Using Wikipedia for conflict forecasting

Curated on Posted on by Stefaan Verhulst

Article by Christian Oswald and Daniel Ohrenhofer: “How can we improve our ability to predict conflicts? Scholars have struggled with this question for a long time. However, as a discipline, and especially over the last two decades, political science has made substantial progress. In general, what we need to improve predictions are advances in data and methodology. Data advances involve both improving the quality of existing data and developing new data sources. We propose a new data source for conflict forecasting efforts: Wikipedia.

The number of country page views indicates international salience of, or interest in, a country. Meanwhile, the number of changes to a country page indicate political controversy between opposing political views.

We took part in the Violence Early-Warning System’s friendly competition to predict changes in battle-related deaths. In our work, we evaluate our findings with out-of-sample predictions using held-out, previously unseen data, and true forecasts into the future. We find support for the predictive power of country page views, whereas we do not for page changes…

Globally available data, updated monthly, are ideal for (near) real-time forecasting. However, many commonly used data sources are available only annually. They are updated once a year, often with considerable delay.

Some of these variables, such as democracy or GDP, tend to be relatively static over time. Furthermore, many data sources face the problem of missing values. These occur when it is not possible to find reliable data for a variable for a given country.

Wikipedia is updated in real time, unlike many commonly used data sources, which may update only annually and with considerable delay

More recent data sources such as Twitter, images or text as data, or mobile phone data, often do not provide global coverage. What’s more, collecting and manipulating data from such sources is typically computationally and/or financially costly. Wikipedia provides an alternative data source that, to some extent, overcomes many of these limitations…(More)”.

Selected Readings on the LGTBQ+ Community and Data

Curated on Posted on by Uma

By Uma Kalkar, Salwa Mansuri, Marine Ragnet and Andrew J. Zahuranec

As part of an ongoing effort to contribute to current topics in data, technology, and governance, The GovLab’s Selected Readings series provides an annotated and curated collection of recommended works on themes such as open data, data collaboration, and civic technology.

Around the world, LGBTQ+ people face exclusion and discrimination that undermines their capacity to live their lives and succeed. Together with allies, many LGBTQ+ people are fighting to exercise their rights and achieve full equality. However, this struggle has been undermined by a lack of specific, quantifiable information on the challenges they face.

When collected and managed responsibly, data about sexual and gender minorities can be used to protect and empower LGBTQ+ people through informed policy and advocacy work. To this end, this Selected Reading investigates what data is (and is not) collected about LGBTQ+ individuals in the areas within healthcare, education, economics, and public policy and the ramifications of these outcomes. It offers a perspective on some of the existing gaps regarding LGBTQ+ data collection. It also examines the various challenges that LGBTQ+ groups have had to overcome through a data lens. While activism and advocacy has increased the visibility and acceptance of sexual and gender minorities and allowed them to better exercise their rights in society, significant inequities remain. Our literature review puts forward some of these recent efforts.

Most of the papers included in this review, however, conclude with similar findings: data for about LGBTQ+ communities is still lacking and as a result, research on the topic is often times also lagging behind. This is particularly problematic, as detailed in some of our readings, because LGBTQ+ populations are often at the center of discrimination and still face disparate health vulnerabilities. The LGBTQI+ Data Inclusion Act, which recently passed the US House of Representatives and would require over 100 federal agencies to improve data collection and surveying of LGBTQ communities, seeks to address this gap.

We hope this selection of readings can provide some clarity on current data-driven research for and about LGBTQ+ individuals. The readings are presented in alphabetical order.

***

Selected Reading List (in alphabetical order)

***

Annotated Selected Reading List (in alphabetical order):

D’Ignazio, Catherine, and Lauren F. Klein. Data Feminism. MIT Press, 2020. https://mitpress.mit.edu/books/data-feminism.

  • D’Ignazio and Klein investigate how data has been historically used to maintain specific social status quos. To overcome this challenge, they approach data collection and uses through an intersectional, feminist lens that identifies issues in current data handling systems and looks toward solutions for more inclusive data applications.
  • The editors define data feminism as “power, about who has it and who doesn’t, and about how those differentials of power can be challenged and changed using data.” The book centers around seven principles that identify and challenge existing power structures around data and seek pluralist, context-based data processes that illuminate hidden and missed data.

Giblon, Rachel, and Greta R. Bauer. “Health care availability, quality, and unmet need: a comparison of transgender and cisgender residents of Ontario, Canada.” BMC Health Services Research 17, no. 1 (2017): 1–10. https://bmchealthservres.biomedcentral.com/articles/10.1186/s12913-017-2226-z.

  • Canada boasts a universal healthcare and insurance system, yet disparities exist between the treatment quality, services, and knowledge about transgender patients.
  • Data collection on transgender, non-binary, and intersex individuals is not conducted in Canadian health surveys, making it difficult to compare and contrast the healthcare provided to transgender people with that provided to cisgender people. Moreover, a lack of physician knowledge about trans needs and/or refusal to provide hormone therapy/ gender-affirming procedures result in trans individuals explicitly avoiding medical services. The lack of services, comfort, and data about transgender people in Canada demonstrate their severely “unmet health care need.”
  • Using data about Ontario residents from the Canadian Community Health Survey and the Trans PULSE survey, the researchers find that 33% transgender Ontarians had an unmet health need that would not be unmet if they were cisgender. As well, transgender men and women found the quality of healthcare in their community to be poor than compared to cisgender individuals. Twenty-one percent of transgender people avoided going to emergency rooms because of their gender identity.

Bowleg, Lisa, and Stewart Landers. “The need for COVID-19 LGBTQ-specific data.” American Journal of Public Health 111, no. 9 (2021): 1604–1605. https://pubmed.ncbi.nlm.nih.gov/34436923/.

