Prosecutor-Led Diversion Unveiled: Lessons and Insights From Implementing Project Reset in Two Jurisdictions

Introduction

As caseloads and backlogs rise and state budgets decrease, there has been increased recognition that formal system processing might be doing more harm than good, particularly for low-level offenses (Bourgeois et al., 2019; Cauffman et al., 2021; Metcalfe & Kuhns, 2023; Wickman et al., n.d.). In response, there has been an increase in prosecutor-led diversion programs (Wright & Levine, 2021), allowing prosecutors to leverage their discretion to divert lower-level offenses to community-based services while still promoting accountability (Rempel et al., 2018; Wright & Levine, 2021). By diverting individuals from traditional system outcomes (e.g., conviction, probation, or jail time) and into community-based programming, prosecutor-led diversion programs can deliver proportionate responses to reduce the burdens that the system can place on people. Additionally, these programs often aim to address the underlying behavior that led to an arrest, while also allowing the individual to avoid the collateral consequences that result from a conviction. Recent research suggests that diversionary practices (e.g., deferred adjudication, dismissal of charges) for individuals charged with first-time criminal offenses can substantially reduce future criminal legal system involvement compared with those traditionally processed (Agan et al., 2023; Heaton et al., 2017; Mueller-Smith & Schnepel, 2021).

One such program is Project Reset (Reset), which was developed to be a prefiling^[1] diversion model in New York City. In this model, a prosecutor’s office establishes an office-wide policy to divert people charged with certain low-level offenses from traditional case processing. Following the success of the New York City program, two jurisdictions have adopted and adapted the Reset model for their communities: Jefferson County, Alabama, and Westchester County, New York. RTI International (RTI) partnered with the two prosecutors’ offices to evaluate these models and outcomes. This paper highlights the initial implementation of Reset in these two jurisdictions, describes program challenges and successes, and provides recommendations for other jurisdictions interested in implementing Reset in their communities.

Overview of Project Reset

Reset is a prefiling prosecutor-led diversion model to respond to low-level offenses with a brief community-based intervention. Program participation does not require an admission of guilt, and if an individual successfully completes the program, the prosecutor’s office will decline to prosecute the case without any other obligations (e.g., attending court, meeting probationary conditions).

In 2015, New York City district attorneys (DAs) in Manhattan and Brooklyn collaborated with the nonprofit Center for Court Innovation (now the Center for Justice Innovation) and the New York City Police Department to launch the first Reset pilots, serving 16- and 17-year-olds, in two neighborhoods: Harlem and Brownsville. The Reset model was eventually expanded to include all ages and was implemented throughout all five boroughs of New York City (http://www.projectreset.nyc). In the New York City models, an individual’s original arrest is sealed after successful completion of the program, reducing the collateral consequences of system involvement (Dalve & Cadoff, 2019). This program was designed partly to establish a proportionate response to the vast number of nonviolent violations and misdemeanors that pass through the New York City courts each year. Although offices that have implemented Reset have focused on low-level eligible charges that often are not likely to result in incarceration, the obligations associated with traditional case processing even for low-level charges—such as court attendance, supervision requirements and fees, and the criminal record created by a conviction—can be burdensome and lead to a host of consequences (e.g., lost or reduced employment, impacts to familial obligations; see Bergin et al., 2022; National Inventory of Collateral Consequences of Conviction [NICCC], 2024).

Recent evaluations of Reset in Manhattan found high participant satisfaction with the program, as well as success in connecting participants to community-based services and positively shaping participants’ view of the legal system (Anwar et al., 2023; Dalve & Cadoff, 2019). While evaluations of Reset have not found statistically significant effects on rearrest rates at the 1-year follow-up (see Anwar et al., 2023; Dalve & Cadoff, 2019), results suggest no measurable changes in public safety risk (through rearrests and convictions). Furthermore, the evaluation of the youth pilot identified notable benefits for Reset participants ages 16 and 17, who were less likely than matched nonparticipants to experience a new conviction or new arrest (Dalve & Cadoff, 2019).

