Skip to main content
SearchLoginLogin or Signup

Chapter 2 – Official Data, Missing Data, Counterdata

Published onNov 02, 2022
Chapter 2 – Official Data, Missing Data, Counterdata

In 2019, two non-profit organizations Proyecto Matria and Kilómetro Cero released a somber report called The persistence of indolence: feminicides in Puerto Rico 2014-20181. The report compared various sources of data about feminicide with official police records of murdered women and found that, in any given year, police data were missing between 11% and 27% of the year's victims. That is to say, the police did not know about, or have on record, the murders of almost a third of the women killed in Puerto Rico in a given year. Thus, the "indolence" in the title of the report referred to the Puerto Rican state's response to the problem of feminicide on the island. This indolence had grown in the years leading up to the report and was woven together with larger shifts: the Office of the Attorney General of Women was increasingly politicized and rendered ineffective, neoliberal measures gutted public services, and the island experienced increasing environmental shocks2. As Mari Mari Narváez, founder of Kilómetro Cero, explains, "Our interest initially was to collect data for us to be able to further expand our work on the state's response, particularly that of the police, in regards to gender violence. There really was a very basic data problem in Puerto Rico. In other words, it was simply not known how many feminicides there were."3

A flow chart triangulating feminicide cases from 2014-2018 using official records from the state and unofficial records collected from media reports. The top box lists the total feminicides, which is 266. This is sub categorized in the layer below as 257 factual cases and 9 estimated cases. The factual cases are broken further down in the third layer, indicating 68 from the press, 33 from Vital records, and 156 from both.
Figure 2.1. Diagram from The persistence of indolence: feminicides in Puerto Rico 2014-2018. The diagram illustrates how Kilómetro Cero and Proyecto Matria triangulated their case counts across multiple sources of information, comparing news reports and the official demographic registry. Source: Kilómetro Cero and Proyecto Matria.

This chapter explores the official data published by the state in relation to the twin phenomena of missing data – exemplified by the many murders of women in Puerto Rico that are inexplicably absent from official reports – and counterdata – those cases painstakingly compiled by individuals and feminist organizations in Puerto Rico and assembled for analysis in the report. Missing data and counterdata are central concepts for data activism about feminicide, so this chapter provides more background on both, including literature review, practical definitions and further theoretical elaboration. We will also see how both are at work in this specific case of the persistence of indolence report and its impact in Puerto Rico. The production of counterdata is not a simple task, particularly in an information ecosystem characterized by institutional inaction, ignorance, absence and not-knowing. As Alice Driver observes, there is a disquieting parallel here with the disappearances that often mark the murder of women: missing bodies are accompanied by unrecorded violence4. And no one can be held accountable for what is not known.

When the two nonprofit groups decided that they wanted to study feminicide in Puerto Rico, they had to figure out how to get the data. And that was how, in 2018, Proyecto Matria and Kilómetro Cero made a visit to an égida – a senior living facility – in San Juan where a retired social worker named Carmen Castelló lives on her state pension. Carmen retired in 2010 and began caring for her grand-niece, baby Alba. While the baby slept, she would watch the news on TV and read newspapers, and she was alarmed at what appeared to be increasing rates of feminicide and sexual assault. She started applying skills she had learned on the job, "Because we, as social workers, when we serve people, we open a case file and follow up on it, right?"5

Carmen discovers cases through the news, and logs each case of feminicide that she finds in a Word document. She monitors the news everyday but tries not to do it early in the morning because it affects her emotionally to start her day with violence. Carmen has become something of an expert on the media ecosystem in Puerto Rico and in particular looks to lesser known news sources from small towns and villages for details about each case. As the case develops, she follows it through the justice system and updates the Word document with the new information, carefully noting the sources of the information. At the same time, she also publishes all cases and updates on her Facebook page: Seguimiento de Casos (Case Tracking)6. Carmen also compiles separate databases about sexual abuse and missing women. When Proyecto Matria and Kilómetro Cero approached her about their study, Carmen was happy to work with them – she shares her files freely with any organizations working to address the issue.

A middle-aged woman sitting at her desk looking back. Her eyebrows are raised and she is wearing glasses. A computer and photographs on a desk are visible in the background.
Figure 2.2. Carmen Castelló is a retired social worker who produces the most reliable data on feminicide in Puerto Rico while she cares for her grand-niece Alba and other young relatives. Courtesy of Ana María Abruña Reyes and TodasPR.

Carmen's data became the starting point for The persistence of indolence report, though project members did want to understand how thorough Carmen's media monitoring was in order to assess the quality of her information. They dedicated a staff member to monitor feminicides reported in the media for two months. The staff member was unable to find any case which had not already been logged by Carmen, so the group decided to collaborate with Carmen rather than duplicate efforts. Says Debora Upegui Hernandez, a data analyst who joined the project in 2020, "So far [Carmen's data] has been the most reliable and most up to date source I have seen." The group fact-checked and verified Carmen's existing data and set to work on the rest of the study.

Proyecto Matria and Kilómetro Cero then did something quite creative – but, as we will see, also quite common across anti-feminicide data activist groups – to try to assess the undercounting of feminicides. They triangulated data gathered from press reports with data from official records. They combined the data from Carmen with open data from the federal demographic registry which logs births, deaths and marriages across the country. Notably, these data were only made public because of a lawsuit won by the Centro de Periodismo Investigativo (Center for Investigative Journalism) in 2018. This lawsuit had stemmed from another high-profile dispute around numbers: the insistence of the Puerto Rican government that there were only 64 deaths following Hurricane Maria in 2017, versus investigative reports and research studies showing thousands more7. The final number was pegged at 2,975 and accepted by the government only after significant domestic and international pressure.

Cross-referencing Carmen's data from news reports and the demographic registry for the time period of 2014-2018, the researchers found 156 cases in both sources, 68 only noted in news reports, and 33 cases only in the demographic registry (figure 2.1). They were then able to compare these totals with aggregate numbers published by the police to demonstrate that agencies of the state had failed in "rigorously documenting the situation of feminicides in Puerto Rico and in disclosing it to the public."8 The report concluded that the numbers published by the police significantly underreported feminicide – in any given year, police missed and did not count up to almost a third of feminicides that were reported in other sources9. With these techniques – of counterdata production, triangulation across multiple data sources, and systematic case logging – Proyecto Matria and Kilómetro Cero and Carmen were able to quantify missing data in order to build a case for institutional neglect and make demands for visibility, resources and better measurement. They wrote in the report, "Collection and analysis of information on feminicide are fundamental tools to determine its magnitude, to understand its patterns and trends, and to establish international comparisons that serve as an instrument to assess the successes and failures of prevention efforts."10

Feminicide and Missing Data

The data sourcing, analysis and triangulation techniques used by Carmen and by The persistence of indolence report are without doubt creative and rigorous. But it begs the earlier question that we asked about María Salguero: Why is this work being undertaken by individuals, journalists, nonprofits and feminist collectives? Not surprisingly, these data activists are very aware of the irony that, as Marta Perez of the Argentine group Mujeres de Negro states, "we are doing work that belongs to the State."11 According to Irma Lugo, Coordinator of the Gender Equity Observatory of Puerto Rico, the government has reduced funds and dismantled services for gender-related violence over the last decade and so feminist organizations in Puerto Rico have stepped in to fill that void, even though they continue to demand that such work is the government's responsibility12.

Indeed, political demands for data and information are a central aspect of missing data as I am discussing it in this book. Before we go further, it may be helpful to review conceptions of missing data from various fields and disciplines. There is more than a century of literature about missing data in statistical analysis, including theories of missing data and techniques for handling such missing values in the data analysis process. Craig K. Enders has an excellent applied introduction to this work and Sara Fernstad has recently explored how data visualizations should represent missing data13. This literature has a typology of the mechanisms that lead to missingness. Feminicide data, in this typology, can be considered "Missing Not at Random" data – this is to say that there are underlying patterns to the missingness (such as patriarchy) that are not random. In a 2021 paper, statistician Maria Gargiulo outlines two main forms of missing data related to femicide data: deaths that are recorded but information is missing to classify them as femicide and deaths that are not recorded at all. When official or activist datasets about femicide do exist, they represent convenience samples influenced by a number of factors including selection bias (for example, it is easier to document violence in urban areas versus remote rural areas) and reporting bias (media consistently under report cases about racialized women). Gargiulo concludes with a set of recommendations for mitigating some of these biases for researchers undertaking analysis of femicide data14.