  • The adage “no data, no problem” has been magnified during the pandemic, highlighting gaps around data collection for LGBTQ communities, which often intersect with other communities who are disproportionately at-risk for COVID-19, such as minority populations in the service industry and those who smoke.
  • Despite concerns about the stigma facing LBGTQ communities, data collection from these demographics has been relatively feasible, with federal governments drastically increasing their data collection from LGBTQ communities.
  • However, the lack of direction and guidance at a federal level to collect sexual and gender minority data has stunted information about how this demographic has experienced COVID-19 when compared to cis-gender, heterosexual groups. The authors stress the need for data collection from LGBTQ communities and advocacy to encourage these practices to help address the pandemic.

Marshall, Zack, Vivian Welch, Alexa Minichiello, Michelle Swab, Fern Brunger, and Chris Kaposy. “Documenting research with transgender, nonbinary, and other gender diverse (trans) individuals and communities: introducing the global trans research evidence map.” Transgender Health 4, no. 1 (2019): 68–80. https://www.liebertpub.com/doi/10.1089/trgh.2018.0020.

  • Marshall and colleagues study a series of 15 academic databases to assemble a dataset describing 690 trans-focused articles. They then map where and how transgender “have been studied and represented within and across multiple fields of research” to understand the landscape of existing research on transgender people. They find that research around the trans community focused on physical and mental healthcare services and marginalization and were primarily observational research.
  • The authors found that social determinants of health for transgender people were the least studied, along with ethnicity, culture, and race, violence, early life experiences, activism, and education.
  • With this evidence map, researchers have a strong starting point to further explore issues through a LGBTQ lens and better engage with trans people and perspectives when looking at social problems.

Medina, Caroline and Lindsay Mahowald. “Collecting Data about LGBTQI+ and Other Sexual and Gender-Diverse Communities.” Center for American Progress, May 26, 2022. https://www.americanprogress.org/article/collecting-data-about-lgbtqi-and-other-sexual-and-gender-diverse-communities.

  • The paper argues, that despite advances “a persistent lack of routine data collection on sexual orientation, gender identity, and variations in sex characteristics (SOGISC) is still a substantial roadblock for policymakers, researchers, service providers, and advocates seeking to improve the health and well-being of LGBTQI+ people.”
  • Even though various types of data are integral to the experiences of LGBTQI+ people, the report narrows its focus to data collection in two forms of environments: general population surveys & surveys regarding LGBTQI+ people. Specific population surveys such as the latter provide significant advantage to capture specific and sensitive data.
  • It argues that a range of precautions can be adopted from a research design perspective to ensure that personal data and information is handled with care and matches ethical standards as outlined in the Data Ethics Framework of the Federal Data Strategy ranging from privacy and confidentiality to honesty and transparency.

Miner, Michael H., Walter O. Bockting, Rebecca Swinburne Romine, and Sivakumaran Raman. “Conducting internet research with the transgender population: Reaching broad samples and collecting valid data.” Social science computer review 30, no. 2 (2012): 202–211. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3769415/.

  • The internet has the potential to collect information from transgender people, who are “a hard-to-reach, relatively small, and geographically dispersed population” in a diverse and representative manner.
  • To study HIV risk behaviors of transgender individuals in the U.S., Miner et al. developed an online tool that recruited individuals who frequent websites that are important for the transgender community and used quantiative and qualitative methods to learn more about these individuals. They conclude that while online data collection can be difficult to ensure internal validity, careful testing and methods can overcome these issues to improve data quality on transgender people.

Pega, Frank, Sari L. Reisner, Randall L. Sell, and Jaimie F. Veale. “Transgender health: New Zealand’s innovative statistical standard for gender identity.” American journal of public health 107, no. 2 (2017): 217–221. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5227923/.

  • Pega et al. discuss New Zealand’s national statistical standard for gender identity data collection, the first of its kind. More governments in Australia and the United States are now following suit to address the health access and information disparity that transgender people face.
  • Data about transgender people has advanced progressive policy action in New Zealand, and the authors celebrate this statistical standard as a way to collect high quality data for data-driven policies to support these groups.
  • While this move will help uncover LGBTQ individuals currently hidden in data, the authors critique the standard because it does not “promote the two-question method, risking misclassification and undercounts; does promote the use of the ambiguous response category “gender diverse” in standard questions; and is not intersex inclusive.”

Ruberg, Bonnie, and Spencer Ruelos. “Data for Queer Lives: How LGBTQ Gender and Sexuality Identities Challenge Norms of Demographics.” Big Data & Society 7, no. 1 (June 18, 2020): 205395172093328. https://journals.sagepub.com/doi/full/10.1177/2053951720933286.

  • Drawing from the responses of 178 people who identified as non-heterosexual or non-cisgender in a survey, this paper argues that “dominant notions of demographic data, […] that seeks to accurately categorize and “capture” identity do not sufficiently account for the complexities of LGBTQ lives.”
  • Demographic data commonly imagines identity as fixed, singular, and discrete. However, the researchers’ findings suggest that, for LGBTQ people, gender and sexual identities are often multiple and in flux. Most respondents reported their understanding of their identity shifting over time. For many, “gender identity was made up of overlapping factors, including the relationship between gender and transgender identities. These findings challenge researchers to reconsider how identity is understood as and through data.” They argue that considering identities as fixed and discrete are not only exclusionary but also do not wholly represent the dynamic and fluid nature of gender identities.
  • The piece offers several recommendations to address this challenge. Firstly, the researchers argue to remove data discreteness, which will enable users to select multiple identities rather than choose one from a drop-down list. Secondly, create communication and feedback channels for LGBTQ+ to express whether surveys and other data collection methods are sufficiently inclusive and gender-sensitive.

Sell, Randall L. “LGBTQ health surveillance: data = power.” American Journal of Public Health 107, no. 6 (2017): 843–844. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5425894/.