Study Methods

This paper discusses the initial implementation of Reset in two communities: Jefferson County, Alabama, and Westchester County, New York. RTI, the evaluator of the two implementation sites, has met monthly with each of the implementation sites since May 2022. Additionally, RTI researchers conducted in-person, semistructured interviews at each site with the DA, key program staff, partner agencies, and court system actors to understand implementation progress and challenges (Jefferson County, N = 8; Westchester, N = 5).^[2] Information obtained from these meetings and interviews, including illustrative quotations from the qualitative interviews conducted in September 2023, underlies this implementation paper. We present limited case data, which were reported in aggregated counts by the Jefferson County and Westchester County program staff.^[3]

Characteristics of the Two Implementation Sites

Both sites implemented many of the core Reset model elements, but adjusted the model to fit its local process, needs, and priorities. Jefferson County’s model, called Reset Jefferson County, has made multiple adaptations from previous versions of Reset, including shifting the eligible offenses from misdemeanors to felonies, limiting the program to young adults, and eventually expanding the criminal history criteria to include individuals with prior misdemeanor arrests and convictions (except domestic violence). These adaptations were required to adjust for statutory differences between Alabama and New York State (e.g., many of the qualifying misdemeanors in New York would be Class D felonies in Alabama; Alabama uses traditional booking practice and does not have a desk appearance ticket model as in New York).

Westchester County’s version of Reset, renamed Fresh Start, is similar to New York City’s model, with expanded eligibility criteria that include violations such as disorderly conduct, trespassing, and possession of cannabis. Unlike Reset in New York City, Westchester County’s Fresh Start currently excludes people with prior arrests.

Both Jefferson County and Westchester County worked closely with community-based social-service providers in their communities to develop the structure, goals, and curriculum of a workshop that is essential to the model. See Table 1 for an overview of the program elements in Reset Jefferson County and Fresh Start.

Reset Jefferson County

Reset Jefferson County, led by the Jefferson County DA’s Office—Birmingham Division, was launched in early 2022, covering most of the county. The jurisdictional area of Bessemer is not part of the Birmingham Division and does not participate in Reset. One main change from the original Reset model is that Reset Jefferson County functions as a postfiling diversion program, and thus, participants must attend their first court appearance before being recruited. This change largely stemmed from a local need to provide more proportionate responses to Class D felonies (beyond the existing drug and theft courts); offer the office capacity to identify, screen, and contact the eligible participant, law enforcement office, and victim before first appearance; and provide the eligible participant an opportunity to consult with an attorney to ensure the program and requirements are an appropriate fit.

The first participant enrolled in March 2022. Originally, the program was designed for young adults, ages 18–24, who had no prior arrests and had been arrested for selected Class D felonies. The DA shaped the program in collaboration with the nonprofit advocacy organization Alabama Appleseed Center for Law and Justice. Since Reset was implemented, the eligibility criteria have been expanded twice. The first adaptation occurred in June 2022, when individuals arrested for selected Class C felonies were deemed eligible for Reset. In February 2023, the second adaptation occurred: individuals with prior misdemeanor arrests and convictions were also deemed eligible.

In describing the benefit of Reset over other alternatives to incarceration in the county, a public defender noted,

Reset would be highly more preferential than drug court for most of our clients because of the cost, the time. And also, when you do drug court, you have to plead guilty. It’s a conditional plea…. And you have to agree to an aggravator, which means in Alabama, if you fail out of drug court, an offense that you would otherwise not be eligible to go to prison for, they can send you to prison because you did drug court and didn’t finish.

The Reset program has a dedicated coordinator, employed by the DA’s office, who manages all aspects of the program, including recruiting participants; gathering consent from arresting police officers and, if applicable, victims; scheduling workshop sessions; and monitoring program completion. The program coordinator is also responsible for screening cases, which is done by obtaining daily reports of new warrants. Using this report, the coordinator flags eligible individuals and conducts criminal history checks. During the intake process, the Reset coordinator informally assesses any social, behavioral, or health needs and refers the participant to relevant services. The Reset coordinator also attends the workshop, which allows them to connect individuals with newly identified service needs.

Reset Workshop

The Reset workshop, facilitated by an external nonprofit, includes the Write to Heal/Re-Storied program, which is a 4-hour trauma-informed workshop guided by tenets of restorative justice. The workshop comprises videos related to restorative justice, group discussion and sharing, and writing prompts. In the group discussion, individuals discuss the meaning of justice, as well as the meaning of restorative justice broadly and on a personal level. Although the group discussion focuses on the legal system and its impact on their lives, program staff, including the workshop facilitator, described the workshop as a “healing space.” Participants have the opportunity to talk about their offenses but are not required to do so. Instead, the group discussion allows individuals to choose what they share, in a space that emphasizes feeling heard by others and engaging in personal reflection. Individuals are tasked with writing their goals and the services needed to achieve these goals, and then they are asked to share their responses with the group.

