Share:
Senate Co-Sponsorship Memoranda
Senate of Pennsylvania
Session of 2019 - 2020
Regular Session
MEMORANDUM
| Posted: | March 4, 2019 04:54 PM |
|---|---|
| From: | Senator Sharif Street |
| To: | All Senate members |
| Subject: | Repealing the Mandate to Develop a Pre-Sentencing Risk Assessment Tool |
| In the near future, I will introduce a bill to repeal the mandate on the Pennsylvania Commission on Sentencing to develop a pre-sentencing risk assessment tool. In 2010, the legislature passed Act 95 requiring the Sentencing Commission to develop a risk assessment tool to assist courts at sentencing. While the original intent was to help the court divert appropriate non-violent offenders to alternative sentencing programs rather than prison, the version that passed created a mandate to develop a “2nd generation tool” which relies only on static factors used to predict likelihood of recidivism. It does not evaluate dynamic needs factors’ such as substance abuse, family/marital relationships, education/employment, financial status, and mental health. These factors provide important information necessary in guiding judges to appropriate alternative sentencing programs. Since the passage of the mandate, the Commission has worked hard to create an automated tool that is statistically predictive of risk and does not show bias against any protected class. After reviewing more than 8 years of thorough research and development conducted by the Commission, many, including members of the Commission itself, have serious concerns that such an automated tool is possible. In 2018, the Commission released an algorithm for public comment. There was broad criticism of the tool by public advocates and experts in criminal justice. Many had concerns that an algorithm designed to make predictions about future behavior based on static factors such as age, sex and criminal history could have a disparate impact based on race or gender. Specifically, the current algorithm fails on two important counts. First, by the Commissions own conclusion, it is not statistically predictive in many cases. When determining whether someone is a “high risk” to commit a violent crime in the future, it is only correct 11% of the time. Additionally, it is only 52% accurate when it labels someone a “high risk” to recommit any crime. That is basically the same as flipping a coin. The algorithm may also have a constitutional problem, due to its use of race and gender. While the Commission solicited a legal opinion in 2015 which advised that considering race or gender as a factor in the tool may not be constitutional, a decision then was made to continue to account for gender. The current algorithm assigns an additional point to all men, and not women, thus assuming all men are “riskier” than women. Additionally, the algorithm likely has a disparate impact on race due to the utilization of arrest records amassed during a history of over policing in communities of color. As a member of the Pennsylvania Senate serving on the Pennsylvania Sentencing Commission, I hope you will join me in co-sponsoring the repeal of this unattainable mandate so that the Commission can get back to its other important work. Please contact Micah Mahjoubian at micah@pasenate.com if you have any questions. |
Introduced as SB474