The number of people held in pretrial detention increased by more than 400% between 1970 and 2015. Yet research indicates that detaining individuals in jail while awaiting trial does not improve pretrial outcomes, such as lowering rates of rearrest or failure to appear in court, and instead significantly increases the probability of conviction while reducing employment outcomes. Additionally, pretrial supervision can be burdensome and expensive for both clients and government agencies, as clients face costs associated with monitoring, such as traveling for regular drug tests, and agencies must pay for new monitoring technology and hire additional staff to manage growing caseloads.
Many jurisdictions are reforming their pretrial systems to right-size their pretrial caseloads, including adopting data-driven approaches to decisions about pretrial release and “step-downs” in the conditions of pretrial release. To effectively implement these data-driven approaches, pretrial agencies and courts need to understand if their pretrial clients are successfully showing up to court appearances and remaining free of violations while awaiting trial. These measures of pretrial success are commonly captured with two metrics: appearance rates and public safety rates.
Many pretrial agencies establish success metrics using guidance from a report by the U.S. Department of Justice National Institute of Corrections called Measuring What Matters. In the report, these metrics are defined as:
With an understanding of pretrial client success rooted in data, pretrial staff and judges can better make the case for pretrial release instead of detention and can use less restrictive pretrial supervision conditions. This can benefit clients, who will face fewer financial and time burdens as they seek to remain compliant with supervision conditions, and pretrial staff, who may have more manageable caseloads, all while potentially resulting in public cost savings.
Through the Harvard Kennedy School Government Performance Lab’s (GPL) work with six pretrial agencies, pretrial staff often tell us that despite the potential benefits of measuring pretrial success rates, they face significant challenges collecting data to measure pretrial success due to limited data collection tools and staff capacity. The Pretrial Justice Initiative (PJI) has also identified challenges related to measuring pretrial success. In a 2019 survey by PJI, the majority of responding jurisdictions did not know their court appearance rate, either because the rates were unknown or the respondents were unsure if the data was collected.
In this brief, the GPL answers common questions from judges and pretrial staff, such as:
As pretrial agencies work to better measure the success of pretrial clients, consider the following takeaways from the GPL’s pretrial technical assistance:
1. Start collecting data to establish estimates rather than waiting to create the perfect measurements.
Even with incomplete data sets, pretrial staff can begin generating estimates based on available For example, if agencies have a record of which pretrial supervision cases were closed successfully or a record of individuals who have had any rearrests, they can create estimates of metrics like public safety rates as a starting point. These estimates can immediately inform the agency’s work and judges’ decision-making as staff members make ongoing improvements to the data system.
2. Standardize how front-line staff define violations and enter data.
Inconsistencies in how staff define violations and enter data into a case management system can create significant challenges in quickly and accurately measuring pretrial success First, agencies should align on which types of violations actually relate to pretrial outcomes in order to develop shared definitions and ensure they are capturing the most relevant information. Then, agencies should ensure officers enter data consistently. For example, one officer may only enter a violation if it was reported to the court, and another officer may track every violation even if it is handled internally via a warning. This can result in over- or under-reporting of violations since officers are using vastly different protocols for when they should enter violation data.
3. Establish protocols that track client success in real time.
As pretrial caseloads shift constantly, judges and pretrial staff need to have the ability to review pretrial success rates at a high frequency to identify trends and shape decision-making. By collecting data on violations as they happen, pretrial agencies are better positioned to share frequent, accurate data with judges.
For more on developing a standardized system for tracking client compliance in real time, read about our work in Harris County, Texas.
Many pretrial agencies have an internal case management system in which they track outcomes for people under their supervision. Some pretrial agencies also have access to additional data sets with information on pretrial client interactions with the justice system, such as from the sheriff’s office or court case management system. Promising approaches to measuring pretrial success differ between agencies that have access to additional data sets and those that do not.
Agencies that have access to external data sets can more comprehensively measure pretrial success rates in a way that places less burden on staff. In this scenario, pretrial staff can pull information on the same client across multiple systems, providing higher-quality data on instances of court appearances and rearrests and requiring less time for manual tracking.
However, many agencies only have access to their own case management system. In this scenario, there are still ways to ensure agencies can track pretrial client success, including officers tracking violations in a violations-specific part of the case management system, officers tracking violations more generally in the case management system, and officers selecting whether the pretrial client was successful or not when closing the case file. For both scenarios, the GPL has identified approaches that pretrial agencies can take to continuously measure pretrial success. The following tables provide details for each scenario.
There are three ways pretrial officers can collect data within existing case management systems, depending on what the agency’s data system allows for.
Once pretrial agencies have regular access to pretrial success metrics like appearance rates and public safety rates, staff and judges can use this information to right-size pretrial supervision and reduce burdens on clients and caseworkers. For example, with access to ongoing client success data, judges can monitor and adjust the intensity of pretrial supervision requirements — such as in-person check-ins, drug tests, and electronic monitoring — through the course of a client’s supervision period. Additionally, this is often the first time that judges can see the impacts of their release decisions and receive regular feedback about what is and is not going according to plan. Establishing this consistent feedback loop between the decisions judges are making and client outcomes can help inform judges’ future pretrial release decisions.
With support from the GPL, pretrial staff in Harris County, Texas, began tracking violations using a standardized case note process. Pretrial staff then analyzed compliance data weekly to determine individuals eligible to receive a step-down. From this analysis, judges received an individualized list of clients in their courtroom who were eligible for reduced supervision requirements. Because judges in the county typically have more than 1,000 individuals on their dockets at any time, reviewing a targeted list of clients allowed them to revisit conditions more efficiently and frequently. From October 2020 to June 2022, the agency successfully adjusted supervision conditions for more than 2,200 clients with no changes in client compliance or rearrest rates during that time.
The GPL is now actively applying its learning from Harris County to other jurisdictions, including the Superior Court of Alameda County (California), Clark County District Court (Washington), Illinois Office of Statewide Pretrial Services, Las Vegas Justice Court (Nevada), and Santa Cruz County Probation Department’s Pretrial Division (California).
One growing component of the GPL’s technical assistance is supporting pretrial agencies in redesigning how they measure pretrial success, allowing for regular tracking of pretrial outcomes to support the agency’s ability to right-size its supervision practices. This support often includes:
