CWRU researchers work to analyze “signaling language” in sexual assault cases

Virginia Behmer, Staff Reporter

Researchers from the Begun Center for Violence Prevention Research and Education of the Jack, Joseph and Morton Mandel School of Applied Social Sciences (MSASS) have received a $715,000 grant from the National Institute of Justice (NIJ). This grant will fund their project to use artificial intelligence to find signaling-language patterns in police reports included in Sexual Assault Kits (SAKs, also known as “rape kits”). This is an expansion on their original project of evaluating untested SAKs from Cuyahoga County.

The Begun Center is made up of 50 people from MSASS. According to their website, they have a mission of “social justice and community development by conducting applied, community-based and interdisciplinary research on the causes and prevention of violence, and by educating and training social workers, teachers, law enforcement and other professionals in the principles of effective violence prevention.”

The center currently has a Sexual Assault Kit Initiative (SAKI), where individuals from MSASS along with the Cuyahoga County Sexual Assault Kit Task Force are committed to evaluating untested kits in Cuyahoga County.

This was the original project at hand for Rachel Lovell, principal investigator and senior research associate with the Begun Center, and her team. Both DNA evidence and police reports are being evaluated from a sample of 7,000 unsubmitted SAKs from the years 1993-2009 in Cuyahoga County.

“The idea for the [NIJ] grant came out of the project,” said Lovell. “In reading a lot of those [police reports] we noticed early on that there was very blatant victim blaming and doubting of what the victim was saying.”

Reporting officers of the assault are often those who write the original report. Reports are then used by potential detectives and prosecutors depending on how the case proceeds, so these reports hold inordinate amounts of power.

Such “signaling language” used in police reports might affect survivors’ experience and credibility, how the case is handled by future law enforcement and court case proceedings. According to Lovell, most officers writing the reports for the assaults are not trained in neurobiology of trauma and often seem “confused by victims’ behavior.”

The current project, funded with the $715,000 grant, utilizes machine learning technology to track and compile patterns in signaling language in police reports.

“I think this is a case for how open data can help in ways people might not expect,” said Misty Luminais, a senior research associate with the Begun Center also working on the project. “Without the access we’ve been given, we couldn’t have dreamt of proposing such work.”

Cities around the country have embarked on examining their backlog of untested SAKs, and according to Lovell, “Cleveland is not unique nationally.”

“When sexual assault kits were first collected, DNA wasn’t around or available,” Lovell said. DNA was also very expensive and often took years to process. As such, only certain SAKs were processed according to many provisions, including if perpetrators were strangers or if survivors wanted to prosecute.

As researchers, we thrive on data, but we also have a responsibility to do work that positively impacts the world we live in. Partnerships such as these can demonstrate how researchers are woven into their communities, not just outside, disinterested observers,” said Luminais.

The team knows “the power of words” and recognizes that their research is groundbreaking, as a small amount of similar research has been done. They see future projects stemming from this one, including possible opportunities to inform police report writing training in reporting officers.

“One of the most exciting implications I see for this project [is to] go beyond the criminal justice realm and pushes the envelope on how research will be conducted in the social sciences going forward,” Luminais said. “This approach allows us to leverage the nuance of qualitative data with the wide scope of quantitative data to truly capture the strengths of both approaches. In the future, I hope to see numbers enhanced by narrative.”