  • Sell recounts his motto: ‘data = power;’ ‘silence = death’ and how LGBTQ people have been victims of this situation. He argues that health research and surveillance has systemically ignored sexual and gender minorities, leading to gaps in administrative understanding and policies for LGBTQ population.
  • He laments that very few surveys on American health collect sexual and gender orientation data, and the lack of standardization around this data collection muddies researchers’ ability to collate and utilize the information meaningfully.
  • He calls for legislation that mandates the National Institutes of Health to include sexual and gender minorities in all publicly funded research similar to the specific inclusion requirement of women and racial and ethnic minorities in studies. Despite concerns about surveillance and targeting of LGBTQ minorities, Sell argues that data collection is imperative now for a long-scale understanding of the needs of the community, transcending political terms.

Snapp, Shannon D., Stephen T. Russell, Mariella Arredondo, and Russell Skiba. “A right to disclose: LGBTQ youth representation in data, science, and policy.” Advances in child development and behavior 50 (2016): 135–159. https://pubmed.ncbi.nlm.nih.gov/26956072/.

  • Despite significant and positive reforms such as the legalization of same-sex marriages and protection from intersectional sexual harrasment (Webb, 2011) in the United States, there is a striking gap in literature on evidence-based practices that support LGBTQ+ Youth (Kosciw & Pizmony-Levy, 2013). The lack of data-driven solutions stifle the creation of inclusive environments where members of the LGBTQI+ community feel heard and seen. There is a striking gap in literature on evidence-based practices that support LGBTQ+ Youth (also see Kosciw & Pizmony-Levy, 2013; Mustanski, 2011).
  • At present federal and local state data-states do not include SOGI (Sexual Oreintation & Gender Identity) in demographic questions. Data sets that do have spaces to disclose SOGI are largely in a health-related setting such as the Centre for Disease Control or Youth Risk Behavior. As such learning and education disparities and outcomes are not accurately measured.
  • Missing systematic SOGI data renders members of the LGBTQ+ community invisible and sidelined. As such several members of civil society have therefore demanded for the need to gather SOGI data in the Department of Health, Education & Justice. Such data is therefore central to holistically encapsulate the discriminatory experiencees LGBTQ+ Youth face in an education setting, integral to well-being and development. Scholars and research teams have thusfar overcome the barriers of data reliability and validity (see Ridolfo, Miller, & Maitland, 2012) by collating the most effective methods for data collection (Sexual Minority Assessment Research Team, 2009).

Wimberly, George L. “Chapter 10: Use of large-scale data sets and LGBTQ education.” LGBTQ issues in education: Advancing a research agenda (2015): 175–218. https://ebooks.aera.net/LGBTQCH10.

  • This book chapter highlighs the importance of large-scale data sets to gain understanding about LGBTQ students, school experiences, and academic achievement.
  • Young people who identify as LGBTQ tend to be generalized and ways that LGBTQ identification questions are asked by surveys change across years, making it important to disaggregate large-scale data for more granular knowledge about LGBTQ people in education.
  • Wimberly provides information about multiple datasets that collect this information, how they ask questions on LGBTQ identity, and ways in which the datasets have been used or have the potential to be leveraged for a more comprehensive understanding of students. He also points out the limitations of existing data sets, namely that they tend to be retrospective of the LGBTQ adolescent experience and collected from convenience samples, such as college students. This limitation also impacts the external validity of the data, especially with regard to rural, racialized, and lower-income LGBTQ students.

Africa: regulate surveillance technologies and personal data

Curated on Posted on by Stefaan Verhulst


Bulelani Jili in Nature: “…For more than a decade, African governments have installed thousands of closed-circuit television (CCTV) cameras and surveillance devices across cities, along with artificial-intelligence (AI) systems for facial recognition and other uses. Such technologies are often part of state-led initiatives to reduce crime rates and strengthen national security against terrorism. For instance, in Uganda in 2019, Kampala’s police force procured digital cameras and facial-recognition technology worth US$126 million to help it address a rise in homicides and kidnappings (see go.nature.com/3nx2tfk).

However, digital surveillance tools also raise privacy concerns. Citizens, academics and activists in Kampala contend that these tools, if linked to malicious spyware and malware programs, could be used to track and target citizens. In August 2019, an investigation by The Wall Street Journal found that Ugandan intelligence officials had used spyware to penetrate encrypted communications from the political opposition leader Bobi Wine1.

Around half of African countries have laws on data protection. But these are often outdated and lack clear enforcement mechanisms and strategies for secure handling of biometric data, including face, fingerprint and voice records. Inspections, safeguards and other standards for monitoring goods and services that use information and communications technology (ICT) are necessary to address cybersecurity and privacy risks.

The African Union has begun efforts to create a continent-wide legislative framework on this topic. As of March this year, only 13 of the 55 member states have ratified its 2014 Convention on Cyber Security and Personal Data Protection; 15 countries must do so before it can take effect. Whereas nations grappling with food insecurity, conflict and inequality might not view cybersecurity as a priority, some, such as Ghana, are keen to address this vulnerability so that they can expand their information societies.

The risks of using surveillance technologies in places with inadequate laws are great, however, particularly in a region with established problems at the intersections of inequality, crime, governance, race, corruption and policing. Without robust checks and balances, I contend, such tools could encourage political repression, particularly in countries with a history of human-rights violations….(More)”.

What Germany’s Lack of Race Data Means During a Pandemic

Curated on Posted on by Stefaan Verhulst

Article by Edna Bonhomme: “What do you think the rate of Covid-19 is for us?” This is the question that many Black people living in Berlin asked me at the beginning of March 2020. The answer: We don’t know. Unlike other countries, notably the United States and the United Kingdom, the German government does not record racial identity information in official documents and statistics. Due to the country’s history with the Holocaust, calling Rasse (race) by its name has long been contested.

