Beyond statistics:
Readiness in the face
of uncertainty

At HRDAG, we are responding to a period of incessant uncertainty and chaos with readiness, focus and determination. Experience has taught us to remain relentless in the pursuit of undeniable truth — and so, we are. Statistics deliver insight, and our insights pave the way to clarity, accountability and ultimately, justice.

We know that perpetrators of human rights violations create false narratives, hiding evidence of crimes by distorting data, creating gaps. We illuminate those gaps. We know that science can confuse the public when vital data on pandemic risk and safety is unclear. We clarify the facts and explain the science. We know that justice seeking truth commissions require reliable evidence of undocumented deaths and abuses. We calculate what’s missing — and provide the scientific evidence.

We start with what is known, determine what is missing, and provide what is desperately needed: clear answers to urgent questions in the midst of uncertainty. We are equally relentless in preparing for our next challenge. We innovate new statistical methods, monitor world events, and initiate dialogues with prospective partner organizations. When an opening appears, we move rapidly to organize, preserve and protect relevant data. We achieve our goal when legitimate questions of fact yield to informed demands for justice and accountability.

At HRDAG, we are passionate about data and relentless in the pursuit of a better future.

But first, the public needed the facts.

HRDAG statisticians Patrick Ball and Kristian Lum tackled one underlying question in 2015: How many people in the US have been killed by police, including undocumented homicides? Using statistical estimation, we calculated: nearly 10,000 Americans over an 8-year period.

In 2019, newly-available data created an opportunity for us to analyze police misconduct more broadly. Following a multi-year battle, our partners at the Invisible Institute gained access to records of citizen complaints against police in Chicago. We worked with them to design and maintain a data pipeline to process enormous “data dumps” of hundreds of thousands of documents. These complaints typically contain large blocks of text, which make quantitative analysis challenging, but not impossible. Our combined efforts are linking formal complaints to individual officers, making it possible to quantify the number and types of complaints-per-officer.

The result? The most accurate, comprehensive picture yet of police misconduct in Chicago.

That is only the beginning.

We anticipate a substantial influx of new data on police behavior in response to lawsuits, legislation, and formal requests. New York State just repealed a section of its legal code, thereby making police misconduct records more accessible. California is strengthening legislation to make information about police use of force publicly available. More and more records of this type are going to become available in the months and years to come.

Alongside our partners, we will be ready to uncover the truth — bringing facts into sharp focus, and demands for justice and reform to the forefront.

How reliable are Covid-19 test results? In the chaos of a worldwide pandemic, fueled by the spread of misinformation, we all feel an urgent need for accurate, understandable information. The appearance of scientific uncertainty feeds public confusion — hindering individual decision-making, and rippling out in invisible waves of community impact. How can HRDAG help? We bring clarity to the uncertainty.

First, we focus on asking the right question: How many people are infected? Public health departments can only measure how many people receive a positive test result, which is not the same thing. Raw tallies of diagnosed infections reflect some combination of the number of infections in the population, the rate of testing in the population, and the proportion of infections that cause symptoms. Nothing in these variables holds constant from city-to-city or state-to-state.

Test results are also difficult to interpret on an individual level. A positive result does not prove sickness; nor does a negative one ensure health. But, a specific mathematical equation, called Bayes Theorem, helps explain both results.

Consider screening tests looking for SARS-CoV-2 proteins or antibodies (potential indicators of a past infection, but not a current one). What does a positive test result indicate? What is the likelihood that it is correct and accurately recognizes antibodies? Answering these questions requires knowledge of the test (its “sensitivity” and “specificity”) AND the infection rate within a community. Counterintuitively, the lower the infection rate in the community, the higher the odds that a positive test is incorrect. Our team unpacked all the math in an essay in Granta magazine in April 2020.

While the implications for public policy are debated nationwide, our essay sidesteps the politics to provide a valuable public service: equipping individuals to interpret their own test results accurately, and make informed personal decisions.

What does justice require? Countries managing the transition from injustice-in-conflict to accountability-in-peace require data to determine patterns of violence, assess culpability, and prescribe justice. We have a long history of contributing to these transitions; Patrick Ball’s 2015 testimony in the trial of Chad's former president Hissène Habré is just one recent example.

This year, we are in the midst of advising two projects in countries as they end long civil wars. The sensitive process necessitates discretion; we cannot identify our partners currently. While our role is straightforward — we clarify questions, analyze data, provide reliably precise answers, and (when necessary) testify to them — the transitional justice efforts are complicated and long-term. This is often the case. Our analyses cited by the judges when they found Habré guilty of crimes against humanity were decades in the making.

We may not know yet the specific outcomes of our current international projects. But we do know that inaccurate statistics can damage the credibility of human rights claims — and that is why we strive to ensure that statistics about human rights violations are generated with as much scientific rigor as possible. The evidence we produce for our partners is just that: evidence. Documented in its methodologies, conclusive in its results, sufficient for courtrooms and the press. Ready for delivery at those pivotal moments when what seemed impossible becomes possible: resolution and accountability.

HRDAG is relentless in our pursuit of human rights and seeking justice around the world. This work is at the foundation of who we are. As we witness funding for and attention to such critical work shrinking, we remain passionately committed to partnering with truth commissions and international peace processes in the years to come.

The central tenet of HRDAG’s work is a deep-seated commitment to justice, accountability, and truth. Every year, the pursuit of these fundamental principles becomes more fraught and more necessary. This past year was no exception with extraordinary challenges posed by the Covid-19 pandemic, widespread worldwide protests due to police violence, and the very concept of evidence-based science being questioned by so many people around the world. The times are both unprecedented and uncertain.

In the midst of this uncertainty, I am proud to say that my colleagues and I at HRDAG have been steadily and systematically doing the work we do best — using statistics to advance international human rights and justice. As data scientists in the field we created nearly 30 years ago, we continue our critical work with extraordinary focus and commitment. We are profoundly grateful to our loyal and generous supporters for sharing our belief in rigorous scientific methods and dedication to human rights and justice.

Megan Price
Executive Director

WE THANK THIS YEAR’S GENEROUS SUPPORTERS