Excepting this excellent case study, the statistical literature on missing data does not generally try to explain – in any causal or contextual way – why observations or values or whole datasets do not exist. These explanations have come from scholars in the social sciences, international development, humanities and the arts. For example, there is a long feminist literature about women's invisibility in economic and labor data due to their relegation to the domestic sphere. Since the 1970s, social scientists have been pointing out how these systemic biases result in missing data: "Quite simply, governments and international agencies produce data only for those aspects of social life they deem important."15 Geographer Joni Seager has spoken about the "data voids" that result because of institutional neglect of women's lives, a perspective recently popularized by Caroline Criodo Perez' book Invisible Women: Data Bias in a World Designed for Men16. And through a 2016 artwork and accompanying essay, Mimi Ọnụọha posited the idea of missing datasets – whole collections and categories of data that are disregarded, overlooked, and discounted. In her project, missing datasets include such topics as "People excluded from public housing because of criminal records" and "Trans people killed or injured in instances of hate crime."17

Ọnụọha links these missing datasets to the study of ignorance: "That which we ignore reveals more than what we give our attention to. It’s in these things that we find cultural and colloquial hints of what is deemed important. Spots that we've left blank reveal our hidden social biases and indifferences."18 In fact, there is a whole body of work called agnotology – the study of ignorance19. While philosophy tends to think deeply about how people come to know things, work on the epistemology of ignorance examines how people and societies come to not know things. For Charles W. Mills, who brought critical race theory to bear on this work, this is exemplified by the ignorance of white people of the effects of their system of racial domination in North America20. Feminist philosopher Nancy Tuana applies this line of thinking to knowledge about women's health, where she develops a taxonomy of types of ignorance, including "willful ignorance" – where missing data and missing knowledge about gender violence would be situated– “a systematic process of self-deception, a willful embrace of ignorance that infects those who are in positions of privilege, an active ignoring of the oppression of others and one’s role in that exploitation”21. Indeed, in a reflection on lessons to be learned from Audre Lorde's life and work, health researcher Lisa Bowleg stated that "epistemological ignorance is one of the master's most formidable tools."22 And through her framing of "data silences", Indigenous studies scholar Bronwyn Carlson renders this process of ignorance even more active – she links the production of not-knowing directly to gendered colonial violence whereby the settler state actively silences accounts of violence perpetrated on Indigenous bodies23. There are also cases when the state has a strategic interest in denying the existence of not only a widespread phenomenon but also certain types of human beings. For example, in her book The Uncounted, Sara Davis outlines how governments try to simultaneously fight the AIDS epidemic while at the same time denying the existence of the people most vulnerable to AIDS: sex workers, men who have sex with men, drug users, and transgender people24.

While Donna Haraway taught us that knowledge is situated, Tuana writes about how ignorance, too, is situated25. This can lead us towards Carlson's more active framing of data silences: the absence of data is not merely due to oversight, accident or "unconscious bias"26. Rather, the absence of data can be part of an active and on-going production of ignorance and maintenance of ignorance, requiring loads of labor to simply not know and not understand and not see a systemic phenomenon and not count it.

Missing data and power

The working definition of missing data that I am using in this book derives from these social, political and structural considerations on the presence and absence of knowledge. Missing data are those data that are neglected to be prioritized, collected, maintained and published by institutions, despite political demands that such data should be collected and made available. As Ọnụọha explains, to frame data as missing is to make a normative assertion: "It implies both a lack and an ought: something does not exist, but it should."27 Data are never missing in any absolute sense. Rather, data are missing because there is a political demand that such data should exist and be made available to specific groups. Moreover, missing data includes not only the total absence of data but also the production of sparse, incomplete, unreliable, misclassified, inaccessible, contested and untimely information.

Missing data, therefore, is a concept that is situated, relational and political. In the case of official data about feminicide, there is missing data precisely because grassroots feminist, Indigenous, Black, queer and women's groups and social movements make demands that such data be produced. Thus in relation to missing data we must ask, for whom are the data missing? Which groups are making the political demand for data collection and knowledge production? Who is named as the entity responsible for (or negligent in) data collection? The answers to these questions matter for understanding how power is at work throughout the sociotechnical environment of missing data. Indeed, the first two principles of data feminism call on us to examine power and to challenge power in relation to data.

Yet what exactly do we mean by "power", especially in relation to data produced about feminicide? Lawyer and trans activist Dean Spade makes a strong case that power is decentralized – there is no one person or institution responsible for administering and distributing systemic oppression. And conversely, to challenge power one cannot only look to one person or even one institution such as the law. He writes, "This way of understanding the dispersion of power helps us realize that power is not simply about certain individuals being targeted for death or exclusion by a ruler, but instead about the creation of norms that distribute vulnerability and security."28 Sociologist Patricia Hill Collins developed a Black feminist conceptual model for this multi-sited operation of power called the matrix of domination. This refers to the uneven allocation of privilege – increased life chances – for some groups, and oppression – decreased life chances – for other groups. The matrix of domination is a conceptual model that, among other things, outlines a taxonomy of four domains of oppression: structural, disciplinary, hegemonic, and interpersonal (Table 2.1)29. This model can help to pinpoint some of the diffuse and decentralized operations of power that result in the non-production of feminicide data. Missing data about feminicide is produced across these four domains, and anti-feminicide activists have also taken action in and across these four domains. This is a high-level framework because it attempts to grasp broad patterns in the non-production of feminicide data across contexts. It is one beginning to answering the question "Why and how does official data about feminicide and gender-related violence come to be missing?"

Laws and policies

The structural domain organizes oppression and inequality.

Implementation of laws & policies

The disciplinary domain manages oppression and inequality.

Media & culture

The hegemonic domain circulates oppressive ideas.

Interpersonal

Individuals experience oppression in the interpersonal domain.

Table 2.1. The four domains of the matrix of domination. Based on concepts introduced by Patricia Hill Collins in Black Feminist Thought: Knowledge, Consciousness and the Politics of Empowerment. Partially adapted from Data Feminism with permission of the authors.

Two domains in the matrix of domination can be framed as the purview of the state – laws and policies and their implementation. The first is where oppressive laws are enacted or where, conversely, lack of or inadequate legislation reinforces oppression of certain groups. As discussed in Chapter 1, thanks to persistent efforts by feminist, women-led and Indigenous-led movements, many countries have passed legislation criminalizing femicide, feminicide or strengthening protections for MMIWG2 and fatal violence against women. But this legislation often retains a narrow definition of the phenomenon and only in rare cases includes transfeminicide (see Table 1.1 in the prior chapter). Until recently, for example, Puerto Rico only counted cisgender women's murders in the context of domestic violence – what is known as intimate feminicide – but lumped in all other gender-motivated killings into the category of homicides 30. The lack of adequate legal definitions and categories thus underpins missing data - without these the state simply does not recognize a systemic form of violence, let alone count it or ensure justice for those who experience it.

The implementation of laws & policies happens in the disciplinary domain of Collins' matrix, and here many governments have failed to implement adequate information collection, failed to devote resources to accurately classify deaths, and failed to create supportive climates for survivors and families to report violence. This is primarily where the The persistence of indolence report was directed, "the Police force is indolent in its lack of diligence to renew its practices of collecting, analyzing, interpreting and disclosing statistics of murdered women in accordance with widely used international standards."31 Their main finding was the systemic undercounting of women's murders. This came to be through various factors: different state agencies were responsible for different types of violence, so there is an office for sexual assault, an office for domestic violence, and another office for homicide. Each was counting differently and their numbers on feminicide didn't correspond with each other32. There are also errors in those data that do exist. For example, reports about MMIWG2 in North America have consistently highlighted the racial misclassification of Native women33. In Puerto Rico, as the The persistence of indolence team was going through the demographic registry, there were more than a dozen cases of women murdered in a single year which were inexplicably and incorrectly marked as "suicides"34. These bureaucratic incompetencies constitute some of the more pernicious ways that the disciplinary domain contributes to missing data – through what Menjívar and Walsh have called “state acts of omission and commission,” wherein the state either indirectly or directly contributes to underreporting and allows violence to go unpunished35.

The implementation of laws and policies can also see the active suppression of data and information. High numbers of feminicides make the police and the state look bad. As Carmen recounts, "[The government] was telling us they couldn't do that [collect data about feminicide]. They themselves said: 'we cannot do that, because people will think that the relevant agencies are not doing their job.' And, we told them: 'Well, exactly, you are not doing your job.'”36 Leaders of the The persistence of indolence team also described how feminicide data requested from police came in only in aggregated form and were irregularly reported – sometimes reported by week, sometimes by month, sometimes by year. Often there were no updates for months at a time. And Debora expressed great skepticism at the official reports published about domestic violence. These in particular drew her attention because they were so "supremely consistent over the years."37 For more than ten years, the official data about domestic violence complaints appeared to be decreasing with fewer and fewer complaints each year, a trend she found extremely hard to believe given her work with grassroots survivor-led groups.

