. Gabrielle Foreman and her team at the University of Delaware are creating a historical dataset of women who would otherwise go uncounted, and therefore unrecognized for their work. The team’s focus is on the women who attended but were not named as participants in the nineteenth-century Colored Conventions, organizing meetings in which Black Americans, fugitive and free, met to strategize about how to achieve educational, economic, and legal justice. Because these women often worked behind the scenes, packing the lunches and watching the children so that their husbands could attend; running the boarding-houses where out-of-town delegates stayed during the conventions; or even, as research has shown, standing in the back of the meeting hall in order to make their presence known, their contributions were not considered as participation in the events. But as continues to be true today--think back to the issue of maternal mortality mentioned at the beginning of Chapter One, or to the issue of sexual assault, as we discuss more in The Numbers Don’t Speak for Themselves—the systems of power that place women below men in patriarchal societies such as ours are the same that ensure that the types of contributions that women make to those societies are valued less, and therefore less likely to be counted.

But counting is not always an unmitigated good. Sometimes counting can have unintended consequences, really bad ones, especially for marginalized groups. Some transgender people, for example, prefer not to disclose the sex they were assigned at birth, keeping their identity as a trans person private. Even for those who generally choose to make their trans identity public, being visibly identified as trans on a map, or in a database, for example, could expose them to violence. Even in a big dataset, there is no additional strength in numbers. Compared to cisgendered people (folks whose genders match the sex they were assigned at birth), trans people are so small a group that they are more exposed, and therefore more vulnerable.

A similar paradox of exposure is evident among undocumented immigrants; visualizing the precise locations of undocumented immigrants may, on the one hand, help make an argument for directing additional resources to a particular area, but on the other, may alert ICE officials of the locations of their homes or schools, making the threat of deportation more likely. In cases where lives are at stake, and the security of the data can’t be guaranteed, not collecting statistical outliers can be the best way to go, as Catherine has argued in some of her other work. In other cases, however, the decision to exclude outliers can be viewed as “demographic malpractice,” since it completely erases the record of those whose experiences are already marginalized in their everyday lives, and forecloses any future analysis for good or ill.

Is there any way out of this paradox? Feminist geographer Joni Seager has studied this issue for decades, and in 2004, experienced its effects firsthand when she began what she thought would be an easy project: making a map of women doctors for her monumental Atlas of Women in the World. But she hit a wall when she discovered that the World Health Organization data on medical professionals did not include a field for gender. Seager had to abandon the map, and as a result, she could not include any information about women doctors in her Atlas. Ever since, her approach has been to always collect gender data according to the most precise possible categories, and also to always ask-- before the analysis phase-- whether the data should be aggregated or otherwise anonymized in order to mask any potential adverse effects.  

[ IMAGE FROM SEAGER ATLAS TK ]

Seager’s research is focused on the collection practices associated with global and nation-wide data, where she has found that gender data is often collected but rarely made available or analyzed in disaggregated form. For example, in 2015, the Pew Research Center published a report about cell phone use in Africa. “Cell Phone Ownership Surges in Africa,” was the title of the report; and the first chart showed the growth in cell phone ownership in the United States compared with several African countries. But buried in the text of report was a surprising finding: “Men are more likely than women to own a cell phone in six of the seven countries surveyed.” Now, this would seem like an important distinction-- and perhaps one tied to other inequities-- but because gender was not treated as a primary category of analysis, those who didn’t read the fine print might not come away with one of its most important findings. In the case of this study, it wasn’t a question of what got counted that turned out to matter, but how that counting was put to use. 

Sometimes, however, questions about counting shouldn’t be answered by the survey designer, or by the data analyst, or even by the most careful reader of this book. As a final example helps to show, questions about counting often go hand-in-hand with questions of consent. Flash back to another era -- 2006 -- when another debate about a border wall was underway. Its source was the Secure Fence Act, a bill signed into law by then President George W. Bush, which authorized the construction of 700-mile-fence along the US-Mexico border. But for the fence to be completed, it would have to pass through the Tohono O'odham Nation, which straddles both countries. Recognizing that they would have to build around several sacred burial sites, the U.S. Government requested that the O’odham nation provide them with the locations of those remains.

In O'odham tradition, however, the locations of burial sites constitute sacred knowledge, and cannot be shared with outsiders under any circumstances. The O'odham Nation refused to violate its own laws by divulging information about its burial sites to the U.S. government, but it could not oppose the legal or political power of the United States. The United States built the fence, unearthing many O’odham remains in the process, and the tribe spent months attempting to get the US to return them.

But why should it be assumed that the O'odham Nation, which has existed for thousands of years, weigh its own laws less heavily than those of the United States, which-- after all-- has existed for less than two hundred fifty? Who has the right to demand that information be made public, and who has the right to protect it? And what are the cultural assumptions-- and not just the logistical considerations-- that go along with making knowledge visible and information known.

We’ve all heard the phrase “knowledge is power,” and the example of the border wall shows how this is undeniably true. But the range of examples in this chapter, we hope, also help to show how knowledge can be used to contest power, and to begin transform it. By paying attention to the politics of data collection, and to the systems of power that influence how that data is collected, we can work to rebalance some of the relationships that would otherwise contribute to their force. We might look to large institutions like the National Library of New Zealand, which began the Ngā Upoko Tukutuku Reo Māori Working Group to develop new subject headings for the Māori materials in its collections, ensuring that those materials would be classified in terms of subjects that make sense within a Māori worldview. We might look to small research groups like Mobilized Humanities, which aggregated and visualized dozens of public data sets relating to the U.S.’s “Zero Tolerance” policy, in order to call attention to the humanitarian crisis that unfolded along the U.S./Mexico border in Summer 2018. We might look to individual artists like Caroline Sinders, who is developing a data set of intersectional feminist content that can be used to train the next generation of feminist AI. Or we might look to distributed movements like #SayHerName, which employed that Twitter hashtag to create a digital record of the police violence against Black women that would otherwise go unrecorded.

These are each projects that recognized that what gets counted counts, and how the act of counting, and how we decide to show our results, profoundly influences the ideas we’re able to take away. An intersectional feminist approach to counting, like the one we’ve demonstrated here, insists that you always ask questions about the categories that structure your data, and the systems of power that might, in turn, have structured them.