During the interviews, DA Danny Carr stated that the motivation for implementing the program was

[J]ust me wanting to have more of an understanding of why [the individual] did what they did and wanting them to have an understanding of how it affects the community. And I think that’s way more effective than a guy standing in front of the judge or a woman standing in front of the judge in a crowded courtroom—a judge saying, “Hey, this is what you’re sentenced to now. Don’t you get [in] any more trouble. If you get [in] any more trouble, I’m going to send you away for that year.”

Once the participant completes the workshop and pays any restitution and fees, the DA’s office enters the case as nolle prossed (i.e., nolle prosequi), meaning that prosecutors inform the court they are no longer pursuing the case.

Fresh Start

The Westchester County DA’s Office, in collaboration with Westchester County’s Department of Community Mental Health (DCMH), launched Fresh Start in two localities, White Plains and Greenburgh, in November 2021; implementation then rolled out across other municipalities. By June 2022, Fresh Start was available countywide. Fresh Start is designed primarily for individuals ages 18 and older, with no prior arrests, who are arrested for selected nonviolent misdemeanors and violations. (Individuals ages 16 and 17 who are charged with a more selected set of violations are also eligible for Fresh Start.) Fresh Start is both a prefiling and a postfiling diversion program (i.e., a hybrid preadjudication program), as the filing of charges is dependent on (1) the jurisdiction, and (2) whether DCMH was able to make contact with the individual before their court date. In large jurisdictions, police departments send desk appearance tickets to their assigned DA branch offices. The case is then screened for eligibility into Fresh Start by the assistant DA (ADA), who, if the case is eligible, sends the case to DCMH to begin recruitment efforts. In smaller jurisdictions, the DA branch offices do not receive the desk appearance ticket until the individual’s first appearance.

Regardless of the jurisdiction, DCMH recruits eligible participants once the organization receives the referral from prosecution and conducts the intake process over the phone to individuals interested in participating, which provides an opportunity to assess any issues or needs for workshop accommodation (e.g., whether participants need the workshop conducted in English or Spanish). DCMH also developed and delivers the Fresh Start workshop. Since its implementation, Fresh Start has not modified its eligibility criteria.

Fresh Start Workshop

The workshop is a 3-hour group-based program that uses elements and principles of restorative justice and cognitive behavioral therapy. Although the workshop is overseen by the DCMH, each session is led by a DCMH-employed facilitator and a formerly justice-involved peer, who functions as the co-facilitator. The workshop has four main goals:

1. Increase knowledge of the criminal legal system and use programming to address root behaviors driving criminal activities.

2. Help participants gain an understanding of how early diversion differs from conventional court processing.

3. Facilitate an opportunity for individuals to share with others and gain insight into their own experiences.

4. Make individuals aware of what resources are available in the community and how they can access them.

Similar to the workshop in the Reset Jefferson County model, the Fresh Start workshop is not focused solely on the pending charge; instead, individuals are given an opportunity to tell their stories. The DCMH program staff described a highly collaborative and positive relationship with the DA’s office and noted,

[At the onset of the program,] the DA really spoke a lot about trauma-informed care, recognizing the social determinants of health, and mental health and behavioral health and the impact that it has on our residents and not compromising safety, but really ensuring that we have diversion [to give] people an opportunity.

Staff further described adapting the existing Reset model implemented in New York City to fit Westchester County’s priorities and service needs:

A lot of [the original Reset model] was retained… the whole restorative model and how the criminal justice system is seen. [We] just added in more of the didactic piece so that it could be very interactive and wanted [participants] to be able to look at their own situation and identify how they could make the changes.

Notable Characteristics of Each Site

Several characteristics of the implementation sites are worth mentioning: (1) shared goals and objectives, (2) staffing, (3) court appearances, (4) program fees, (5) eligible offenses, (6) handling of restitution, and (7) sealing criminal history.

To start, both DAs’ goals and purposes in adopting the Reset model were expressly to provide a fairer and more proportionate response to lower-level offenses than traditional processing. Neither of the two implementation sites focuses its goals on cost savings to the court system or the DA’s office. This approach aligns with previous evaluations of the Reset model in New York City, which found little or no court cost savings (Anwar et al., 2023). In the site visit interviews, most of those interviewed did not feel that the program would lead to any substantial system-level cost savings, because the program required an investment of DA staff time and the use of county and community-based social services. At the time of the September 2023 site visit interviews, both implementation sites were beginning to think about promoting the program and its success to the wider community.

Both implementation sites use a staffing model in which the DA’s office conducts all eligibility screening, executes formal case dismissal/declination, and manages partnerships with behavioral health providers. This model helps keep the power, accountability, and discretion within the DA’s office while allowing trained behavioral and social service professionals to work with participants. Program participants in both sites are encouraged to work with connected service providers as needed after program completion and are provided information for how to obtain these services at little or no cost.

As previously noted, the two implementation sites differ in terms of court appearances. Although the New York City models of Reset were prefiling diversion programs, both implementation sites adapted the timing of recruitment to include or entirely adopt a postfiling, preadjudication model. In Jefferson County’s Reset, individuals must attend their first court appearance before being recruited. In Westchester County, Fresh Start aims to enroll participants before the first hearing. With such an approach, if an individual successfully completes the program, their arrest is sealed automatically, and the individual often never has to appear in court for this charge. For the most part, this approach occurs in larger jurisdictions; however, in smaller jurisdictions, individuals must attend their first court appearance before recruitment by DCMH begins.

Regarding program fees, neither implementation site charges participants to enroll or complete the program, which is a unique and important feature of the Reset model. For example, a Reset Jefferson County consultant who has assisted with program design and implementation stated,

I see [Reset] as unique in one way, in that it is a diversion program where the participant doesn’t have to pay. My organization looked into a range of drug court programs, other diversion programs…. And the biggest problem that we saw is they were cost prohibitive for the people who needed them the most, or interaction with the program dragged on for so long because it was dependent on paying off a debt or some sort of administrative or program cost.

The same sentiment was shared by individuals in Westchester County, who highlighted that offering a program free of charge may help eliminate inherent disparities in the criminal legal system.

Regarding eligibility, both programs are focused on similar types of charges, such as petit larceny/petty theft and simple possession of cannabis. Even though the underlying behavior and offenses are similar in nature, these charges are considered felonies in Alabama, whereas in New York they are misdemeanors or violations. In Westchester County, most of the Fresh Start eligible cases would be disposed of with a violation or an adjournment in contemplation of dismissal with traditional processing. As such, having a conviction and criminal history likely has more impact in Jefferson County, as felony convictions are associated with greater repercussions than misdemeanors (National Inventory of Collateral Consequences of Conviction [NICCC], 2024).

Restitution is handled differently across the two sites. Restitution is court-ordered compensation to be paid by the person charged to a crime victim. Depending on the jurisdiction, restitution can be set by the court or prosecutor’s office before the person has been convicted of the offense. In the Reset Jefferson County model, individuals’ cases are dismissed, which may remove the DA’s ability to enforce an individual’s compliance with restitution. As such, cases are not dismissed until restitution has been paid in full. It is worth highlighting that the amount of restitution owed by Reset Jefferson County participants is solely dependent on the amount that was not recovered (e.g., in a theft case, restitution is dependent on the value of nonrecovered stolen items). In Jefferson County, although the program does not exclude those owing restitution, it does often prolong the time between the booking event and case dismissal. In Westchester County, program staff reported that eligible cases thus far have not included restitution but individuals ordered to pay restitution for the pending charge may be considered for inclusion in the program on a case-by-case basis, as long as an arrangement for payment can be made or if the victim agrees to forgo restitution.^[4]

Finally, in both programs, the charges will either be dismissed or formally declined upon completion. In Jefferson County, the DA’s office can dismiss the case; however, there are no automatic procedures to seal or expunge the arrest from the individual’s criminal history. Under Alabama law, for the arrest to be removed, individuals must apply for expungement—a complicated, bureaucratic process that may take months or even years, incur fees, and be difficult to navigate without the assistance of an attorney. However, under New York State law, certain felonies and misdemeanors are automatically sealed once the case is dismissed or declined, meaning that these criminal records are not accessible to anyone other than certain law enforcement personnel (NYCourts.gov, n.d.)—an important mechanism for avoiding long-term collateral consequences. As such, all Fresh Start participants’ records are automatically sealed, with sealing initiated upon program completion.