A third domain is the purview of media and culture: the hegemonic domain is where stereotypes and harmful cultural ideas circulate and this, too, plays a role in producing missing data. Media coverage of feminicide tends to reproduce gender and racial stereotypes, blames victims for their own deaths, reinforces stigma, and ultimately normalizes gender-related violence. This is typified by coverage that paints intimate partner violence as a "crime of passion", or emphasizes details about how victims were dressed or which industries they work in or how late they were out at night or how drunk they were or their prior history of incarceration, creating stratified categories of worthy victims (typically whiter, richer, settlers, gender-conforming, heterosexual, students or professionally employed) and unworthy victims38. In Guatemala, Anthropologist Sarah England found that newspaper coverage of feminicides typically provides “mainly speculation on the part of the people interviewed, the police, and the reporter.”39 In Brazil, a study found that news articles provided little information on women and the context of their deaths, and often relied on common tropes such as “jealousy” and “violent emotion” to describe the motivation for crimes—an approach that de-emphasizes the structural and systemic character of gender-related violence40. Media often directly import the framing and perspective of the police to report on cases and thus can function as a kind of cultural arm of the state. For example, in the case of Puerto Rico, when the police reported out their feminicide data one of the categories they used is "crime of passion" a category that authors of The persistence of indolence rejected as "offensive and obsolete"41. Finally, media often misgender transgender victims of violence, resulting in what Sofia Vanoli Imperiale and Eloi Leones have separately called a "double murder" in which the person's lived identity is murdered again in the media coverage of the event42.

When media circulate harmful stereotypes that dehumanize victims and normalize gender violence, this further reduces political and public will for 1) seeing fatal gender violence as a systemic violation of human rights and 2) measuring and counting fatal gender violence. But an equally harmful way that media contribute to missing data is in their unequal coverage of killed people. For example, it is so well-documented that murders and disappearances of white women in the US receive more media attention than those of women of color that there is a name for the phenomenon: "Missing White Woman Syndrome", created by Black journalist Gwen Ifill in 200443. This was evidenced in the 2021 case of Gabby Petito, a young white woman, who went missing and was later determined to have been murdered by her fiancé in Wyoming. There was extensive media coverage throughout the period of her disappearance, drawing criticism from many groups who put forward that more than 700 Native women had disappeared or been murdered in the same state with virtually no public attention44. Their point is not that Petito didn't deserve public attention and justice, but rather that Native and Black women who experience violence are being systematically denied that public attention and justice. The domain of media and culture thus also contributes to missing data through systematically declining to pay attention to the killings of certain groups of women. They fail to "ignite any form of public outrage," which Carlson and other Indigenous scholars have linked to a form of symbolic annihilation45. This prevents the public from knowing more and demanding more from its institutions. The bias in coverage also poses a significant problem for counterdata producers like Carmen and other data activists who rely heavily on media sources to compile cases that are not being tracked elsewhere, as we will discuss in detail in Chapter 4.

Finally, the interpersonal domain is where individuals experience oppression directly as discrimination and violence. In relation to feminicide, the violence extends beyond the act itself, leaving families devastated, demoralized and reluctant to report violence. In a climate of bureaucratic ineptitude (whether intentional or not), open hostility and victim-blaming, families and communities may understandably want to seek as little involvement with the state as possible, perpetuating the cycle of missing data by not reporting46. In places where there are high levels of corruption and organized crime, many police officers may be on criminal payrolls, so families themselves can face retribution and violence simply for reporting. For example, political scientist Mneesha Gellman has read many US asylum cases where women have to state their reasons for not reporting gender violence in their home country: "Reasons include the police being either gang-involved themselves or colleagues or friends with the abuser. Other reasons were that the police would either further the abuse themselves, or dismiss the abuse as to be expected given the cultural context of machismo."47

As is evident, missing data are enmeshed in power relations. Missing data are actively produced in the domains of law and policy; the implementation of law and policy; media and culture; and the interpersonal realm. Indeed, the production of silence and un-knowledge across multiple domains is central to the maintenance of the matrix itself. Latin American activists often talk about "visibilizing" feminicide and it is precisely this multi-sited production of ignorance that they are up against.

Feminicide and Counterdata Science

When the state and its institutions fail to produce or make available important information and when the media reproduce oppressive narratives and when families are demoralized and disempowered, activists are increasingly stepping into those data gaps and producing their own counterdata in order to challenge power, the second principle of data feminism. Just as missing data is produced across all four domains of the matrix of domination, data-driven acts of resistance and refusal happen across these four domains as well.

First, a short genealogy of counterdata as a concept. The term "counter-data" was introduced in 2014 by geographers Craig Dalton and Jim Thatcher who drew from "counter-mapping" to describe acts of resistance that use data to upset dominant power relations. They call for examining both speculative and existing practices: "We must ask what counter-data actions are possible? What counter-data actions are already happening?"48 As such, counterdata can be viewed as part of the larger universe of data activism practices discussed in Chapter 1. Case studies have built on this concept of counterdata. For example, Morgan Currie and colleagues described a Hackathon on Police Brutality conducted in Los Angeles in which participants worked with official and citizen-collected data about police killings as a way of interrogating what actions data do and do not make possible49. Amanda Meng and Carl DiSalvo demonstrate that grassroots counterdata production around housing in Atlanta is rooted in a long history of Black community organizing.50 In their case study on a community group in Los Angeles that documents police killings, Roderic Crooks and Morgan Currie place the group's work in a larger field of agonistic data practices. These practices constitute efforts by communities to use data to document harm, but the group they profile also uses their data in affective and narrative ways – "to motivate people to act on their passions and imagination."51 In Data Feminism, Lauren and I discussed numerous examples of counterdata production projects emerging from academia, data journalism, community organizing, and activism52.

Counterdata production is an apt term for describing the work that groups do to compile spreadsheets and databases, yet anti-feminicide data activists do much more than produce data. Their work comprises more than simply creating counts and tallies of fatal violence. They verify and triangulate information; they undertake exploratory and explanatory analysis; they draw from data to produce reports, artworks, infographics, social media campaigns, family support systems and protests in public space; they develop new theoretical concepts from their analysis. This is why I make the case for considering counterdata science as a concept that encompasses the care, rigor and systematization that grassroots groups enact as they work to count feminicide and gender-related violence. Counterdata science means mounting an explicit, and usually collective, challenge to the data practices (measurement, collection, analysis, publication) of mainstream, well-resourced "counting institutions" such as governments and corporations. It involves the production of counterdata – systematic counting – and additionally involves what happens with those records downstream, i.e. all of the data science practices that groups use to secure, systematize, analyze, publish and circulate their data.

Counterdata science has resonance with the many other theories of "counters": counterpublics and counternarratives (out of political philosophy and media studies), counter-mapping and counter-cartography (from geography and activist mapping), counter-storytelling (specifically the technique developed in critical race theory), counter archives and counter memorials (via cultural studies)53. All of these directly acknowledge –and then attempt to counteract– asymmetrical power relationships. They often appropriate hegemonic cultural forms such as the map, the archive, or the memorial and shift them into the service of minoritized groups and subjugated knowledge. Counterdata science appropriates conventional data science practices and uses them to produce knowledge about a phenomena from outside the official counting institutions. Counterdata science is not only countering institutional data (or lack thereof) but also countering hegemonic data science by enacting alternate politics and producing alternate imaginaries. In the case of anti-feminicide data activists, they posit the idea of using data to achieve a more just world free from gender-related violence, and in the process they enact alternate imaginaries of what data science is, who does it, and who it benefits. Meaning – data activists are claiming data science for uses other than wealth hoarding and they are demanding that it operate in the interests of people other than elite, white, settler men from the Global North. In this sense, counterdata science, like data activism more broadly, is a citizenship practice. It is an informatic form of enacting democratic dissent, prompting protest and insisting on political engagement.