To some, data that focuses on race without considering intersecting factors such as class, neighborhood, environment, or genetics rings with furtive deception, because it might fail to encapsulate the multitude of elements that impact well-being. Similarly, some information makes it difficult to categorize a person into one identity: A multiracial person may not wish to choose one racial group, one of many conundrums that complicate the denotation of demographics. There is also the element of trust. If there are reliable statistics that document racial data and health in Germany, what will be done about it, and what does it mean for the government to potentially access, collect, or use this information? As with the history of artificial intelligence, figures often poorly capture the experiences of Black people, or are often misused. Would people have confidence in the German government to prioritize the interests of ethnic or racial minorities and other marginalized groups, specifically with respect to health and medicine?

Nevertheless, the absence of data collection around racial identity may conceal how certain groups might be disproportionately impacted by a malady. Racial self-identities can be a marker for data scientists and public health officials to understand the rates or trends of diseases, whether it’s breast cancer or Covid-19. Race data has been helpful for understanding inequities in many contexts. In the US, statistics on maternal mortality and race have been a portent for exposing how African Americans are disproportionately affected, and have since been a persuasive foundation for shifting behavior, resources, and policy on birthing practices.

In 2020, the educational association Each One Teach One, in partnership with Citizens for Europe, launched The Afrozensus, the first large-scale sociological study on Black people living in Germany, inquiring about employment, housing, and health—part of deepening insight into the ethnic makeup of this group and the institutional discrimination that they might face. Of the 5,000 people that took part in the survey, a little over 70 percent were born in Germany, with the other top four being the United States, Nigeria, Ghana, and Kenya. Germany’s Afro-German population is heterogenous, a reflection of an African diaspora that hails from various migrations, whether it be Fulani people from Senegal or the descendants of slaves from the Americas. “Black,” as an identity, does not and cannot grasp the cultural and linguistic richness that exists among the people who fit into this category, but it may be part of a tableau for gathering shared experiences or systematic inequities.“I think that the Afrozensus didn’t reveal anything that Black people didn’t already know,” said Jeff Kwasi Klein, Project Manager of Each One Teach. “Yes, there is discrimination in all walks of life.” The results from this first attempt at race-based data collection show that ignoring Rasse has not allowed racial minorities to elide prejudice in Germany….(More)”.

The Privacy Elasticity of Behavior: Conceptualization and Application

Curated on Posted on by Stefaan Verhulst

Paper by Inbal Dekel, Rachel Cummings, Ori Heffetz & Katrina Ligett: “We propose and initiate the study of privacy elasticity—the responsiveness of economic variables to small changes in the level of privacy given to participants in an economic system. Individuals rarely experience either full privacy or a complete lack of privacy; we propose to use differential privacy—a computer-science theory increasingly adopted by industry and government—as a standardized means of quantifying continuous privacy changes. The resulting privacy measure implies a privacy-elasticity notion that is portable and comparable across contexts. We demonstrate the feasibility of this approach by estimating the privacy elasticity of public-good contributions in a lab experiment…(More)”.

Selected Readings on the Intersection of Data, Abortion Care, and Women’s Health

Curated on Posted on by Uma

By: Uma Kalkar, Salwa Mansuri, Andrew J. Zahuranec

As part of an ongoing effort to contribute to current topics in data, technology, and governance, The GovLab’s Selected Readings series provides an annotated and curated collection of recommended readings on themes such as open data, data collaboration, and civic technology.

In this edition, we reflect on the intersection between data, abortion, and women’s health following the United States Supreme Court ruling regarding Dobbs v. Jackson Women’s Health Organization which held that there was no constitutional right to abortion and decided that individual states have the authority to regulate access to abortion services. In the days before and since the decision, a large amount of literature has been produced both on the implications of this ruling for individuals’ data privacy and the effects on women’s social and economic lives. It is clear that, while opinions on access to abortion services are often influenced by deeply held attitudes about women’s bodily autonomy and when life begins, data has critical importance both as a potential source of risk and as a tool to understand the decision’s impact.

Below we curate some stories from news sources and academic papers on the role of data in abortion services as well as data-driven research by institutions into the effects of abortion. We hope this selection of readings provides a broader perspective on how data and women’s rights and health intersect.

As well, we urge that anyone seeking further information about abortion access visit www.ineedana.com via a secure site, and preferably via a VPN. For those looking for menstrual apps, Spot On by the Planned Parenthood Federation of America saves data locally on phones, does not provide information to third parties, and allows for anonymous accounts.

The readings are presented in alphabetical order.

***

Data & Privacy Concerns

Conti-Cook, Cynthia. “Surveilling the Digital Abortion Diary: A Preview of How Anti-Abortion Prosecutors Will Weaponize Commonly-Used Digital Devices As Criminal Evidence Against Pregnant People and Abortion Providers in a Post-Roe America.” University of Baltimore Law Review, forthcoming. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3666305

  • In this four-part article, Conti-Cook discusses the history of health data rights and the long-standing ways in which digital evidence produced by pregnant people has been used to prosecute their actions. She discusses how digital technologies help prosecutors lay charges against those seeking abortions and how they help “ the state see[k] control over [them] by virtue of their pregnancy status” by digitally surveilling them.
  • The author examines how “digital, biometric, and genetic surveillance” serves as a vehicle to “microtarget” historically oppressed communities” under a patriarchal and racist social structure.
  • She also discusses how online searches relating to pregnancy termination and abortion, location and tracking data, site history, wearable devices, and app data can be factored into risk assessment tools to assess social service outcomes and federal prosecutions.
  • Conti-Cook ends by reviewing digital hygiene strategies to stop the use of personal data against oneself and foster a more critical use of digital tools for reproductive and pregnancy-related health needs.