Jefferson County and Westchester County Case Flow

As of November 2023, both programs had screened hundreds of charge-eligible cases (see Table 2). Jefferson County had graduated 33 individuals from the Reset Jefferson County program (which represents about 64 percent of those deemed eligible for the program). Westchester County had graduated 528 individuals from the Fresh Start program (which represents 71 percent of those deemed eligible for the program). The differences in completion numbers between the two models are mostly attributed to differences in eligibility (e.g., age restrictions for Reset Jefferson County) and case volume.

Implementation Successes

Early Buy-in From System Actors

Both sites described great success with their initial implementation of Reset to date and credit strong buy-in from local system actors. Most of the effort to implement the program was in adapting the Reset model for each jurisdiction. Sites mentioned requiring substantial time and discussions with system actors to create the parameters of the program, establish community and service-providing partnerships, and determine the initial communities in which to implement. Reset Jefferson County has received strong support from the local district court judges, the police, and the county’s public defender office. Reset staff credit their success in enrolling all eligible participants (as of November 2023) to the supportive relationship with the public defender’s office. Similarly, in Westchester County, the DA’s office has received support from local police, particularly in a few of the localities experiencing substantial petit larceny. For example, the City of White Plains police saw Fresh Start as an opportunity to respond to the growing calls from retail businesses to curb shoplifting while addressing the social and behavioral factors that likely drive petit larceny.

One of the interviewed law enforcement officers described their support of Westchester County’s Fresh Start:

You don’t want to ruin an 18-year-old’s entire life for a shoplifting arrest, which I think sometimes those things can be tied to behavioral health. Maybe they’re reaching out for some reason, maybe there’s a defiance phase or something like that. Do they need to have that on their record when they’re trying to get a job for the next 10 years and they’re labeled as a thief? Fresh Start makes sense. Kind of hit the reset button…

Staff and other interviewed stakeholders in each site noted that the benefits of the program were to increase desistance or avoidance from long-term system engagement. For example, as one prosecutor stated, the program “is less focused on caseload but instead [is] a way to provide needed services to people.”

Recruitment, Enrollment, and Completion

Both sites described great success in program recruitment, enrollment, and completion. As of November 2023, the Westchester County DA’s Office had declined to prosecute 528 Fresh Start cases, and additional cases were pending workshop attendance. When individuals were contacted, a very high percentage of them not only enrolled in but also completed the program. Of all screened eligible participants in Westchester County, approximately 75 percent have enrolled in the program, and 71 percent have completed it. The DCMH runs one or two Fresh Start workshops each week, which allows for quick resolution of the pending charge.

The case flow numbers in Jefferson County are lower than those in Westchester County, but Jefferson County has noted success in recruitment, enrollment, and program completion. In fact, all eligible participants have enrolled in the program. As previously noted, the program often has high-quality contact information and can recruit eligible individuals directly as well as receive referrals from defense attorneys.

Model Expansion

As a result of buy-in and relative ease of initial implementation, both sites have successfully expanded their program models since initial implementation. In Jefferson County, the program has expanded its eligibility criteria twice: once to include a slightly more serious class of offenses and again to allow participants who have some prior nonviolent misdemeanors. Similarly, Westchester County was able to expand its program countywide within a year of implementation. Sites also attributed their implementation success to starting with a conservative program model. Such an approach allowed them to iron out implementation logistics, garner system actor buy-in, and monitor program changes. With more implementation success came a stronger interest by the DAs’ offices in expanding the program model.

In discussing plans for model expansion, one of the DAs stated,

We felt like this was a comfortable starting spot because… we were trying to sell this to [the] community foundation, to the county commission, to the police chief. So we had to make it kind of restrictive for their buy-in. And now that we have [it] rolling, and ultimately if we get good numbers and show that the recidivism is down and all these things, then obviously we can kind of expand a little bit more. And you may get more buy-in from more community partners as well.