For example, in the case of Puerto Rico, Proyecto Matria and Kilómetro Cero developed a collaboration with Carmen and carried their project from data collection through analysis and public distribution. While the police had claimed there were no feminicides in 2018, their report demonstrated that in fact a feminicide occurred every seven days on the island54. When the team published their results in The persistence of indolence report, they framed feminicide as a structural public health problem which was being systematically neglected by the state. The report included background and definitional information, infographics, and narrative and statistical analysis. It demanded that the state's definition of feminicide include transfeminicides, cases under investigation and feminicides relating to drugs and organized crime. The report's six recommendations targeted changes in law and policy (broader definition of feminicide; more regulation for firearms); the implementation of law and policy (more training on gender-related violence for police, courts, and health professionals) as well as media and culture (promote a culture of non-violence)55. On its publication, the report received wide press coverage in English and Spanish, including a feature in the Intercept and television coverage on Telemundo56. In August 2021, the Governor signed Senate Bill PS 130 into law which defines feminicide and transfeminicide as crimes of first-degree murder. The text of the legislation specifically names and quotes The persistence of indolence report as a motivation for the necessity of the law to recognize and define feminicide and transfeminicide57.

Here counterdata science is connected to a feminist politics of refusal which declines to accept a status quo that delivers violence disproportionately to women. This politics also declines to accept a system that makes such violence invisible, interpersonal, and normal. At the same time that counterdata science about feminicide appropriates the methods, tools and evidentiary models of hegemonic data science, it bends and hacks these towards shifting power across multiple domains of the matrix of domination.

Counterdata Caveats

Counterdata science should be regarded as one tactic amongst many in the action repertoires of activists and social movements to effect structural change58. States Paola Maldonado, a member of the Alliance for Mapping Feminicide from Civil Society in Ecuador, "We play a minimal role in this process, which is basically a defense, a fight for rights and for a dignified life for women and for the eradication of violence, but this is collective work, networked work, work that it has to continue expanding upwards, downwards, towards all the edges that we can give it."59 Not all groups working on feminicide and gender-related violence work with data. Many grassroots groups focus on family or survivor support, accompany families through the justice system, develop programs focused on prevention, or utilize legal strategies to push for policy change. Even among those data activists who do produce counterdata there is variation. For some, like Carmen and many individually-led efforts, the production and circulation of datasets is the sole focus and goal, and they are happy when others use their data (with consent) for media reports, policy change or to secure grants for prevention programs. But for other monitoring groups, particularly those with more people and financial resources, the production of data is one tactical aspect of their larger strategic efforts to "visibilize" feminicide and gender-related violence and hold institutions accountable. For example, while Sovereign Bodies Institute began as a research and data production effort, it now provides direct services to families, produces reports to influence policy makers and engages in lobbying efforts around MMIWG2. In addition to running a national feminicide observatory, the Red Feminista Antimilitarista in Colombia also works in the space of prevention and organizes community protection circles for women facing intimate partner violence. For these groups, counterdata play a supporting role in a larger action repertoire of social change tactics. It is important to underline that in none of these cases – absolutely zero – do activists think that more data alone can lead to social change.

Indeed, there are certainly plenty of reasons to be skeptical of counting and counterdata production. These come from the data activists themselves, many of whom have told our research team "no somos números" – "we [women] are not numbers" – and asserted that a fundamental tension of this work is that producing data about feminicide should be someone else's job. Skepticism also comes from scholars. In 2015, legal scholar Sarah Deer – citizen of the Muscogee (Creek) Nation of Oklahoma – published an award-winning book: The Beginning and End of Rape: Confronting Sexual Violence in Native America. The first chapter reflects at length on the benefits and drawbacks of data. Deer outlines what is known, such as federal numbers that one in three Native women will be raped in their lifetime, and what is simultaneously known and unknown, such as grassroots advocates' assertions that the actual prevalence of rape is much, much higher. While she affirms the importance of numbers and statistics about rape as rhetorical devices that have helped usher in reforms in federal law, she raises questions about fixating on numbers. She asks a provocative question: "But do we need more data in order to move forward?" Her answer is clear, "A continued emphasis on the aggregate data about the rate of rape committed against Native women may serve to eclipse long-term victim-centered solutions."60 At what point does a fixation on data just divert resources and attention from taking action? How much evidence is "enough" and for whom? Especially since Native women, and two spirit people, and Indigenous advocates already understand the scope and scale of the problem – because they live it.

Feminist sociologists also offer a powerful critique of the way in which reducing violence against women into counts, ratios and indicators – particularly at the global scale – works against a rich understanding of the circumstances of violence. In her book, The Seduction of Quantification, Sally Merry compares four frameworks for producing global indicators about gender violence and finds that because each is based on a different conceptual background, each would lead to very different policy actions and "...none of these approaches is comprehensive. Each one is insufficient without detailed qualitative studies that reveal the social and cultural context of the violence."61 Saide Mobayed draws from Merry's work to elucidate the situation about feminicide data in Mexico, specifically, and how the War on Drugs has led to more public feminicides and more feminicides with firearms. This narrative is only visible when combining government data with activist data like that of María Salguero and Data Cívica62. Thus, Mobayed argues that data about feminicide should not be viewed as micro versus macro scales but rather as an "interconnected web."63

Walklate and her colleagues share Merry's concerns about decontextualization, and outline further risks of counting femicide specifically. They assert that a serious risk of creating counts is that the available data are so poor, with so many femicides left out – especially those of women at the intersection of forces of domination – that they may possibly do harm by appearing to be objective while obscuring the deeply underreported reality. A paper by statistician Patrick Ball aligns with this argument. An expert in using statistical methods to determine the scope of wide-scale human rights violations, Ball challenges the use of convenience data – data collected from what is readily available such as press reports rather than a representative sample. "After more than 20 years creating statistics for human rights analysis," he writes, "I have come to believe that descriptive statistics of convenience samples are worse than no statistics." In his experience, decision-makers will give more weight to deeply flawed numbers, often bypassing grounded, qualitative accounts from insiders, no matter how many caveats and disclaimers the statisticians themselves may make about those numbers64.

Finally, Os Keyes, reflecting on the utility of data science for queer and trans people, finds fundamental dissonance: "trans existences are built around fluidity, contextuality, and autonomy, and administrative systems are fundamentally opposed to that. Attempts to negotiate and compromise with those systems (and the state that oversees them) tend to just legitimize the state, while leaving the most vulnerable among us out in the cold." In other words, helping the state better understand and represent queer and trans lives might help the state better target and oppress queer and trans people65. And, moreover, participating in the datafication of trans lives risks legitimizing the very knowledge methods (quantification, counting) that have led to structural violence in the first place. This represents a kind of epistemological complicity with the matrix of domination. As Walklate and colleagues write, "When we focus our attention and activism on counting the killing of women, we then become part of these knowledge projects that have so successfully maintained the invisibility of gendered violence of all types."66

What are the implications of these critiques of counting? Are efforts at reforming official data collection – like #NiUnaMenos's demand for a federal registry of feminicide in Argentina or The persistence of indolence report's call for state accountability – fundamentally flawed? Rather than resolving these tensions, I invite us to hold them close and carry them forward. I offer them because they are important reasons to pause and resist the linear liberal "solution" to missing data, which is to go out and collect more data, fill the gaps, and tweak the counting methodology.

Following the data feminism principle of rethinking binaries and hierarchies, this book will continue both to work with and to trouble the interrelated concepts of official data, missing data and counterdata. Not all data that society does not have is missing. As I stated previously, "missing data" is constituted by the political demand that such data should exist. The political demand comes from specific groups and is directed towards specific institutions. Naming official feminicide data as missing – as virtually all activism, reports, guides, and policy recommendations about the topic do – is in line with feminist efforts to tie such violence back to the state's involvement in enabling human rights violations against significant groups of its own citizens and residents. "Feminicide is a state crime," Lagarde y de los Ríos states plainly67. But Margaret Pearce, Indigenous cartographer and scholar and member of the Citizen Band Potawatomi, makes the point that sometimes data are withheld and sometimes data are refused68. Sometimes dominant groups and their institutions do not deserve to know things about others because they have proved themselves unworthy of that knowledge. Some knowledge is sacred. Some communities must actively organize to obfuscate institutional knowledge about themselves because they are so profoundly targeted by those same institutions69. Not all data gaps should be filled. This brings us back to the who questions that Lauren and I referenced often in Data Feminism: Who collects data? About whom? Who benefits (and who is overlooked or actively harmed)? Whose values guide the process?70 It is not a given that counterdata science is emancipatory. It is not a given that counterdata production will lead to desirable social change (or any change at all). And it is not a given that producing counterdata will not, in fact, harm communities it may intend to help. Navigating these questions is crucial for people who are exploring the utility of counterdata science for their efforts to shift power, and for this reason I included the final chapter of this book, A Toolkit fo Counterdata Science, which supports teams to ask and answer hard questions about their current or future counterdata science projects.

What is a given is that more than 150 grassroots individuals and groups across the Americas have turned to counterdata science to document and challenge feminicide and fatal gender-related violence. As we learn more about their work, I invite us to think about how and whether these caveats about counterdata may shift and change (or not) when the people doing the counting are not state agencies but impacted communities themselves.