Diamant, Jeff, and Besheer Mohamed. “What the Data Says about Abortion in the U.S.” Pew Research Center, June 24, 2022. https://www.pewresearch.org/fact-tank/2022/06/24/what-the-data-says-about-abortion-in-the-u-s-2

  • In the aftermath of the overturn of Roe v. Wade (1973), the Pew Research Center published a compilation of facts and statistics about abortion care in the United States obtained through the Centers for Disease Control and Prevention and Guttmacher Institute.
  • The piece describes shifting trends pertaining to the number of legal abortions conducted each year in the United States since the 1970s, the abortion rate among women, the most common types of abortions, and the number of abortion providers over time. It describes, for example, how the procedure has generally declined at “a slow yet steady pace” since the early 1990s. It also notes that the number of providers has declined over time.

Paul, Kari. “Tech Firms under Pressure to Safeguard User Data as Abortion Prosecutions Loom.” The Guardian, June 25, 2022, sec. US news. https://www.theguardian.com/us-news/2022/jun/25/tech-companies-health-data-security-abortion-prosecution

  • Paul writes about the concerns of abortion and civil rights activists on how data collected about individuals through apps and online searches might incriminate those seeking or providing abortion services. It notes how geo-location data used by tech companies can make “it easy for law enforcement officials to access incriminating data on location, internet searches, and communication history.”
  • While period tracking apps have received significant attention, the article notes that companies such as Meta, Uber, Lyft, Google, and Apple have yet to publicly announce how they would respond to law enforcement requests on abortion evidence.
  • The piece finally includes a recommendation from the digital rights advocacy group Electronic Frontier Foundation that companies preemptively prepare “for a future in which they are served with subpoenas and warrants seeking user data to prosecute abortion seekers and providers.” It suggests end-to-end encryption as a default, refraining from collecting location information, and allowing anonymous or pseudonymous access to apps.

Nguyen, Nicole, and Cordilia James. “How Period-Tracker Apps Treat Your Data, and What That Means If Roe v. Wade Is Overturned.” Wall Street Journal, June 21, 2022. https://www.wsj.com/articles/how-period-tracker-apps-treat-your-data-and-what-that-means-if-roe-v-wade-is-overturned-11655561595

  • Nguyen and James provide an extensive analysis of the ways that period tracking apps track, collect, store, and share data about women’s fertility and menstrual cycle. Following Dobbs v. Jackson Women’s Health Organization (2022), which overturned Roe v. Wade (1973), there has been significant public concern about the (re)use of the data these apps collect.
  • They detail different kinds of data that could be subpoenaed from period trackers and the terminology that users can search for in an app’s privacy policy to understand how their data will be used. It describes, for example, what it means to when Terms & Conditions outline how they will “encrypt” (that is, to scramble into an incoherent string of code), “share” or “sell” (data can be given to third parties such as advertisers), and respond to “requests” (companies may notify the user when a court or government data asks for data).
  • The article closes with an overview of the most-downloaded fertility apps — including Flo, Apple Health, Clue, FitBit, Glow, and Natural Cycles — and where they stand on data privacy.

Sherman, Jenna. “How Abortion Misinformation and Disinformation Spread Online.” Scientific American, June 24, 2022. https://www.scientificamerican.com/article/how-abortion-misinformation-and-disinformation-spread-online/

  • In Scientific American, Sherman writes an opinion piece on the growth of online dis- and misinformation in the aftermath of Dobbs. She summarizes how, according to current data-driven research, much of the information people find online about abortion is not reliable and that the highest volume of online searches about abortion tends to be in those states with the most restricted access.
  • Despite much research on abortion, Sherman notes “a lack of access to quality information or care” online, especially for marginalized communities. She also summarizes the results of studies on social media and search engines. In one 2021 study, searches for “abortion pill” tended not to yield scientifically accurate and moderately accessible information.
  • Another study cited in the article found that half of the web pages surfaced by Google on abortion contained misinformation. This appears to be by design — with false information about “abortion pill reversal” and abortion practices generating large revenues for platforms like Facebook.

Data on the Impact of Abortion Access

Amador, Diego. “The Consequences of Abortion and Contraception Policies on Young Women’s Reproductive Choices, Schooling and Labor Supply.” Documento CEDE №2017–43 (2017). https://ssrn.com/abstract=2987367

  • Amador analyzes aggregate provider data from the Guttmacher Institute to assess the relationship between contraceptive use, abortion, schooling, and labor decisions of US women. The dataset follows a sample of women born between 1980 and 1984, with data from interviews starting in 1997 and ending in 2011.
  • A counterfactual model based on the data suggests that a perfectly enforced ban on abortions would raise the rate of standard contraceptive use for women 9.1%. The fraction of children born to single mothers would increase from 30% to 34% while the average amount of schooling after high school would decrease by 3.1%. The number of women with college degrees would drop by 1.8% age points. The estimated average loss in lifetime earnings for women who would have at least had one abortion was estimated at USD 39,172.
  • The author also assesses the impact that free contraception would have, suggesting a 15.7 decrease in pregnancies per 1000 women and an 11.6 reduction in abortions per 1000 women. Accumulated schooling after high school increased by an estimated 3%. An assessment of mandatory counseling laws found that the long-run effect of these laws on women ages 18 to 30 was a 10% decrease in abortion rates.
  • The author concludes that policies such as an abortion ban and free contraception have important effects on schooling and lifetime earnings but only a moderate impact on labor supply.

