Arnold Ventures’ New Paid Mouthpieces Conclude The Public Safety Assessment Worked In New Jersey Despite A Feeble, Tenuous, And Unconvincing Justification And Strong Evidence To The Contrary
Arnold Ventures (formerly the Laura and John Arnold Foundation), as we have repeatedly said, cannot concede defeat regardless of how bad things get. Their attempt to “moneyball” criminal justice using their Public Safety Assessment Tool (PSA) has been ongoing for nearly seven years now. Landmark research showed their moneyballing didn’t work in Kentucky, and now the Chief Justice and other leaders in Kentucky are calling for bail reform. This is the only peer-reviewed study to review the results of the Arnold Ventures PSA, based on five years of historical data.
Twenty-seven prominent academics have recently jumped in to the fray now calling on jurisdictions to stop using pretrial risk assessment tools. The Leadership Conference of Civil Rights has said these are problematic. The ACLU recently testified at a hearing in front of the U.S. Congress and argued against the use of risk assessment tools. And, of course, eighty digital groups that make up the Partnership on AI have also questioned the race neutrality and efficacy of these tools, including lesser known companies such as Facebook, Twitter, IBM, Samsung, etc.
Even Arnold Ventures recently conceded that of the top five solutions to bail reform issues, their own Public Safety Assessment did not crack the top five as a solution.
Of course, we have previously noted that the Arnold Foundation relies on paid hacks posing as academics in order to conduct “independent” studies. As it turns out, every independent study commissioned by Arnold Ventures concluded, you guessed it, that the Arnold Ventures PSA is working. They also conclude that the tool is racially righteous, by, in particular changing the methodology to evaluate for racial and other protected-class bias to then conclude everything is all good at Arnold Ventures.
Arnold relied on a fake institute that has no professors, no students, and no website called the Center for Safety, Justice and Resilience (“RTI”).
RELATED: Arnold Foundations “Pay to Play” Pays Off with New Validation Study of Risk Assessment Tool
Apparently, RTI was long on justice and short on resilience because Arnold Ventures has a new vendor to conduct its apparent independent studies. The new paid mouthpiece of Arnold Ventures is called the Manpower Demonstration Research Corporation (“MDRC”). They recently issued a new report proclaiming the success of the New Jersey system, but they fundamentally fail to respond to the clear statistics that the new system is worse in terms of outcomes, locking up and throwing away the key to thousands of defendants recommended for preventative detention and, at best is a wash with the State having expended hundreds of millions of dollars to implement it.
At least MDRC was clear who was paying for the report: “With funding from Arnold Ventures, MDRC is conducting an independent study of how CJR was implemented and assessing its effects on case dispositions, new criminal charges, and other important outcomes.” We call that not an “independent study” but instead a conflict of interest: if the researchers were to conclude that the Arnold Ventures PSA had a negligible effect in New Jersey, then MDRC would not get funded by Arnold Ventures. We already burned Arnold for giving RTI $761,381.00 to manufacture “independent” studies before. And here we are again. So, the question is, how much is MDRC taking from Arnold Ventures to conclude that Arnold Ventures is great?
Let’s get to the point—this research is not peer-reviewed, does not appear to have been conducted by anyone with a doctoral level education, and offers little if any evidence that the Public Safety Assessment had any impact. The biggest problem, as the report concedes, is that their data on the PSA did not include any of the cases where officers issued a summons to the defendant. The report’s key piece of evidence that the PSA was working was that judges followed it 63% of the time: “In almost two-thirds of cases (63 percent), the release conditions resulting from first appearance hearings matched the DMF [PSA “decision-making framework] recommendations.” That is flimsy evidence as best, and yet the key finding in the report regarding the PSA.
Then, there is one other key problem—this report does nothing to debunk the fact that the system is not performing better than the old system. We broke down the data provided by the New Jersey Courts, and as we noted, the results were not good. Obviously, there’s one other key question—was it worth the money? Apparently doing any sort of benefit-cost analysis is too expensive for Arnold Ventures despite the billions it has in the bank.
So, here we are again—Arnold Ventures is trying to once again re-arrange deck chairs on the Titanic to argue that their PSA works, when it clearly doesn’t. Indeed, RTI got off the ship early in a lifeboat to leave MDRC and their paid hack researchers to do the re-arranging. We’ve seen this movie before.