Who are the anti-feminicide data activists?

Anti-feminicide data activism has been growing in the past decades, and especially since the 2015 #NiUnaMenos mobilizations. One of the first questions that the Data Against Feminicide team had as we embarked on our project in 2019 was: How widespread are practices of counterdata production about feminicide and fatal gender-related violence in the Americas? From the experiences of Helena Suárez Val, co-lead on the project and a data activist herself who runs Feminicidio Uruguay, we knew there efforts in numerous Latin American countries. Helena had been maintaining a list of anti-feminicide mapping projects on her own website and had interviewed several other mapeadoras for her research71. We initially started a spreadsheet with that list. As our team read research papers or hosted community events, each time we came across a mention of a group producing data about feminicide or fatal gender-related violence, we did additional research about them and added them. Thanks to dedicated researchers on the team, especially Angeles Martinez Cuba, this informal spreadsheet became more systematized and now includes more than 150 efforts from around the world. It includes individuals and groups who monitor feminicide, femicide, MMIWG2, LGBTQ+ killings, Black women killed in police violence and other forms of fatal gender-related violence. The focus is on those efforts that produce data about feminicide. There are many more projects that use data about feminicide in a secondary way (e.g. for social media campaigns, journalism stories, protests, artworks, drafting policy, supporting families), but these efforts were not included unless they specifically engage in data production themselves. And since our focus was on data activism, we did not include government-led efforts on this list.

Our area of interest in the Data Against Feminicide project is the Americas – much of the reading and community-building we have been doing is in this region; we speak many of the languages prevalent in this region – and so the list's geographic patterns reflect that focus. It would not be appropriate, for example, to use our list to make geographic comparisons like "there is more feminicide data activism in Latin America versus Asia" since we did not equally scan all world regions. And yet, it is striking that there are so many efforts globally – at minimum 150 (and there are certainly many more which are unknown to us) – which are producing grassroots data about feminicide and fatal gender-related violence.

Infographic titled ‘WHO ARE THE ANTI-FEMINICIDE DATA-ACTIVISTS’ with a paragraph below the title that reads ‘“in 2019, our Data Against Feminicide team asked; how widespread are practices of counter-data production about feminicide and fatal gender-related violence is the Americas?  We began  a spreadsheet that now includes more than 150 efforts  from around the world (and there are certainly many more which are unknown to us). It includes individuals and groups who monitor feminicide, femicide, MMIWG2,  LGBTQ+ killings, Black women killed in police violence,  and other forms of fatal gender-related violence.”

 Below the paragraph are three adjacent pie-charts. The first one is titled ‘World Region of Data Activists’ and shows a breakdown of 95 Latin American and Caribbean, 23 Europe, 16 US and Canada, 9 Asia, 5 Africa and 1 Australia and Oceania. The second is titled ‘Sector of Data Activism Projects’ and shows a breakdown of  55 ‘Feminist and/or Political Collective,’ 41 ‘Nonprofit’, 28 ‘Data Journalism’, 21 ‘Individual’, and 7 ‘Academia’. The third pie-chart is titled ‘Status of Violence-Monitoring Projects’ and shows 111 as ‘Active’ and 41 as ‘Inactive/Ended’.

Below the pie-charts is a bar-graph titled ‘Types of Gender-related Violence Monitored by Data Activists’ showing their counts; with 69 as ‘Feminicide’, 53 as ‘Femicide’, 13 as ‘Trans Killing’, 11 as Missing Women, 7 as ‘LGBTQ+ Killing’, 6 as Black or Afro-descendent feminicide, 6 as ‘LGBTQ+ hate-crimes’, 6 as ‘MMIG2S/MMIP’, 4 as ‘Transgender hate crimes’and 1 as ‘Police-killings’.
Figure 2.3 This infographic characterizes the data activists that the Data Against Feminicide team has cataloged over the course of three years. It is important to note that this is a convenience sample and should not be regarded as comprehensive. For example, it reflects our project's geographic focus on the Americas and not on Asia, Africa or Oceania, where there are most certainly more groups doing this work than we have cataloged here. Courtesy of the author. Analysis by Angelez Martinez Cuba. Infographic by Melissa Teng.

The infographic in figure 2.3 broadly characterizes this set of counterdata groups and projects. The vast majority of what we cataloged monitor femicide or feminicide in the Americas, with more than 150 total projects that we have encountered in Latin America and the Caribbean. We cataloged at least one feminicide monitoring project in 26 out of 35 total countries in the Americas. Most counterdata projects emerge from either feminist collectives, political collectives or nonprofit organizations, but there are significant numbers of efforts undertaken by individuals as well as a smaller number of projects coming from academia and journalism. Not all of the counterdata projects we cataloged are active and on-going monitoring efforts. Some efforts – like the The persistence of indolence report in Puerto Rico, or the Uma por Uma data journalism project out of Brazil, or the Urban Indian Health Institute's 2018 research study on Missing and Murdered Indigenous Women and Girls in the US – had a specific time period and have ended72. Others, like the Columbian Observatory of Feminicide, operate as "observatories" and continuously monitor cases of fatal gender-related violence in a specific geography73. Here the case of Puerto Rico is interesting - after the publication of The persistence of indolence report, a group of eight feminist organizations came together to form an on-going coalition called the Observatory for Gender Equity Puerto Rico74. This group continues to work closely with Carmen, drawing from her data to monitor feminicides in an on-going way, as well as collecting and publishing data about other gender equity concerns.

In addition to scanning the landscape of anti-feminicide data activism, the Data Against Feminicide team also had questions about activists' motivations and data practices: Why do groups and individuals begin their monitoring projects? What sources do they use for information? How do they classify and categorize cases? How do they publish and circulate their data and with what impacts and effects? What challenges do they face in collecting, analyzing and using the data they produce? To answer these questions, we conducted interviews with data activists based mainly in the Americas starting in 2020. We began with ten interviews, but it quickly became apparent that this form of data activism was more common than we had previously understood, so we continued interviewing diverse individuals and groups. To date, our research team has conducted in-depth interviews with the 35 grassroots data activism efforts listed in Appendix 1.

We went in seeking answers to our questions, yet as we analyzed these interviews four themes emerged that were common across the varied contexts and geographies in which data activists were working75. This is the descriptive model you see in figure 2.4, called "Anatomy of a Feminicide Counterdata Science Project." You might immediately notice some significant differences between this process model and the typical diagrams that circulate about the process of doing data science. If you go out on the web and search images for "data science process" you will find lots of diagrams and charts with variations on linear processes like "Collect > Clean > Explore > Analyze > Visualize." These models, while they can be helpful for teaching and I have certainly used them in technical classes and workshops, are too abstract, too generalized and too technocratic to describe the work of data activists. Counterdata science involving feminicide data takes a fundamentally more specific and grounded form. It starts with an individual or group resolving to take action on the issue of feminicide and developing a theory of change that involves data; then a systematic process of researching various official and other information sources to find cases; then recording cases through an information extraction and categorization process; and finally, multiple goals for reaching diverse audiences with data. Power and inequality permeate all stages of this process and anti-feminicide data activists are creative, careful and rigorous in navigating such ethical quandaries. They also articulate many unresolved tensions. You can think about figure 2.4 as a wayfinding guide, since the next four chapters will explore each of these workflow stages in detail and analyze how grassroots data activist practices depart from those of hegemonic data science.

Graphic indicating the ‘The Anatomy of a Feminicide Counterdata Science Project’. It is visually divided into four panels to represent the processes of Resolving, Researching, Recording, Refusing + Using. It is decorated with tonal reds and other images of activists, maps, infographics, media headlines about feminicide, and protests (previously in this book). 

The Resolving panel has a subheading ‘Starting a Monitoring Effort’ and reads "Background and motivation for why activists start a database of cases. Their theory of power and conceptual influences; their framing of the problem; how they encountered missing data; why they believe counting feminicides or gender-related killings may help to challenge the problem"

The Researching panel has a subheading ‘Finding + verifying information’  and reads "Activists seek relevant information to add to their database. This can include sourcing existing data-sets, mining media and other sources of information, and triangulating across sources to verify details. Such research either discovers new cases or adds information to existing cases in the database."

The Recording panel has subheading ‘Information extraction + classification’ and reads “Activists transform unstructured data from various sources into structured datasets located in databases, spreadsheets and/or text documents. They classify cases according to diverse typologies. They manage data, including ethics, access and governance of the database. 

