Criminology

Removing human bias from predictive modeling

Predictive modeling is supposed to be neutral, a way to help remove personal prejudices from decision-making. But the algorithms are packed with the same biases that are built into the real-world data used to create them. 

Penn Today Staff

Documentary shows discriminatory impact of state legal assistance provision

The Penn Program on Documentaries and the Law has produced a new documentary that exposes the discriminatory impact of a provision of the Pennsylvania Victims Assistance Compensation Program law that denies assistance to victims who contribute, or are suspected of contributing, to their own death or injury.

Penn Today Staff



Media Contact


In the News


The New York Times

The lasting effects of stop-and-frisk in Bloomberg’s New York

John MacDonald of the School of Arts and Sciences discussed the unproductive majority of street stops made by police under stop-and-frisk. “Who’s being affected by that?” he asked. “It’s going to be people who, for example, may be likely voters, who are trying to go to school, who are afraid because they normally wouldn’t have interactions with police that are intrusive. That’s not your average offender. That’s your average citizen.”

FULL STORY →



BBC

Democratic debate: Mike Bloomberg's record on crime in New York fact-checked

Research by John MacDonald of the School of Arts and Science was cited in an article fact-checking the most recent Democratic debate. The study found that stop and frisk “made almost no difference” to New York City’s crime rates.

FULL STORY →



Newsweek

Bloomberg stop and frisk comments resurface, said he put ‘all the cops’ in minority neighborhoods ‘where all the crime is’

John MacDonald of the School of Arts and Sciences was cited for his analysis of the reduction in New York City crime rates, which has been attributed to stop and frisk policies. “Saturating high crime neighborhoods with extra police helped reduce crime in New York, but the bulk of investigative stops did not play a meaningful role in the crime reduction,” he wrote.

FULL STORY →



The New York Times

An algorithm that grants freedom, or takes it away

An algorithm created for the Philadelphia Adult Probation and Parole Department by Richard Berk of the School of Arts and Sciences tries to improve on human judgement by excluding data that could be a proxy for race. “All machine-learning algorithms are black boxes, but the human brain is also a black box,” he said.

FULL STORY →



The New York Times

Trump wants law and order front and center

Research by Aurélie Ouss of the School of Arts and Sciences on the consequences of Philadelphia’s “no cash bail” policies was cited.

FULL STORY →



The Washington Post

Algorithms were supposed to make Virginia judges fairer. What actually happened was far more complicated

Aurélie Ouss of the School of Arts and Sciences praised a study that proved that algorithms are imperfect tools when it comes to predicting crime. She said that, ultimately, the usefulness of algorithms in criminal justice comes down to implementation: “It may be a case that a different tool that’s designed differently—that judges use differently—would yield different results.”

FULL STORY →