How Big Data Reveals Bias in Virginia Courts

Jun 6, 2016

Credit luckey_sun/Flickr

The justice system is supposed to be impartial and fair. Many, though, assume justice isn’t blind when it comes to matters of race or income. But now an unlikely pair -- a lawyer, and a software engineer -- have used data to uncover bias in Virginia’s courts. 

David Colarusso is a lawyer turned data scientist in Boston. A couple months ago, some of his colleagues were arguing about which is a bigger problem in criminal justice: bias against people of color? Or people who are poor? 

“That question was sort of, lingering in my mind. And the first thing that came to my mind was ‘That’s an empirical question! If we had the right data we could answer that!” says Colarusso.

At just about the same time, Ben Schoenfeld, a software engineer in southwest Virginia, was finding and publishing massive amounts of bulk data -- more than 2 million Virginia criminal court records.

"I believe in the power of open data, and as someone who knows how to write software, and it's easy and fun for me, it's a great way to contribute by writing this software and getting this data and giving it to people who can do really amazing things with it," says Schoenfeld.

So that’s what Colarusso did. A couple weeks, and a lot of statistical analysis later, he was able to find the answer to his colleague’s original question.

“In the end, the finding was that for a black person in Virginia courts to get the same treatment as their white peers they would have to make an additional $90,000 a year,” Colarusso says. “So that really sort of puts in stark contrast the power of class and race as it relates to outcomes.”

Colarusso's conclusion is limited, he found race is a small factor among many in determining case outcomes. But he’s gotten positive feedback from data scientists and lawyers alike. They hope his analysis can provide a model for others.