The Refusing + Using panel has a subheading ‘Where data go, who uses them’ and list the following: Reform/Working with state to formulate new laws, policies and practices. 
Remember/Memorializing killed people, Grieving in public
Revolt/Protesting, mobilizing, performing resistance in public space
Reframe/Story-telling that challenges stigma, Reframing violence as structural
Repair/Supporting families and communities who have lost beloved members
Figure 2.4 The Anatomy of a Feminicide Counterdata Science Project is a process model that describes the high-level workflow stages that data activists engage in to monitor feminicide. The graphics highlight individual and groups of data activists, as well as the multidisciplinary counterdata science methods used in social movements across the Americas, including #SayHerName, #NiUnaMenos, and #MMIWG2.

Pictured include: Protesters in Buenos Aires, 2015, with signs reading “HARTAS” (“FED UP”) and the city’s Palacio del Congreso lit purple for the now-annual and multi-city #NiUnaMenos public demonstration; María Salguero’s map of feminicides in Mexico (2016–present), viewed on Google Earth; Carmen Castelló, a retired social worker who produces the most reliable data on feminicide in Puerto Rico while she cares for her grand-niece Alba and other young relatives; recent news headlines of reported feminicides from around the Americas in Spanish, English, Portuguese, French, Quechua; a database by Dawn Wilcox from Women Count USA, who spends many hours seeking photos of killed women for it; a map of counterdata activist groups created by the Data Against Feminicide initiative, to support and sustain the practices of activists caring for femicide data in their own contexts; one of many installations of The REDress Project (2010–present) by artist Jaime Black, with each dress representing one of the thousands of missing or murdered Aboriginal women across Canada; the music video for “Say Her Name (Hell You Talmbout),” commissioned by the African American Policy Forum and performed by Janelle Monáe and Black women activists, to raise awareness about the Black Women victims of police brutality and anti-Black violence in the U.S.; one of many installations of Zapatos Rojos (2009–Present) by artist Elina Chauvet, where each shoe represents one victim of feminicide, started in Cuidad Juárez, Mexico; and pink crosses, which started in Juárez as a symbol of justice for victims—many of whom work in low-wage, dangerous factory jobs—but have now transcended borders.

Credits: Courtesy of La Nación. Photo by Fabian Marelli. [pending] Courtesy of María Salguero. Courtesy of Dawn Wilcox / Women Count USA: Femicide Accountability Project. Courtesy of Ana María Abruña Reyes / Todas / todaspr.com. Courtesy of Jaime Black. Photo by J. Addington. Courtesy of African American Policy Forum. Song performed by Janelle Monaé feat. various artists. Video by 351 Studio. [pending]

Graphic design by Melissa Q. Teng.

Conclusion

This chapter makes the case that missing data and counterdata are central concepts for understanding data activism about feminicide. While there is a good amount of literature on both concepts, it comes from a variety of fields and case studies. Here I have reviewed that literature as well as formulated precise definitions that can serve us as we consider anti-feminicide data activism. In particular, missing data as a concept is relational, contextual and political. Activists, statisticians and theorists of ignorance would point out that there are structural patterns to missing data about feminicide – data and other forms of knowledge about feminicide and gender-related violence are actively neglected, silenced, and suppressed.

Yet missing data do not exist outside the political demand that they should exist. In the case of feminicide, these demands come from activist and advocacy groups. When they are not fulfilled (which is most of the time) activists across the Americas are taking action to produce counterdata – resistant records that attempt to tally cases of feminicide and fatal gender-related violence. Moreover, data activists do more than count, as is evidenced by the work of advocates in Puerto Rico. They verify and triangulate information; they produce analyses; and they circulate results in a variety of forms to challenge power across the matrix of domination. These practices constitute a counterdata science that mounts an explicit and political challenge to the data practices of state agencies and other "counting institutions". As we will see as we go deeper into the data practices of anti-feminicide groups in the following chapters, their counterdata science practices challenge hegemonic data science practices and invite us to imagine a data science grounded in care, memory justice, rigor and rage.

Comments
144
?
Alessandra Jungs de Almeida:

Catherine, 2 points that maybe you could incorporate from the interviews: 1) using as an example of counterdata production the dossier from I37 and; 2) when you talk about care, rigor and systematization you can incorporate I38’s perspective on rigor when collecting data as a social movement.

?
anonymous:

Note to add Sara’s reference

?
Paola Ricaurte:

Creo que el diagrama se puede explicar en palabras. Ponerlo tan al inicio del capítulo rompe un poco el flujo de la lectura.

?
Paola Ricaurte:

Me parece más una gramática que una anatomía.

?
Paola Ricaurte:

Homologaría esto con la mención de arriba porque decía situado, relacional y político.

?
Paola Ricaurte:

Can you talk a little bit about this Alliance?

?
Paola Ricaurte:

A lo mejor desde antes se podría retomar la idea de visibilidad/invisibilidad, saber/no saber y saber/poder.

?
Paola Ricaurte:

También me suena mucho a Foucault

?
Alessandra Jungs de Almeida:

Yes, I thought the same. There is an interview with Foucault about power, where he says that power only exists because there is resistance, otherwise it would be mere obedience. And when there is resistance, the power relations are reconfigured.

FOUCAULT, Michel. Michel Foucault, uma entrevista: sexo, poder e as políticas da identidade. Verve, v. 5, pp. 260-277, 2004.

+ 1 more...
?
Paola Ricaurte:

Esto también de alguna manera lo había explicado Foucault.

?
Paola Ricaurte:

Aquí yo creo que sería fundamental citar a Rita Segato. Además, la idea de contradatos también está alineada con la idea de Segato de generar contrapedagogías.

?
Paola Ricaurte:

is

?
Paola Ricaurte:

Me parece que es importante discutir críticamente la noción de ignorancia en este contexto. No creo que es políticamente útil hablar de ignorancia y desviarlo hacia la cuestión filosófica y subjetiva. Me parece que es importante hablar de “epistemically produced invisibility”, más que de ignorancia. Es un poco volver a la idea de que no es una cuestión individual. Es una de las estrategias en las que se basa el poder para borrar a lxs otrxs. En este caso es feminicidio, pero se ha hecho desde siempre. Nosotrxs nunca hemos existido en realidad.

Bhattacharyya, S. (2022). Epistemically produced invisibility (pp. 3-14). Minneapolis, MN: University of Minnesota Press.

?
Paola Ricaurte:

Criado

?
Paola Ricaurte:

Proyecto Matria, Kilómetro Cero and

?
Paola Ricaurte:

Podría decir simplemente: has become an expert, para no minimizar el valor de su expertise.

?
Paola Ricaurte:

Parecería como una actitud un poco desde la soberbia, “vamos a revisar si lo que usted hace está bien.” Me sigue aquí faltando contexto sobre las organizaciones: ¿A qué se dedicaban? ¿Qué antigüedad tienen? Incluso quién las financia. No sé si sea necesario tanto, es solamente que tal como aparece aquí no hay contexto sobre la colaboración a partir del trabajo de Carmen. ¿Cuál fue el acuerdo que hicieron?¿La apoyaron luego con su trabajo?

?
Paola Ricaurte:

Me parece que ignorar usado aquí como una de las posibilidades, demerita el hecho de que es una ignorancia producida. Es decir, no es que ignoren sino que deliberadamente deciden ignorar. Coincido en eso con el comentario anterior. Creo que valdría la pena que ordenar los términos de acuerdo a cómo opera el sistema: la inacción es en el sentido de no producir información, no buscar justicia, legitimar la impunidad. Las ausencias que son producidas por una deliberada voluntad de no saber, que si lo quieres poner en términos más de Foucault, está asociada con la voluntad de verdad. Es decir, se construye una verdad oficial a partir de la ausencia, a partir del poder que ellos ejercen con el no saber producido de la ciudadanía.

?
Paola Ricaurte:

Para contexto, añadiría a qué están dedicadas las organizaciones.

?
Saide MV:

Love this!

?
Saide MV:

Love these!!

?
Saide MV:

How did you collectively coded these data? How was that process? How many people were involved? Which infrastructures supported the emergence and development of this project?

?
Saide MV:

Just saw footnote 75! :)

?
Saide MV:

When you say projects, what do you mean? Do you touch on the platforms they use to communicate their work (i.e., social media vs Ushahidi), is this list publicly available?

?
Saide MV:

Yes, totally!

?
Saide MV:

And, perhaps, closeness could mean to further question, dissect and trace who is making what and how? for what reasons?

?
Saide MV:

I don’t know if you have more space / time, but there is a huge body of work Sally Merry draws from which are seminal in their critiques of statistics / enumeration / numbering coming from history, STS, sociology of quantification, and others:

Desrosières A (1998) The Politics of Large Numbers: A History of Statistical Reasoning. Cambridge, Mass: Harvard University Press.
*He was basically a pioneer on that (drawing on Foucault).