ANSIRH. “Introduction to the Turnaway Study.” ANSIRH, March 2020. https://www.ansirh.org/sites/default/files/publications/files/turnawaystudyannotatedbibliography.pdf

  • This fact sheet summarizes various analyses stemming from the Turnaway Study, the first study to rigorously examine the effects of receiving abortion services versus being denied access to them. The study is an initiative by Advancing New Standards in Reproductive Health (ANSIRH), a program within the UCSF Bixby Center for Global Reproductive Health. It examines 1,000 women seeking abortion from 30 facilities around the country, with interviews conducted over five years.
  • Studies conducted with the dataset find that the most common reason for women to seek an abortion was not being able to afford a child and/or not having a suitable partner/parent involved to assist with childrearing. Most women don’t feel pressured by counseling that occurs in clinics but find it less helpful when it is state-mandated. Half of all women report seeing anti-abortion protestors at clinics and greater contact with them tends to be more upsetting.
  • Studies also suggest no evidence that abortion causes negative mental health outcomes, although being denied an abortion is associated with elevated anxiety and stress and lower self-esteem. Those who receive an abortion experience “a mix of positive and negative emotions in the days after […] with relief predominating.” The intensity of the emotion diminishes over time but over 95% of women report “abortion was the right decision for them at all times over five years after.”
  • Carrying an unwanted pregnancy tended to be associated with worse outcomes for women’s physical health and socioeconomic status. Women denied abortion who later gave birth reported more chronic pain and rated their overall health as worse. Economic insecurity for women and their families increased almost four-fold. In terms of education, women who received abortions tended to have higher odds of having positive one-year plans while women denied abortions were no more or less likely to drop out of school.

Donohue, John J., and Steven D. Levitt. “The Impact of Legalized Abortion on Crime Over the Last Two Decades.” The University of Chicago, Becker Friedman Institute for Economics Working Paper №2019–75 (May 2017). https://ssrn.com/abstract=3391510

  • This paper primarily argues that legalizing abortion in the 1970s had positive consequences in the significant reduction of crime even two decades later, in the 1990s. In particular, the paper suggested an approximate 20% decrease in crime rates between 1997 and 2014. Not only is abortion legalization a crucial factor but perhaps one of the most crucial ones in the significant reduction in crime rates (see Donohue and Levitt, 2001).
  • A particularly crucial aspect of the data collected was that it took close to a decade for the “number of abortions performed to reach a steady-state” attributed to the variability and heterogeneity of state-level data due to the variability and dynamic nature of evolving abortion legislation and abortion reform.
  • Moreover, the effect of abortion on crime rates was only incrementally visible as “crime-aged cohorts” were gradually exposed to legalized abortion. Donohue and Levitt’s work supports the abortion-crime hypothesis — that increased access to abortion would decrease crime.

Frost, Jennifer J., Jennifer Mueller, and Zoe H. Pleasure. “Trends and Differentials in Receipt of Sexual and Reproductive Health Services in the United States: Services Received and Sources of Care, 2006–2019.” The Guttmacher Institute, June 24, 2021. https://doi.org/10.1363/2021.33017

  • This report describes trends in reproductive and sexual health care across the United States over a 13-year period as told by the National Survey of Family Growth, the only national data source that contains detailed information on sexual and reproductive health. It finds that some 7 in 10 women of reproductive age (44 million people) make at least one medical visit for sexual and reproductive health care each year. However, disparities exist — Hispanic women are less likely to receive care than White women, and the uninsured are substantially less likely to receive care than privately insured women.
  • It further finds that publicly funded clinics were a critical source of care for young women, lower-income women, women of color, foreign-born women, women on Medicaid, and women without insurance.
  • The report also finds that the Affordable Care Act increased the number of women receiving contraceptive services by 8% among women with private providers. There was a complimentary drop among women receiving contraceptive care from publicly funded clinics.

Hill, J. Jackson IV. “The Need for a National Abortion Reporting Requirement: Why Both Sides Should Be in Support of Better Data.” Available at SSRN (May 2, 2014). https://ssrn.com/abstract=2306667.

  • Hill writes a paper urging organizations to improve the status of abortion reporting in the United States. Examining statistics collected by the Centers for Disease Control and the Guttmacher Institute, the author finds serious deficiencies, including a lack of voluntary reporting from states, conflicting requirements (or unenforced requirements) about what data is collected, and an absence of timely data.
  • After the passage of Roe, state legislatures attempted to mandate abortion reporting and monitoring; however, concerns over the safety of women’s choice, undue administrative hurdles, and issues over pervasive data collection made it difficult to impose a standardized, non-intrusive, and anonymized data collection practice.
  • Hill argues that these data gaps and paternalistic methods of collecting data have had consequences on the ability of policymakers to make decisions around abortion policy and undermine the public’s knowledge on the issue. He assesses the feasibility of federally regulated abortion data and potential other strategies for achieving reliable, uniform data. He proposes two avenues for a “comprehensive, uniform abortion data” set: a ‘command’ option that requires states to provide and collect abortion information for a federal database or a ‘bribe’ option that monetarily incentivizes states to provide this information.

Knowles Myers, Caitlin, and Morgan Welch. “What Can Economic Research Tell Us about the Effect of Abortion Access on Women’s Lives?” Brookings, November 30, 2021. https://www.brookings.edu/research/what-can-economic-research-tell-us-about-the-effect-of-abortion-access-on-womens-lives/

  • Knowles Myers and Welch write on what current economic research suggests about abortion access on women’s reproductive, social, and economic outcomes.
  • Comparing Alaska, California, Hawaii, New York, Washington, and the District of Columbia (states which repealed abortion bans prior to Roe) to other states, research suggests states that repealed abortion bans had between a 4–11% decline in births relative to the rest of the country — with effects particularly large for teens and women of color. Studies also suggest that abortion legalization reduced the number of teen mothers by 34% and reduced maternal mortality by 30–40%, with little impact on white women.
  • Additional studies indicate that abortion access has a large impact on the circumstances under which children are born. Various studies find that abortion legalization reduced the number of unwanted children, cases of neglect and abuse, and the number of children living in poverty. It also improved long-term outcomes by increasing the likelihood of child attendance in college.
  • Other studies find that abortion and pregnancy have substantial impact on women’s economic and social lives, with pregnancy frequently lowering women’s wages. This fact has substantial implications for “low-income mothers experiencing disruptive life events.” Based on various studies, the authors argue that “access to abortion could be pivotal to these women’s financial lives.”
  • While abortion is driven by views on women’s bodily autonomy and when life begins, the authors find a clear causal link between access to abortion and “whether, when, and under what circumstances women become mothers.” All studies suggest that access to abortion can have substantial implications on education, earnings, careers, and life outcomes. Restricting or eliminating access would diminish women’s personal and economic lives along with that of their families.

