Research Paper: Restorative Justice

profilefarra004
Darling-Hammond2020_Article_EffectivenessOfRestorativeJust.pdf

Effectiveness of Restorative Justice in US K-12 Schools: a Review of Quantitative Research

Sean Darling-Hammond1 & Trevor A. Fronius2 & Hannah Sutherland2 & Sarah Guckenburg2 & Anthony Petrosino3 & Nancy Hurley4

# California Association of School Psychologists 2020

Abstract This article provides an overview of restorative justice (RJ) in US K-12 schools, discusses implementation challenges, and summarizes the most recent two decades of quantitative studies regarding the effectiveness of RJ at achieving a range of outcomes. While RJ has become increasingly popular, there is still relatively little quantitative research regarding its effective- ness. Still, available evidence suggests that RJ programs can improve school climates and reduce student misbehavior and school discipline. Results are more mixed regarding RJ’s impact on bullying, student absenteeism, and academic performance.

Keywords Restorative justice . Restorative practices . K-12 schools . Quantitative research . Quantitative effects . Student misbehavior . School discipline . Bullying . Discipline disparities . Attendance . Absenteeism . School climate . School safety .

Social and Emotional Learning . Academic outcomes

Restorative justice (RJ) is a broad term that encompasses an array of non-punitive, relationship-centered approaches for addressing and avoiding harm. While K-12 schools in the USA have increasingly turned to RJ as a means of improving school climate, reducing discipline and discipline disparities, and achieving other educational goals, quantitative research regarding the effectiveness of RJ in schools is nascent and has only recently grown more sophisticated. This article provides an overview of RJ in US K-12 schools, discusses implemen- tation challenges, and summarizes the most recent two

decades of quantitative studies regarding the effectiveness of RJ at achieving a range of outcomes.

Overview of Restorative Justice

Experts generally agree that the first recorded uses of RJ oc- curred in the pre-modern era in the South Pacific and Americas (e.g., Zehr 2002). These cultures had an approach to conflict and social ills that emphasized the offender’s ac- countability for the social harm they caused, along with a plan for repairing the hurt and restoring the offender to acceptance. The emphasis on the harm done rather than the act is a widely recognized principle across the RJ literature. Perhaps the most well-known framework for understanding RJ is called “rein- tegrative shaming theory” (Braithwaite 2004). Critically, rein- tegrative shaming is distinct from typical “negative shaming” in that the ultimate goal is not simply deterrence, but recon- ciliation between all parties involved and reintegration of the offender back into the community. Reintegrative shaming acknowledges the impact of wrongdoing on both the offender and those who were harmed. Zehr (2002) suggests that RJ requires society to move away from a system that emphasizes traditional retributive justice (“an eye for an eye”). Morrison and Vaandering (2012) argue that a system influenced by RJ would define “laws and rules as serving

Electronic supplementary material The online version of this article (https://doi.org/10.1007/s40688-020-00290-0) contains supplementary material, which is available to authorized users.

* Sean Darling-Hammond [email protected]

Trevor A. Fronius [email protected]

1 University of California, Berkeley, Goldman School of Public Policy, 2607 Hearst Ave., Berkeley, CA 94720, USA

2 WestEd, 300 Unicorn Park Drive, 5th Floor, Woburn, MA 01801, USA

3 WestEd, 1140 3rd St NE Suite 360, Washington, DC 20002, USA 4 WestEd, 9 Hawthorne Place, Unit 5L, Boston, MA 02114, USA

https://doi.org/10.1007/s40688-020-00290-0

Published online: 24 May 2020

Contemporary School Psychology (2020) 24:295–308

people to protect and encourage relationships and relational cultures” (p. 145) rather than protecting the status quo.

Today, many institutions in the USA—such as juvenile courts, domestic violence courts, universities, and K-12 schools—use RJ as a means to divert people from punitive justice systems, reduce rates of recidivism, and repair and improve social connections. While RJ has recently achieved considerable prominence in the USA, other countries have been utilizing RJ for some time. New Zealand, for example, has used RJ as a central framework in its juvenile justice system for more than a quarter century (Zehr 2002). After reviews of RJ systems around the world (e.g., Sherman and Strang 2007) suggested that RJ can reduce recidivism, im- prove victims’ mental health and coping, and cut costs, many US court systems adopted the practice. Bazemore and Schiff reported in 2005 that there were 773 RJ programs in place across the nation. In the years since Bazemore and Schiff’s census, educators have turned to RJ in large measure out of concern that many students suspended or expelled from school become part of a “school-to-prison pipeline” (Losen 2015, p. 241).