Hacking I (1982) Biopower and the avalanche of printed numbers. 5: 279–295. Available at: https://s3.amazonaws.com/arena-attachments/778687/622e0ba69d28d9ff4049b1bc81462079.pdf (accessed 30 May 2020).

Porter TM (1995) Trust in Numbers: The Pursuit of Objectivity in Science and Public Life. Princeton, N.J: Princeton University Press.

Other more contemporary:

Bruno I, Jany-Catrice F and Touchelay B (eds) (2016) The Social Sciences of Quantification: From Politics of Large Numbers to Target-Driven Policies. Logic, Argumentation & Reasoning. Cham: Springer International Publishing. DOI: 10.1007/978-3-319-44000-2.

Guyer JI, Khan N, Obarrio J, et al. (2010) Introduction: Number as Inventive Frontier. Anthropological Theory 10(1–2). SAGE Publications: 36–61. DOI: 10.1177/1463499610365388.

Lippert I and Verran H (2018) After Numbers? Innovations in Science and Technology Studies’ Analytics of Numbers and Numbering. Science & Technology Studies 31(4). 4: 2–12. DOI: 10.23987/sts.76416.

Popp Berman E and Hirschman D (2018) The Sociology of Quantification: Where Are We Now? Contemporary Sociology 47(3). SAGE Publications Inc: 257–266. DOI: 10.1177/0094306118767649.

?
Saide MV:

Diane Nelson’s work might be helpful here too:

Nelson DM (2015) Who Counts? The Mathematics of Death and Life after Genocide. Durham: Duke University Press.

?
Saide MV:

Wow this resonates so much to the ‘global’ feminicide data… so much gets excluded for the sake of rendering comparisons which often makes me think: for what? why do we need to compare feminicide data across countries? who does this benefit and why? for what?

?
Saide MV:

Are you using ‘femicide’ here because that is the term these authors use?

?
Saide MV:

This reminds me so much to this amazing piece by Rinaldo Walcott titled ‘Data or Politics? Why the Answer Still Remains Political’: https://rsc-src.ca/en/covid-19/impact-covid-19-in-racialized-communities/data-or-politics-why-answer-still-remains

?
Saide MV:

Yes, great!

Interestingly, these practices have evolved and grown at the same time as practices of datafication / global logics of quantification.

?
Saide MV:

I wonder whether this report and the data here triangulated included information from the families? That is also a particularity of anti-feminicide data activism: the closeness to the families and their stories.

?
Saide MV:

Wow! I am always amazed at how these practices filter the official narratives of data collection.

?
Saide MV:

Similar to my point above, it would be good to provide your reader with a line or two on how the Puerto Rican state collects feminicide data: how do they define feminicide? How does the Proyecto Matria and Kilómetro Cero define feminicide?

?
Saide MV:

Perhaps one way to strengthen the use of ‘science’ is to go back to its Latin roots ‘scientia’: ‘t originally came from the Latin word scientia which meant knowledge, a knowing, expertness, or experience. By the late 14th century, science meant, in English, collective knowledge’.

https://theconversation.com/the-weighty-history-and-meaning-behind-the-word-science-48280#:~:text=In%20English%2C%20science%20came%20from,%2C%20in%20English%2C%20collective%20knowledge.

?
Saide MV:

Great section!

?
James Scott-Brown:

Something weird is happening with the footnotes: in the text they go up to 75, but the “Footnotes” section of this page only goes up to 9.

Mousing over the superscript 69 in the text shows a tooltip containing the text that is listed next to 63 in the footnote section.

?
James Scott-Brown:

in what sense? I am used to “technocracy” meaning government by (often too narrowly-foused) technical experts.

Do you mean “rigid“?

?
Paola Ricaurte:

+1

?
James Scott-Brown:

“or the” -> “the”

?
James Scott-Brown:

this footnote is not finished

?
James Scott-Brown:

“Paola Maldonado, a member of … states “

?
James Scott-Brown:

“In other words, “

?
James Scott-Brown:

“visions”/”conceptions”?

?
James Scott-Brown:

What determines whether “counter” is followed by a space, hyphen, or neither?

?
James Scott-Brown:

“pointed out“?

?
James Scott-Brown:

“the previous chapter”.

But mentioning the chapter is arguably unnecessary as the first part of the table number is the chapter number.

?
James Scott-Brown:

add ref

?
James Scott-Brown:

Could re-write as:

Feminist philosopher Nancy Tuana applies this line of thinking to knowledge about women's health, and develops a taxonomy of types of ignorance including “willful ignorance” (where missing data and missing knowledge about gender violence would be situated), which she defines as “a systematic process...

?
James Scott-Brown:

“about”?

?
James Scott-Brown:

under-report?

?
James Scott-Brown:

“patriarchy” is a bit vague as an example of a pattern; it’s more of a general cause. A more specific example of the effect of patriarchy on data missingness would be useful here.

Perhaps mention that Chapter 4 contains some detailed discussion of many causes of missingess.

?
James Scott-Brown:

give year of publication rather than saying “recently“

?
James Scott-Brown:

“insist” (or “demand that the government begin to do this work“).

?
James Scott-Brown:

“the final estimate was“ ?

?
James Scott-Brown:

Could link to discussion of Big Tech in chapter 7

?
James Scott-Brown:

“As a case”/”As each case” ?

?
Saide MV:

Here’s some literature you might find helpful in this regard:

https://revistas.uniandes.edu.co/doi/epdf/10.7440/antipoda20.2014.05

Tiscareño-García E and Miranda-Villanueva O-M (2020) Victims and perpetrators of feminicide in the language of the Mexican written press. Comunicar 28(63): 51–60. DOI: 10.3916/C63-2020-05.

?
Saide MV:

This reminds me of the case of Sara Everard vis-a-vis that of Sabina Nessa in the UK in 2021.

https://bust.com/feminism/198576-sarah-everard-sabina-nessa-media-racism.html

?
Saide MV:

And it also hinders prevention. Media could be a great venue to strengthen prevention…

?
Saide MV:

This excellent paper might be interesting for you as well:

Fuentes L (2020) “The Garbage of Society”: Disposable Women and the Socio-Spatial Scripts of Femicide in Guatemala. Antipode 52(6): 1667–1687. DOI: https://doi.org/10.1111/anti.12669.

?
Saide MV:

Interestingly, in the case of Mexico, media (and ‘nota roja’ in particular) plays this sort of dual role where, on the one hand, utterly revictimises women, and, not the other hand, provides activists with loads of valuable data. 

?
Saide MV:

Do you have any info on Puerto Rico’s unreported crimes? (Dark number’ (‘cifra negra’))? Or on the amount of crimes that do not reach any justice or accountability? 

I ask because this might help the reader better situate the conditions of impunity and corruption in Puerto Rico. In Mexico both, the ‘cifra negra’ and the unsolved crimes are over 90% (according to ‘Impunidad Cero’ ‘la probabilidad de que un delito cometido sea resuelto en nuestro país es tan solo de 0.9%’ which is just bonkers)… this certainly informs the way in which civil society contests, counters, and queries official narratives…

?
Saide MV:

Yes, definitely. This goes back to Lagarde’s stressing of the role of the State in the omission and collusion of feminicide. I believe this is a particularity in LATAM which does not necessarily bodes well with the ways in which feminicide has been understood or theorised in countries with more stable / accountable democracies.

?
Saide MV:

Yep, very similar to the Mexican case… also, there is no communication between different institutions responsible for documenting such data.

ILDA came up with similar findings in their project.

?
Saide MV:

In the case of Mexico, the challenge is not only about the lack or adequate legal definitions but the operationalisation of such definitions. Translating these into the criminal / statistical system has proven more complicated than expected…

?
Saide MV:

Or worse, definitions widely vary within the same country, obfuscating national comparisons (like in Mexico).

?
Saide MV:

Yes, excellent questions!

?
Saide MV:

Maybe it would be good to specify who is ignoring what, which in this case is the state/official authorities, right? I think about very local and situated examples where gender-based violence is constantly embodied. I think that rather than ignoring the problem they don’t dare themselves know.

This also connects me to the legitimacy of knowing: who gets to say what knowledge is?

Just some thoughts here!

?
Saide MV:

And also sociologists of knowledge ;).

?
Saide MV:

Other literature that might be interesting for you related to missing data:

López Ruiz D (2018) ‘Desapariciones: usos locales, circulaciones globales’, de Gabriel Gatti (ed.). Kamchatka. Revista de análisis cultural. (12): 564. DOI: 10.7203/KAM.12.11039.