Maxmen, Amy. “Why Hundreds of Scientists Are Weighing in on a High-Stakes US Abortion Case.” Nature 599, no. 7884 (October 26, 2021): 187–89. https://doi.org/10.1038/d41586-021-02834-7

  • A piece by Amy Maxmen for Nature summarizes a recent amicus brief filed by more than 800 scientists and several scientific organizations providing data-driven research into how abortion access is an important aspect of reproductive health.
  • It notes, for example, more than 40 studies suggesting that receiving an abortion does not harm a woman’s mental or physical health but that being denied an abortion can result in negative financial and health outcomes. It also cites a 2019 study of nearly 900 women who “who sought but were unable to get abortions reported higher rates of chronic headaches and joint pain five years later, compared with those who got an abortion,” while a similar 2017 study finds no similar physical or psychological effects.
  • A separate amicus brief submitted to the Court by about 550 public health and reproductive health researchers described how unwanted pregnancies can result in worse health outcomes. It also can disproportionately harm the physical, mental, and economic well-being of Black people according to a separate study.
  • An additional amicus brief filed by economists notes several studies that found that “abortion legalization in the 1970s helped to increase women’s educational attainment, participation in the labor force and earnings — especially for single Black women.”

Myers, Caitlin, and Ladd, Daniel. “Did parental involvement laws grow teeth? The effects of state restrictions on minors’ access to abortion.” Journal of Health Economics, 71, (2020): p.102302. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3029823

  • A paper by Caitlin Knowles Myers of Germany’s IZA Institute of Labor Economics and Daniel Ladd of the University of California, Irvine compiles data on the location of abortion providers and enforcement of parental involvement laws. The researchers seek to assess the impact of laws requiring parental approval for an abortion have on minors seeking abortions.
  • The paper concludes that parental involvement laws may have contributed to a modest decline in teen births (a 1.4% reduction) during the 1980s and 1990s but a 2.8% increase from 1993 to 2014 in women aged 15 to 18.
  • It further finds that laws with an avoidance distance (the distance minors have to travel to avoid parental involvement and can seek an abortion confidentially) have significant effects. In the 1980s, a parental involvement law with an avoidance distance of 100 miles decreased teen births by 1.48%. A parental involvement law with a 400-mile avoidance distance, about a day’s drive, increases the teen birth rate by 4.3%.

Popinchalk, Anna, Cynthia Beavin, and Jonathan Bearak. “The State of Global Abortion Data: An Overview and Call to Action.” BMJ Sexual & Reproductive Health 48, no. 1 (January 1, 2022): 3–6. https://doi.org/10.1136/bmjsrh-2021-201109.

  • Popinchalf and colleagues at the Guttmacher Institute write in the journal BMJ Sexual & Reproductive Health on the urgent need for data on abortion incidents and access to examine disparities in people’s ability to safely terminate a pregnancy.
  • The authors note that the three sources of data on abortion are official statistics, surveys of women, and scientific studies. However, stigmatization and varying legal access undermine the quality of this data and can lead to substantial under-reporting. Even in high-income countries, there can be significant variation in the frequency with which data is published. This variation in quality and availability exacerbates inequities by limiting the number of experiences that can be studied.
  • The authors argue that data availability and quality of abortion care can be improved by investing in country-level surveys and scientific studies. It also argues for reducing stigma through community and provider messaging as it can hinder the accuracy and completeness of datasets.

Tierney, Katherine I. “Abortion Underreporting in Add Health: Findings and Implications.” Population Research and Policy Review 38, no. 3 (June 1, 2019): 417–28. https://doi.org/10.1007/s11113-019-09511-8

  • Tierney notes that there is substantial evidence that abortion is significantly underreported in the United States, especially among Black women and those in lower socioeconomic classes.
  • She supplements this review with her own evaluation of the abortion data in the National Longitudinal Study of Adolescent to Adult Health (Add Health), finding that the dataset captures only 35% of expected abortions. Examining data from 1994–1995, 1996, 2001–2002, and 2008–2009, she found severe abortion underreporting; however, there were no significant differences between race/ethnicity, age, or time of abortion and underreporting.
  • Tierney argues that this fact means that Add Health is no better than other surveys in collecting abortion data. She also argues that this underreporting, likely caused by stigma, has substantial implications for research and that researchers should be cautious with self-reports of abortion. Figures need to be evaluated, contextualized, and used with caution.

10 learnings from considering AI Ethics through global perspectives

Curated on Posted on by Stefaan Verhulst

Blog by Sampriti Saxena and Stefaan G. Verhulst: “Artificial Intelligence (AI) technologies have the potential to solve the world’s biggest challenges. However, they also come with certain risks to individuals and groups. As these technologies become more prevalent around the world, we need to consider the ethical ramifications of AI use to identify and rectify potential harms. Equally, we need to consider the various associated issues from a global perspective, not assuming that a single approach will satisfy different cultural and societal expectations.

