Vol. 3 Issue 11
Trina & Maira on Beneath the Surface; New from Jamie Kalven
Volunteers participate in a group discussion at a Beneath the Surface workshop, which focused on patterns of Chicago Police misconduct complaints that describe police neglect following a reported sexual assault. October 2021.
November 19, 2021
A Conversation with Trina & Maira reflecting on Beneath the Surface:
I sat down with Trina Reynolds-Tyler, our director of data, to reflect on progress made within the Beneath the Surface project this year. Beneath the Surface, our ongoing inquiry into gender-based violence by police, has recently completed six months of a data science investigation with our partners at Human Rights Data Analysis Group (HRDAG) and over 200 volunteers. After months of writing an algorithm that can read and sort thousands of complaint narratives made available through Charles Green v. Chicago Police Department, we have begun to share our early findings with the public.
In the weeks to come, we will share with you our findings on police neglect after instances of sexual assault, harm facing children, and domestic violence. We will also be closely watching Charles Green’s appeal to the Illinois Supreme Court, which could force the release of Chicago Police investigations of police misconduct complaints going back to 1967. As we analyze and discuss this growing library of misconduct, we hope you’ll join us.
Maira Khwaja
director of public strategy
MK: Let's back up. How did you become director of data and director of the Beneath the Surface project?
TRT: When I first began at the Invisible Institute in 2016, I was an AmeriCorps member, in my second year at Public Allies. It was my job to do outreach for the Citizens Police Data Project (CPDP) and work within the classrooms for the youth/police project (y/pp). Over time, as my work around CPDP grew and my conversations with the y/pp alumni deepened, I became interested in better understanding how I could use data to answer questions that I had about the world around me. Over time, my interest grew, as did my roles and responsibilities at the Invisible Institute. I ultimately went to the University of Chicago to get my master's in public policy. While I was there, I sharpened my technical skills to do machine learning data science projects.
MK: I remember in the youth/police project the way young Black women would often gravitate to you to discuss their experiences with gender-based violence. How do those conversations inform or relate to some of the investigative questions you're working on today?
TRT: In the youth/police project, the young people are going to share with you what their lived experience was, or what the lived experience was of someone that they knew, or the stories that they heard in the myths of their neighborhood. What I learned while talking to these young women was that there are so many shared experiences that young women have, and those shared experiences come into conflict or contact with the government or the state during specific times. Maybe some of them have a stalker, or someone is harassing them, or maybe they got into a fight at school. And there are so many of those stories that often repeat themselves. I think you'll find many of those stories within the complaint data.
MK: Now, shifting to the present - you've been leading this project for six months, a collaboration with HRDAG to build a machine learning model that can help us dig deeper into tens of thousands of complaints. Can you explain why we needed to create a machine learning model to dig deeper into complaints? Why is the data we already have on CPDP not enough?
TRT: What you see on CPDP is that the Invisible Institute ultimately gained access to high level [police] data that didn't include the underlying complaint narratives unless we made a specific Freedom of Information Act request for them. The category a complaint was given could mean many things, or nothing. For example, they have a category for “conduct unbecoming - miscellaneous,” but then they also have a category called “miscellaneous - conduct unbecoming.” These categories don't actually provide much insight. Because we now have access to the narratives around the complaints between 2011-2015 because of Charles Green’s lawsuit, we were able to dive deeper into the complaints. But we're talking about over 27,000 unique complaints. Where would we even start? And so, we built this algorithm as a means to help us. Instead of looking for a needle in the haystack while reading thousands and thousands of complaints, we instead built an algorithm so that we would identify a haystack full of needles. The algorithm separates the documents into buckets where there's a high probability that it is associated with the topic that we're interested in or researching.
MK: How would you describe those needles?
TRT: When we first started the project, we prioritized sexual violations at the hands of the police and because we knew there would be instances of it. But because we were using machine learning, we realized that we didn't have to just look for sexual violations. We could actually look for some larger themes or other context around complaints and so that's what we did. We identified sexual violations, but there's also complaints like “home cases,” where officers entered a home, or “disabled cases” where a disabled person was present or impacted in some way. We embraced this kind of flexibility, this opportunity to not just look into sexual violations, but ultimately identify other forms of gender based violence.
MK: Over the past month you've had a series of events--some just for Black women and non-binary people, some just for Black men, and some with anybody who was interested in the work, who'd volunteered with us or had an interest in working on similar things. Can you tell us why you decided to have events before releasing an investigation?
TRT: The decision was deeply rooted in respect for the people who have not only come forward about these things, but also who have lived through these things and haven't actually spoken to anybody at all about it. Statistically, if you look at the National Center on Violence Against Women in the Black community, they have these really interesting stats. It talks about how based on your sex and gender, you are more likely to have experienced these things. And so I knew that if we were to put Black women and nonbinary folks in the room together, if we were to put Black men in the room together, and discuss some of the topics that we're covering, that they would all be reminded of someone that they knew or even maybe be survivors themselves. Before sharing the information with the rest of the world, I felt it was important to consult with survivors themselves. It was an opportunity to get advice from them about how they think it would be best to talk about these things, or to better contextualize some of the things that I've found. Also, we were able to participate in a kind of future-thinking or healing practice, to ask, “what do we imagine this survivor could have experienced differently?”
MK: You were saying there were actually so many potential topics or themes we could have started with, but we chose to start with neglect by police. Why start with neglect as the topic you discuss when you bring people together?
TRT: We can all identify with the experience of needing help and having to call somebody. And I know a lot of people who are actively not calling police. And then the question that people ponder is like, Well, why don't you call police? And it's like, Well, so here are some examples of what people literally said happened to them when they call police. This is just evidence, right? We don't know who these people are. We can't. We only have access to limited information. This text operates as a body of evidence where we can identify patterns and maybe learn about some of the ways that people, when they're calling police for help, have been neglected.
MK: Last question - what did you learn in this series of discussions in October?
TRT: I was so grateful that we were sharing with other people because I think, as a person who works within this field, there's this bit of desensitization that happens. We're always looking at death and murder. We're not necessarily looking at the survivors who remain alive and have to actively function as a part of society. A lot of people at the events had shared experiences. There’s so much sexual trauma that people don't do not discuss because of shame. I also heard from people who were thinking of actual solutions, things that they could do in real time or ways that they could support their friends and family. They were imagining, how organized does a community need to be to respond to a crisis? Within your neighborhood, within your community, who is coming? Who do we want that person to look like? What does comfort look like after trauma? How much time would it take to move people to trust each other?
Receive detailed updates on BTS and opportunities to volunteer at bit.ly/btssignup
Jamie Kalven wrote about the logic of the "split second decision," and the ways it has shaped the law and public discourse. He calls for sharply delineating the limits of law enforcement.
Our Virtual Supporters Convening has been postponed to Tuesday, January 18, 2022 from 12:00 -1:15pm CT. We look forward to beginning the new year with you.
RSVP at bit.ly/iisupporters2021→
Khalid Gibran Muhammad and Ben Austen, a former Invisible Invisible Institute fellow, wrote about the need for parole in Illinois.
Read “Let the Punishment Fit the Crime,” in the New York Times →
Gabriel Mugar featured CPDP.co as an example for how to design for human experience while “accounting for invisible systems of power.”
Read more in Fast Company→
Last month, the Institute for Research on Race & Public Policy at University of Illinois - Chicago hosted a “Data for Justice” panel, featuring Jamie Kalven, Joe Ferguson, Janae Bonsu, Andy Clarno, and Matthew Saniie.
Watch the recording here →