Implementation Challenges

Difficulty Collecting Comprehensive and Relevant Data to Measure Implementation, Disparities, and Outcomes

In both sites, program staff are collecting detailed enrollment, recruitment, and program completion data. Although these data are useful for understanding how the program is being implemented, both sites noted challenges in routinely gathering and analyzing comprehensive implementation data, particularly regarding all cases eligible for charges during the screening process. While the sites track information on enrolled program participants, they have had challenges to systematically tracking and easily describing those deemed ineligible due to criminal history or age—often due to limitations with case management system querying or limited staffing capacity to manually track. Without these data, it is challenging to evaluate any missed opportunities; identify disparities between populations who are offered the program and those who are not; identify implementation gaps, such as cases being referred only from specific localities within the jurisdiction; and ensure that the necessary data are available to analyze participants’ short- and long-term outcomes.

Case Flow Challenges in Jefferson County

In Jefferson County, the program has had notable periods with low month-to-month enrollment. This result is often attributed to court backlogs with first appearance and to a low number of individuals who have eligible charges and meet all the eligibility requirements—particularly the age restriction. Since expanding the eligibility criteria twice, Jefferson County has seen an increase in case flow; however, as noted earlier, there are still sizeable differences between the two sites, which is mostly attributed to different eligibility criteria, screening processes, and case volume. Additionally, as Table 2 shows, only about 6 percent of charge-eligible cases are ultimately offered the Reset Jefferson County program (51 deemed fully eligible out of 801 individuals screened and met the charge and age criteria). Interviewed program staff attributed stark reduction of eligible cases to most charge-eligible individuals having a disqualifying criminal history.

Recruitment Challenges in Westchester County Due to Poor Contact Information

Working with nonprofit service providers, New York City DAs’ offices chose to create the Project Reset model, which Westchester County adapted, to intervene and divert appropriate people out of the traditional process as early as possible. As such, this intervention happens soon after arrest and before the arraignment, which reduces the likelihood that a person will fail to show up at the first appearance and be subject to a bench warrant. In Westchester County, there has been some difficulty in reaching individuals because of incorrect or missing contact information; however, the program has consistently been able to recruit and enroll a substantial number of individuals once contact is made, which sometimes may be after an individual’s first appearance. These recruitment challenges are also similar to those seen in evaluations of the New York City models (Anwar et al., 2023); many of those eligible are not getting the opportunity to participate because of nonresponse or bad contact information. Anwar and colleagues (2023) suggested that recruitment challenges may further racial disparities between those with eligible charges and those who are actually enrolled.

Impacts of Restitution on Program Outcomes

While both programs have developed protocols for handling cases with restitution, eligible participants may face additional challenges in accessing or fully benefiting from the program. In Jefferson County, individuals owing restitution are allowed into the program, but their cases cannot be dismissed until all restitution is paid. The benefit of the Reset model is designed to be a prompt yet proportionate resolution to charges with minimal court involvement. However, interviewed Reset Jefferson County program staff reported some participants taking up to a year to complete restitution payments. During this period, the case remains open, delaying closure for both participants and victims despite completing all other requirements. Additionally, any new charges during this period could terminate an individual’s participation. In Westchester County, eligible cases with restitution are reviewed on a case-by-case basis for final determination of program invitation. Although no cases were disqualified on the basis of owing restitution as of November 2023, this discretionary review could introduce bias and reduce program availability for otherwise eligible participants (Llorente, 2016; Schlesinger, 2013).

Key Considerations of Implementing Reset

• Develop an intentional and strategic staffing model for prosecutors and community service providers. Jurisdictions interested in implementing Reset should identify the staffing model and associated resources needed to manage program logistics. Prosecutors may be able to provide initial data screening and data gathering support but likely will require additional staff, either internal or external, to carry out much of the day-to-day participant engagement. In each of the various Reset models, prosecutors used nonattorney staff and community-based services to manage the participant-facing operations of the program, connect to services, and deliver the program. This delineation of roles appeared to be optimal for program staff, other system actors, and participants. Memoranda of understanding and data-sharing agreements may be needed to establish role expectations and data-sharing protocols.