In February 2021, The Governance Lab (The GovLab), the NYU Tandon School of Engineering, the Global AI Ethics Consortium (GAIEC), the Center for Responsible AI @ NYU (R/AI), and the Technical University of Munich’s (TUM) Institute for Ethics in Artificial Intelligence (IEAI) launched AI Ethics: Global Perspectives. …A year and a half later, the course has grown to 38 modules, contributed by 40 faculty members representing over 20 countries. Our conversations with faculty members and our experiences with the course modules have yielded a wealth of knowledge about AI ethics. In keeping with the values of openness and transparency that underlie the course, we summarized these insights into ten learnings to share with a broader audience. In what follows, we outline our key lessons from experts around the world.

Our Ten Learnings:

  1. Broaden the Conversation
  2. The Public as a Stakeholder
  3. Centering Diversity and Inclusion in Ethics
  4. Building Effective Systems of Accountability
  5. Establishing Trust
  6. Ask the Right Questions
  7. The Role of Independent Research
  8. Humans at the Center
  9. Our Shared Responsibility
  10. The Challenge and Potential for a Global Framework…(More)”.

China’s Expanding Surveillance State

Curated on Posted on by Stefaan Verhulst

Article by  Isabelle Qian, Muyi Xiao, Paul Mozur and Alexander Cardia in the New York Times: “China’s ambition to collect a staggering amount of personal data from everyday citizens is more expansive than previously known, a Times investigation has found. Phone-tracking devices are now everywhere. The police are creating some of the largest DNA databases in the world. And the authorities are building upon facial recognition technology to collect voice prints from the general public.

The Times’s Visual Investigations team and reporters in Asia spent over a year analyzing more than a hundred thousand government bidding documents. They call for companies to bid on the contracts to provide surveillance technology, and include product requirements and budget size, and sometimes describe at length the strategic thinking behind the purchases. Chinese laws stipulate that agencies must keep records of bids and make them public, but in reality the documents are scattered across hard-to-search web pages that are often taken down quickly without notice. ChinaFile, a digital magazine published by the Asia Society, collected the bids and shared them exclusively with The Times.

This unprecedented access allowed The Times to study China’s surveillance capabilities. The Chinese government’s goal is clear: designing a system to maximize what the state can find out about a person’s identity, activities and social connections, which could ultimately help the government maintain its authoritarian rule.

Here are the investigation’s major revelations.

Analysts estimate that more than half of the world’s nearly one billion surveillance cameras are in China, but it had been difficult to gauge how they were being used, what they captured and how much data they generated. The Times analysis found that the police strategically chose locations to maximize the amount of data their facial recognition cameras could collect…

The Chinese authorities are realistic about their technological limitations. According to one bidding document, the Ministry of Public Security, China’s top police agency, believed the country’s video surveillance systems still lacked analytical capabilities. One of the biggest problems they identified was that the data had not been centralized….(More)”.

The Intersection of Data, Equity, and City Governments

Curated on Posted on by Stefaan Verhulst

Blog by Yuki Mitsuda: “The Open Data Policy Lab’s City Incubator program was established in September 2021 to help realize the Third Wave of Open Data at the subnational level by building data capacity among city intrapreneurs. In its first iteration, the program supported innovators from ten cities around the world to better use data to address the opportunities and challenges they face.

Reflecting on the six-month program, the work enabled participants to meet the needs of their cities and the people within them. They also revealed shared themes across cities — common challenges and issues that defined urban, data-driven work in the 21st century. This blog explores one of the emerging themes we saw from participants in the City Incubator program: the intersection of equity, data, and city governments…

Three of our city incubator participants designed their data innovations around the ways cities and citizens can use data to measure and improve equity. 

  • Jennifer Bodnarchuk, a Senior Data Scientist at the Innovation & Technology Department in the City of Winnipeg, for example, led the development of a Diversity Dashboard that quantified and visualized their municipal government’s workforce representation. The tool can be used to measure the level of diversity represented in city-wide employment to move towards equitable hiring in the public sector. 
  • Henry Xavier Hernandez, the Chief Information Officer at the Information Technology Department in Guayaquil, Ecuador, and his team leveraged the City Incubator to develop Citizen 360, a public market analysis platform that helps businesses, organizations, and individuals identify economic opportunities in the city. This tool can aid small business owners from all backgrounds who are navigating the journey of starting a new business.
  • Andrea Calderon led Albuquerque’s Equity Index, which helps evaluate the reach of city service distribution with the goal of increasing municipal investment in pockets of the city where equitable city service provision has not yet been achieved. Albuquerque’s Equity Index work entailed assessing air quality in the city through the framework of cumulative impacts, which measures “exposures, public health, or environmental effects from the combined emissions in a geographic area” in pursuit of environmental justice…(More)”.

Sweeping Legislation Aims to Ban the Sale of Location Data

Curated on Posted on by Stefaan Verhulst

Article by Joseph Cox and Liz Landers: “Sen. Elizabeth Warren and a group of other Democratic lawmakers have introduced a bill that would essentially outlaw the sale of location data harvested from smartphones. The bill also presents a range of other powers to the Federal Trade Commission (FTC) and individual victims to push back against the multibillion-dollar location data industry.

The move comes after Motherboard reported multiple instances in which companies were selling location data of people who visited abortion clinics, and sometimes making subsets of that data freely available. Such data has taken on a new significance in the wake of the Supreme Court’s looming vote on whether to overturn the protections offered by Roe v. Wade. The bill also follows a wave of reporting from Motherboard and others on various abuses and data sales in the location data industry writ large.

“Data brokers profit from the location data of millions of people, posing serious risks to Americans everywhere by selling their most private information,” Warren told Motherboard in a statement. “With this extremist Supreme Court poised to overturn Roe v. Wade and states seeking to criminalize essential health care, it is more crucial than ever for Congress to protect consumers’ sensitive data. The Health and Location Data Protection Act will ban brokers from selling Americans’ location and health data, rein in giant data brokers, and set some long overdue rules of the road for this $200 billion industry.”…(More)”.