Arnold Foundation Public Safety Assessment tool recommends a Get-Out-of-Jail Free Card on 93 percent of felony weapons offenses
In a recent report, Metro Crime Commission prepares report on Risk Assessment System, an evaluation was conducted by the New Orleans Metropolitan Crime Commission of the Arnold Foundation’s (now Arnold Ventures) public safety risk assessment instrument (PSA) – an algorithm based pretrial risk assessment widely considered the gold standard in determining the pretrial risk and release of criminal defendants. The Chairman of the Commission noted that the use of the Arnold Foundation risk assessment has led to “a flawed system.” He went further and said, “At this point it is a waste of money and we better change it.”
"This group is advocating for offenders nobody’s advocating for law-abiding citizens that are jeopardized by some of these violent offenders."
Why was it flawed? “The report said that free bond releases were recommended for 75 percent of violent felony suspects and 93 percent of weapons felony suspects. The report also stated that 33 percent of violent felony suspects, and 29 percent of weapons felony suspects were rated the lowest risk - Level One, which recommends free release with no supervision.” In fact, the Chairman noted on one example: “he was rated a level one on his most recent bookings and he’s got over 100 charges against him.”
Interestingly, the Arnold Foundation did not comment on the report in the article. Yet, John Arnold supported legislation to make the tool completely transparent for evaluation and inspection, including Arnold’s intellectual property. What would he have to hide? We often wondered how good the data set John Arnold used to build the tool—how good is it, and is it derived from the FBI’s crime files? The State of Louisiana should follow the lead of Idaho, a move John Arnold supported, and force immediate transparency.
Enter risk assessment defender Cherise Fanno Burdeen from the Pretrial Justice Institute, who has tap-danced her way pivoting her position through this issue for so long we are probably the only people that read her quotes and then choose to dignify it with a response. Recall, Ms. Burdeen and her friends have said and supported for years that risk assessments should replace the system of money bail, coupled with the use of reckless policies of preventative detention based on risk to detain defendants. Now, the leading pretrial algorithm in the United States, the Arnold Foundation PSA, she says, still works and does not need to be fixed. The Metropolitan Crime Commission said that their new system to replace the Arnold tool will “identify mental health problems, lack of job preparedness, substance abuse, and other challenges that may be contributors to criminal behaviors.” Not necessary says Ms. Burdeen: “The fact is that the PSA and other tools that have been thoroughly tested through rigorous research have found that these factors and others like them add nothing to the tools’ ability to assess risks.” Sounds like some great science!
Ms. Burdeen then suggests that release from jail pending trial should be based on the “nature of the offense and strength of the evidence.” So, if the PSA is scientific, why does it reject such important factors? We have heard for years from Ms. Burdeen and her ilk that risk assessments should not consider the current charges because the defendant is presumed innocent and the current charge does not correlate with risk. Indeed, what would such factors do to the presumption of innocence if the District Attorney could lock up someone in jail based purely on the nature of the offense charged? Should the strength of the evidence be a factor in deciding bail? Says Ms. Burdeen: “No assessment tool can quantify the nature of the offense and strength of the evidence, which is why it is the judge, not the tool, that makes the decision." We guess so. Yet another pivot over at the Pretrial John Arnold-Defending Justice Institute.
As we have said for years, you can’t have it both ways. Either the tool is scientific, or it isn’t. In this case, it has been shown not to be. Ms. Burdeen’s response is that it’s not supposed to be. Judges take the science, pour some discretion on it, and “poof,” make it more scientific after that. In other words, she was for it before she was against it.
At the end of the day, the PSA doesn’t work. Peer-reviewed academic research has already proven that. As the Chairman of the New Orleans Metropolitan Crime Commission said, in frustration, “If all you want to do is reduce the jail population, you don’t have to waste your time with this program, just tell police not to arrest them.”
It is time for jurisdictions to revisit their reliance on the Arnold Foundation PSA, force complete transparency that John Arnold himself agrees with, and look at other ways to reform the system of pretrial release.