• Facilitate the process for sealing or expunging charges from criminal history and offer support as needed. To best take advantage of Reset and avoid collateral consequences, there should be minimal time, cost, and application barriers for individuals to successfully seal or expunge the current offense from their records. Program or other administrative fees can exacerbate disparities of criminal justice–related outcomes among underresourced communities by creating obstacles to program eligibility and timely completion (Llorente, 2016; Wright & Levine, 2021). Although individual offices may have little to no ability to change statewide policies on sealing or expungement procedures, offices interested in implementing the Reset model can take steps to reduce barriers to clearing program completers’ criminal histories. For example, offices can clearly explain to program participants the differences between case dismissal and declination with case sealing; thoroughly describe the steps participants would need to take to seal or expunge all criminal records; and where possible, provide or connect participants with low-cost legal assistance to help navigate the expungement process. In addition, providing guidance on how sealing an arrest or court case can affect employment, housing, or benefits applications can be crucial to fully leveraging the diversion program.

• Create and implement a robust evaluation to understand program impacts. Jurisdictions implementing Reset or any other prosecutor-led diversion programs should collect data beyond program enrollment and completion information of participants. To comprehensively understand program impact, programs should conduct a detailed evaluation that examines who is participating in the program and who does not (either not eligible, eligible but not reached, or eligible but refused to participate), differences and reasons for delayed or nonsuccessive completion, and participant and stakeholder perceptions of the program. Diversion programs should also evaluate outcomes against a comparison sample. Programs are encouraged to have an evaluator to codesign a data collection system to capture a range of implementation and outcome data (Nguyen & Tallon, 2024). As jurisdictions explore and begin the process of implementing a prosecutor-led diversion program, they are strongly encouraged to assess their data and staffing infrastructure to capture, analyze, and meaningfully update the program based on data results. Sites should explore creating a logic model that outlines program activities, resources, outputs, and expected outcomes. This logic model can assist in identifying key program outcomes and can support planning and resource needs.

• Develop communication and branding strategies to increase buy-in and sustainability. It is essential for the long-term success of the program to have stakeholder buy-in from both the local criminal legal system and the larger community. Program staff are encouraged to identify other important short-, intermediate-, and long-term outcomes to clearly describe the potential theory of change and expected program benefit. Furthermore, programs are encouraged to develop communication strategies to promote themselves and their results throughout the community to ensure greater buy-in to and awareness of the program, both of which may increase recruitment of and trust by eligible individuals.

Summary: Reset and Fresh Start

The original Reset model is a prefiling, prosecutor-led diversion program that provides case declination or dismissal after the completion of a brief, community-based, behavioral health intervention. In adopting this model, Reset Jefferson County and Fresh Start target low-level offenses during the early stages of the prosecution process. Both programs offer individuals the opportunity to avoid a criminal conviction while also receiving services, which may reduce recidivism and the size of the correctional population. Furthermore, as highlighted in both sites, the program itself poses no cost to the participant and provides a host of benefits to the individual, such as avoiding multiple court dates and fees, avoiding a criminal conviction, and potentially having collateral consequences eliminated or reduced. Through a curriculum focused on restorative justice and goal setting, individuals are given an opportunity to reset their lives and prevent long-term engagement with the criminal legal system. After a year of implementation, both sites described notable success in the initial implementation of the program and felt they were providing important second chances and needed services to individuals while promoting accountability.

Data Availability Statement

The data supporting the current study will be made available at the conclusion of this evaluation and will be posted on Open Science Framework under project title “Diversion in Westchester and Jefferson Counties: A Policy Study of Process and Outcomes” (https://osf.io/yvaw7/).

Acknowledgments

This paper was generated with the support of Arnold Ventures. The authors would also like to recognize the supportive and thoughtful contributions of Adam Mansky, who as Director of Criminal Justice at the Center for Justice Innovation (formerly the Center for Court Innovation), helped conceive and implement the initial Project Reset pilots and expand the program across New York City. Mr. Mansky introduced the model to the district attorneys’ offices in Jefferson County, Alabama, and Westchester County, New York, and he continues to provide implementation technical assistance to the sites and subject-matter expertise to the evaluation. We express our gratitude to the implementation site teams at the Jefferson County District Attorney’s Office, Westchester County District Attorney’s Office, and Westchester County Department of Community Mental Health for their cooperation and active participation in the execution of this evaluation. We also extend our sincere appreciation to Alabama Appleseed Center for Law and Justice and other community partners and system actors who participated in our site visit interviews, which greatly enriched the information presented in this paper. We extend our gratitude to Megan Comfort and the RTI Fellows Program for their support in the development of this paper.

RTI Press Associate Editor: Jeri Ropero Miller