Whatcom work group reviews PSA validation, eyes redesign of release-conditions grid
Loading...
Summary
Whatcom County’s pretrial work group reviewed a local validation of the Public Safety Assessment, heard that a Washington ethics advisory opinion limits pre‑court contextual intake, and discussed redesigning the county’s release conditions matrix to better match local data and resources.
Judge Jones opened the Pretrial Processes Work Group meeting by saying the group would review a recent validation of the county’s Public Safety Assessment (PSA) and consider how the results should change local practice. "We commissioned a validation study of the PSA... That was completed last year by Dr. Hamilton," the judge said, and the group had just received the results.
Mallory Hamilton, program director for pretrial services at Whatcom County Superior Court, summarized three recommendations from the validation: recalibrating the PSA for Washington populations, expanding assessment beyond criminal-history measures with a dynamic intake, and redesigning the local release conditions matrix (RCM) that maps PSA scores to supervision and conditions. "He gave us 3 focused recommendations," Hamilton said, and she recommended the RCM redesign as the immediately practical option because of legal and operational constraints.
Hamilton and others told the group that a Washington State judicial ethics advisory opinion (No. 18‑04) restricts court employees from collecting contextual, pre‑appearance personal information that could later be conveyed to a judicial officer. Hamilton said that limitation creates a barrier to adding a pre‑court contextual intake administered under court supervision: "In that ethics opinion 18‑04... it talks about the rights of individuals that are incarcerated and that we not gain any more information on them around their personal life," she said.
Jessica Ireland, senior manager at the Center for Effective Public Policy, explained how an RCM is used after the release decision to match PSA scores to appropriate supervision levels and conditions. "Once you decide release, it helps identify appropriate release conditions that should help improve the likelihood the person will appear in court and not be rearrested," Ireland said. She reviewed research findings the group should consider—court‑date reminders consistently improve appearance; the evidence is mixed for drug testing and electronic monitoring; and supervision should follow the risk principle so limited resources are targeted where they will have the most effect.
Members pressed for more descriptive data before revising the RCM. Several participants asked whether the validation report could show which PSA/NCA score groups were performing poorly and whether charge-based overrides in the existing software affected results. Ireland recommended ‘‘peeling back’’ the data: jurisdictions commonly examine who is failing to appear, how failure varies by case type and time, and whether overrides or other practices skew results.
Allen, a pretrial services officer, described a current quality‑assurance step already implemented locally: "On March 30, we put the assessor's worksheets kind of in practice, and we started sending those out to the judicial officers and the public defenders and the prosecution to show them where we're getting our answers," he said. The worksheet is intended to show the sources used to populate PSA answers and to increase transparency.
The group also discussed how bail and pretrial supervision interact. Some members asked whether the frequent use of cash bail in superior‑court cases confounds validation findings; Ireland said validations typically include everyone who had a PSA completed and was released, and suggested a separate analysis to compare outcomes for those placed on supervision versus those who were not. On the policy question of combining bail with supervision, Ireland said jurisdictions often do both but that money is not a proxy for public safety: "Posting money does not make you a safe release," she said.
Members raised victim‑safety concerns when considering monitoring tools such as GPS for domestic‑violence cases. Carol suggested inviting domestic violence and victim‑survivor stakeholders to future RCM discussions so the matrix can reflect victim‑safety needs alongside defendant‑focused outcomes.
Next steps identified by the group included obtaining more descriptive statistics from Dr. Hamilton’s validation to pinpoint which score groups are having the worst outcomes, continuing external review of the validation by independent organizations, and working with stakeholders to revise the RCM if the data supports changes. The meeting ended with no formal vote or motion taken; participants agreed to continue follow‑up work on data, RCM design, and stakeholder engagement.
