Virginia Pre-Trial Release: the True Risks of Recidivism and Failure to Appear for Trial
Share this article
ADVERTISEMENT
(comments below)
Comments
Comments
19 responses to “Virginia Pre-Trial Release: the True Risks of Recidivism and Failure to Appear for Trial”
-
Fellow that I knew in the 80s had a phone number was similar to the clerk of courts number. He used to get calls at 3AM asking, “Do I have to be in court tomorrow?”
“What’s your last name and DOB? … No. Case continued.”
-
Like the author, callers are waiting for a return call.
-
Say, just outta curiosity; “How many times do you have to be wrong to lose the title ‘expert’?”
This is from Fox Business yesterday: Former Trump economic advisers Larry Kudlow and Kevin Hassett dismissed Wall Street expectations of a 250K jobs report for July, insisting in this recession it would be “closer to 100 … way on the downside.”
Numbers just came out: 528K.
-
Hmmm, Democrats… Deficit Reduction Act.
Republicans… 14-yo rape victims forced to birth and Viktor Orban…Yep, all going as planned.
-
-
-
re: ” In 1995, Pretrial Services Agencies (PSA) were authorized (mandated for those seeking approval of jail projects) in Virginia by statute with the passage of the Pretrial Services Act, § 19.2-152.2.”
Who was Governor of Virginia at that time? George Allen?
-
Plans enacted by Republicans are always good plans until Democrats try to implement them…
-
or just follow along as they were originally implemented. George Allen, as Sherlock and others know was a the “no parole” guy. Serve your entire sentence, no time off for good…. so curious that this PSA was vetted during his administration.
-
-
-
re: “The PSA program as currently constituted has been proven by the state itself in the extensive and scientific Virginia Pre-Trial Data Project Preliminary Findings by the Virginia Crime Commission to have failed at that mission, the annual reports to the contrary by DCJS notwithstanding.”
If crime did not increase , does that constitute a “failure” if the goal was to incarcerate less people and reduce the expenses associated with that.
” VIRGINIA PRE-TRIAL DATA PROJECT: PRELIMINARY STATEWIDE FINDINGS ”
-
“… the annual reports to the contrary by DCJS notwithstanding.”
Who ya gonna believe, me or your damned lying eyes?
-
But, Larry, the goal, a good one, of PSA to reduce crime and FTAs in the pre-trial release program. But that goal has not been realized.
A major Virginia Crime Commission study reported the PSA program made no difference at all compared to courts and pre-release populations not served by PSA. It’s overall success rate was no higher than 62%. The other 38% represent a danger to the citizens of Virginia.
My article was written for two reasons: (1) to get the Youngkin administration to improve the PSA system or try something else; and (2) to expose that the DCJS annual reports to the governor and General Assembly have been at best wrong.
The evidence of ineffectiveness submitted by the Virginia Crime Commission is scientifically derived by external review. The reports of DCJS are derived from internal data. The gaps between the two are impossible to rationalize.
In order to make that case, I needed to introduce evidence of both issues.
I hope you and others support the goals and understand the evidence in those contexts.
You will note that I have updated the article with a new introduction to make those points clearer.
-
I see two goals. One to reduce the number of people held in prison if releasing them results in no increased crime. That saves money and resources.
The second that you like is to decide how to reduce crime which I think may be a questionable premise and would like to see results from other states and studies.
It’s really not a “failure” if there is no change in crime but a substantial savings in incarceration and related court and management costs.
A true “conservative” approach wants to look at the balance between releases, crime and costs and not just focus on one thing – crime and in this case – somehow developing a system that reduce crime. As I said above, lets see some evidence from other states that crime as been actually reduce by keeping more people in prison longer while awaiting their court cases.
-
-
-
This was a topic that I had been planning to write on. However, I have had, or taken, the time to re-acquaint myself with the issues and digest the findings of the State Crime Commission. Therefore, this response will be preliminary.
Most importantly, you have overlooked the main function of pretrial services agencies. The main impetus for their creation was the large number of people being held in jail awaiting trial, particularly those charged with minor, nonviolent offenses. Many in this group were indigent and could not afford to post bail.
The pretrial services agencies were intended to provide the courts an alternative to holding these individuals in jail–release them to the supervision of these agencies which could provide some services and assurances they would appear for trial.
As for some of your specific findings, I, too, was struck by the differences in outcome data between the Crime Commission report and the DCJS annual report. I can think of one rational explanation for the difference, but, rather than propose it with nothing to back it up, I have asked a contact within DCJS for an explanation.