Edkins J (2016) Missing Migrants and the Politics of Naming: Names Without Bodies, Bodies Without Names. Social Research 83(2). The New School: 359–389. Available at: https://www.jstor.org/stable/44282192 (accessed 24 March 2022).

Dawkins S (2020) The problem of the missing dead. Journal of Peace Research. SAGE Publications Ltd: 0022343320962159. DOI: 10.1177/0022343320962159.

?
Saide MV:

Perhaps you could add something related to the challenges of operationalising the ‘gender’ element in feminicide data?

This paper might be interesting:
Fuentes L and Cookson TP (2020) Counting gender (in)equality? a feminist geographical critique of the ‘gender data revolution’. Gender, Place & Culture 27(6): 881–902. DOI: 10.1080/0966369X.2019.1681371.

?
Saide MV:

Reminds me of this paper: Dawson M and Carrigan M (2020) Identifying femicide locally and globally: Understanding the utility and accessibility of sex/gender-related motives and indicators. Current Sociology: 001139212094635. DOI: 10.1177/0011392120946359.

?
Saide MV:

Are these coming more from a statistical framework (as in, quantitative approaches to missing data?)

?
Saide MV:

Something I have been thinking about in my own work and which might be important to address is that many do not call themselves ‘data activists’ or identify with that label.

Catherine D'Ignazio:

I have also been wrestling with this! Sometimes I say “groups” but then not all are groups. And then sometimes I say “civil society actors” but that sounds needlessly complicated…

?
Saide MV:

This is a fantastic case study! It might be worth adding a sentence or two about how the state typifies / quantifies feminicide in Puerto Rico and whether there is any open governmental data on this as it might shed a light on further complexities / frictions of these triangulations. 

?
Saide MV:

Wow! Super interesting these forms of citizen-led necroresistance.

?
Saide MV:

Do you know if the cause of death was also publicly available? Or what kind of variables are included in these open government data? In that regard, it might be worth adding how the state typifies feminicide.

?
Saide MV:

This sentence sounds a bit odd…

?
Saide MV:

This is a wonderful photo of Carmen and her work (posting here because I could not add it to the photo’s description).

?
Saide MV:

It is also fascinating how ‘artisanal’ (in many of the activists’ words) their work is…

?
Saide MV:

Perhaps a different word than ‘information’ here?

?
Saide MV:

I wonder whether you know if she also follows printed press? I know of many local newspapers in Mexico that are not digitalised.

?
Saide MV:

The resonance with activists in other contexts is fascinating!

Interestingly, nobody I know here in Mexico doing data activism against feminicide comes from that social work/nursing background….

?
Saide MV:

Known by who? By the authorities? I ask this because communities often know about these cases.

?
Saide MV:

Perhaps you could specify which so readers know what they were comparing with official police records.

?
Mimi O:

What’s interesting is that missing data can also create activist and advocacy groups, as is evidenced in the example of Carmen at the beginning of this chapter. Or perhaps it’s more accurate to say that the realization of what is not being collected creates the demand for it to exist, which turns citizens and residents into activists and advocates who resolve to collect it themselves.

?
Mimi O:

yes to all of these

?
Mimi O:

This section seems to focus more on the news media than on culture and mass media, which I suspect is not because the news media is worse at creating harmful stereotypes than the other fields but because the the artifacts produced by the news media are easier to track and scrutinize.

With that in mind, I’m not sure if it’s worth adding something about how news media representations are intertwined with other harmful cultural representations of domestic violence and feminicide? Media springs from culture and affects it, and vice versa.

?
Mimi O:

change to colon

?
Mimi O:

The knowing production of this information? Or the production of this information, whether intentionally or not?

?
Mimi O:

Would be helpful to include a quick sentence about why triangulating gives a fuller picture in cases of multiple data sources.

?
Mimi O:

I wonder if it’s worth including in that section examples of when counterdata has been incorporated into state data and the process has produced friction as well.

I say this because the concept of “counter” is about asymmetrical power relations, but it’s important to hold the idea and reality that power relations can shift.

?
Susana Galan:

Add comma

?
Susana Galan:

“when” to avoid repetitions

?
Susana Galan:

Add comma

?
Susana Galan:

In the caption above: “Pictures include”

?
Susana Galan:

Also “(2009-present)”

?
Susana Galan:

Replace by “:”

?
Susana Galan:

Add “were”

?
Susana Galan:

Seductions

?
Susana Galan:

“to”

?
Susana Galan:

“Deer”

?
Susana Galan:

“calculations”?

?
James Scott-Brown:

“statistics”?

?
Susana Galan:

“according to which”

?
Susana Galan:

I would flesh out these arguments, which constitute two separate (while both valid and important) criticisms. The first one refers to the reduction of the murdered women (with their complex identities and rich lives) to numbers, which erases their individuality as well as the multiple relationships they were part of. The reduction of women to numbers has the ultimate effect of negating their humanity (in line with how the use of numbers to identify prisoners in camps, for example, has a dehumanizing effect). The second question, which has been addressed in the text before, refers to state responsibility and accountability and can be linked (while not reduced) to a neoliberal move towards a “lean” state, whereas non-governmental organizations are expected to “fill the gaps” left by the retreating state

?
Susana Galan:

Add “as a motor for change/agent of change by itself”

?
Susana Galan:

Note that in the text you sometimes refer to it as “the The persistence of indolence report” and other times (like here) as just “The persistence of indolence report”. I prefer the latter, but the former may be more correct

?
Susana Galan:

This part should be moved earlier one, when the concept of counterdata science is introduced.

?
Susana Galan:

Add comma

?
Susana Galan:

Add comma

?
Susana Galan:

“operates”

?
Susana Galan:

“that”

?
Susana Galan:

“This means that”

?
Susana Galan:

“a phenomenon” or “phenomena”

?
Susana Galan:

Consider using em dash at the beginning and the end

?
Susana Galan:

“as well as”

?
Susana Galan:

Why “science”? What does “science” bring to the argument? Is it due to the identification of science with rigor and systematization? Would there be a less-fraught term that similarly evokes these ideas while also centering care?

?
Susana Galan:

Add comma

?
Susana Galan:

I wouldnuse “ignorance” here, in line with the previous elaboration

?
Susana Galan:

“The extensive media coverage of her disappearance drew criticism”

?
Susana Galan:

“A case in point was Gabby Petito”

?
Susana Galan:

Delete comma

?
Susana Galan:

“it is well documented that murders and disappearances of white women in the US receive more media attention than those of women of color (cite references). This phenomenon has even received a name, ‘Missing White Woman Syndrome’, coined by Black journalist Geen Ifillmin 2004”

?
Susana Galan:

Change to comma

?
Susana Galan:

Add comma

?
Susana Galan:

Add comma

?
Susana Galan:

Should not be capitalized

?
Susana Galan:

“step towards”

?
Susana Galan:

Delete comma

?
Susana Galan:

Same as before. It seems that the stage of identification/realization (becoming aware of the presence of that absence) is a precondition for the political demand and an important part of the process

?
Susana Galan:

I would say that “data are missing because specific groups/individuals have identified/recognized them as such (missing data) and have demanded that they should exist and be made available”

?
Susana Galan:

“try to simultaneously fight … and deny” or “try to fight … while at the same time denying”

?
Susana Galan:

Add “. This category is described/defined as”

?
Susana Galan:

Did they identify any particular patterns regarding the cases that were only in news reports or only in the demographic registry (with regard to, i.e., class, sexual orientation, gender identity, age, etc.)?

?
Susana Galan:

“Up-to-date”

?
Susana Galan:

“. However, project members wanted to check/confirm the thoroughness of Carmen's media monitoring in order to …, so they dedicated a staff member”

?
Susana Galan:

How does this work? Is it through public information or informal contacts with people working in the system?

?
Susana Galan:

How did they know about Carmen? Were they already familiar with her work?

?
Susana Galan:

“lesser-known”

?
Susana Galan:

“hears of”?

?
Susana Galan:

How is not-knowing conceptualized differently than ignorance here? Is it intentional (not-wanting-to-know)?

?
Susana Galan:

Add “the”. Or add commas before and after the two non-profit organizations”

?
Alexis Henshaw:

Paisley Currah (esp. “Legally Sexed) and Kevin Guyan (“Queer Data”) are also great citations for this paragraph.

?
Alexis Henshaw:

Reference to country should be “Colombian”— even in English. :)

?
Alexis Henshaw:

“for”

?
James Scott-Brown:

The text in this figure is quite small: will it be legible in the book?