You assert that the Crime Commission preliminary report showed “VPRAI was wrong as a predictive tool nearly half of the time.” I have gone over that report and could find no discussion of VPRAI. Is that a finding of the Crime Commission or your conclusion? If it is a finding of the Crime Commission, please provide the specific source. If it is your conclusion, please provide the rationale for that conclusion.
-
VPRAI was, and still is, the predictive tool in use by PSAs.
The predictions were faulty at least 38% of the time, higher if you take into account the limitations to scope of the data collected by the Virginia Crime Commission.
Perhaps because the tool was given data that assumed outcomes that did not come true. Perhaps because the pre-trial supervision was faulty. I don’t know, but more importantly, neither does the government of Virginia.
-
I will expand.
What we know from Virginia Crime Commission study is that at minimum of 38% of criminals released based on an assessment that used VPRAI as a predictive tool either committed another crime pretrial or failed to appear.
Such tools are algorithms fed data. The algorithm is a set of rules to be followed in computers in problem-solving operations.
This VPRAI algorithm is designed to predict the success of the PSA system of supervision given a specific person to supervise in pre-trial release. It is fed risk assessment data as in the example in my article. It’s output is supposed to predict whether the accused will commit another crime and show up for trial.
A very high percentage of the accused released as a result of that VPRAI assessment in fact committed another crime or failed to appear or both.
With that known result, both the algorithm and data inputs of VPRAI must be examined.
The contractor tasked to do that will have to do two things:
First, reenter the original data and modify the algorithm until it provides predictions that eliminate the errors.
Then, use the original algorithm and enter additional data until the tool provides predictions that eliminate the errors.
That is a professional way to isolate the causes of the errors. They may prove to be some combination of algorithm and data.
I suspect but do not know that the algorithm as written may significantly overestimate the success of the PSA pre-trial supervision program. That could explain how jurisdictions without PSAs performed the same with pre-trial releases as jurisdictions with them.
At any event, the PSA system, using the VPRAI tool, is dangerously ineffective. Those results are already in. That system needs to be fixed.
-
So, your assertion is based on your analysis, rather than a finding by the Crime Commission.
Your conclusion seems to assume that, for each defendant who failed to appear or who was arrested for a new crime, the VPRAI predicted no risk with a recommendation that the defendant be released.
The process is more complicated than that. First, the VPRAI process measures the degree of risk and assigns a level of risk, ranging from 0-6, for each defendant. The higher the level of risk, the greater the chance that the chance of failure if the charged person is released, i.e. failure to appear or arrest for new crime.
Second, VPRAI is a tool for judicial officers (magistrates and judges) to use in deciding whether to release a person who has been charged with an offense, whether to required secured bond, whether to release the person to supervision by a pretrial services agency, and what conditions, if any, to place upon the release.
The judicial officer is not required to follow any recommendation produced by VPRAI. That tool provides an assessment of risk. The risk level may be high, but the magistrate may choose to release the charged person, especially if a substantial bail is posted. The magistrate may choose not to release the person to supervision by a pretrial services agency.
Therefore, before making an assertion the VPRAI’s predictions are wrong half the time, one would have to analyze the level of risk assigned to each charged person and the result of the release of that person.
DCJS has conducted such a study, called validation, and has determined that the instrument reliably predicts success and failure. For example, for someone with a risk level of 6, it predicts any failure 37 percent of the time.
VPRAI has been adopted by jurisdictions throughout the United States and adapted to each individual jurisdiction’s situation and validated.
-
Thanks, Dick, but the process of assessment of risk and supervision of those released pre-trial has failed in Virginia a minimum of 38% of the time. The State Crime Commission report asserted it, not me. The DCJS can “validate” it until the cows come home. The PSA system is failing us. It is a public safety issue. The state needs to fix it.
-
You ignore the discretion that judicial officers have to use VPRAI. If VPRAI rates a defendant a level 6 risk and the magistrate or judge decides to release the individual on secured bond without any supervision and the person is arrested for a new crime, that is hardly a “failure” of the process of risk assessment.
-
How often does that scenario happen, Dick?
-
I don’t know if anyone has collected and analyzed the data in that way. I will check with my contact. I do know that circuit courts have shown some inclination to deviate from sentencing and probation guidelines recommendations.
-
-
-
-
Leave a Reply
You must be logged in to post a comment.