According to Zehr (2002) and others (e.g., Karp and Breslin 2001; González 2012), in the K-12 setting, RJ is meant to bring together all stakeholders—students, teachers, staff, and parents, as appropriate—to resolve issues and build rela- tionships rather than control student misbehavior through pu- nitive exclusionary approaches. RJ programs in schools range from training for teachers in relationship-building dialogue techniques to professionally guided restorative conferences with students and staff, and sometimes involving community and family members.1 RJ is often used as an add-on to existing school discipline approaches. A common “add-on” approach is to divert some students who would typically be suspended for their offenses to engage in a restorative process, such as a harm repair circle, where they are guided to understand the harm their actions caused, take steps to repair the harm, and make a plan to avoid making the same mistakes in the future. More recently, some schools have embraced a more expan- sive, “whole school” model of RJ as a preventative approach for building an interconnected school community and healthy school climate in which punishable transgressions are less common (e.g., Brown 2017). Such RJ programs often include universal training of staff and students in RJ principles. Both “add-on” and “whole school” RJ programs are often embed- ded within other schoolwide initiatives, such as Social and Emotional Learning (SEL) and Positive Behavioral Interventions and Supports (PBIS).

As documented by Bazemore and Schiff (2009), there are a variety of practices that fall under the RJ umbrella that schools may implement. These practices include victim-offender me- diation conferences; group conferences; and various “circles”

that include peace-making (i.e., restorative), talking (i.e., classroom/community), and re-entry (i.e., welcome). Conferences and circles fall in two categories: community- building circles, which are preemptive and designed to help students and staff deepen relationships and trust; and peace- making circles, which bring together parties who were in- volved in or impacted by harmful actions. In the latter case, participants include the victim(s), offender(s), and facilita- tor(s), but may also include other community members (e.g., witnesses, friends, family). The victims could also include members of the school community who represent the school that was harmed by the perpetrator’s actions (e.g., in the case of vandalism). Together, the conference participants aim to determine a reasonable restorative sanction for the offender. Restorative sanctions are sought out during these justice pro- cesses rather than employing traditional punitive sanctions such as suspension. Restorative sanctions could include com- munity service, restitution, apologies, or agreements to change specific behaviors, such as the offender agreeing to comply with certain conditions, sometimes in exchange for incentives (Stinchcomb et al. 2006). As evidenced above, RJ can take many forms, and while some models appear nearly identical to what one might find in the juvenile justice system (such as the Balance and Restorative Justice, or BARJ, model), others were developed specifically for school communities (Mirsky 2007; Mirsky and Wachtel 2007).

RJ is being implemented in schools and districts across many states to varying degrees. However, in a few states (e.g., California, Colorado, Illinois, Minnesota, and Pennsylvania), RJ has been implemented in schools for many years, evidenced by the presence of more large-scale and, thus far, sustainable programs. It has also been implemented in various environments and at varying scales—in public, pri- vate, or alternative schools; urban, suburban, or rural schools; and at the scale of a single classroom or across entire, large school districts such as the Denver Unified School District.

Proponents often turn to RJ as a means to avoid exclusionary disciplinary actions, which, research sug- gests, are associated with harmful consequences for children, such as academic disengagement (Losen 2015), racial and socioeconomic disparities in punish- ment (Skiba et al. 2002), and subsequent increased youth involvement with the criminal justice system (i.e., the “school to prison pipeline”) (Petrosino et al. 2012). Researchers have found that, controlling for a number of other relevant factors, suspensions alone ex- plain 20% of the Black-White academic achievement gap (Morris and Perry 2016) and that students suspended just one time are twice as likely to drop out of school, 44% less likely to graduate, and 49% less likely to attend a post-secondary institution (Balfanz et al., 2015). Proponents argue that RJ ad- dresses root causes of misbehavior without reliance on1 See the Appendix A for a list of restorative justice terms.

296 Contemp School Psychol (2020) 24:295–308

harmful exclusionary discipline, all while improving school climate and academic engagement.

RJ has also gained popularity as a means of addressing “disproportionalities in exclusionary discipline”—a phenom- enon whereby some groups of students receive exclusionary punishment at higher rates than others. A recent report by the US Government Accountability Office found that while Black students represent 15.5% of all students in the country, they represent 39% of students suspended from schools, and that while students with disabilities represent 13.7% of all stu- dents, they represent 25.9% of those suspended (Government Accountability Office 2018). Gregory et al. (2016) report that Black students are 26.2% more likely to receive out-of-school suspension for their first offense than White students. Many other studies indicate disproportionalities related to race and disability status (Losen 2015). Advocates argue that RJ can help reduce these disparities by facilitating positive student-teacher relations re- gardless of student demographics (e.g., Gregory et al. 2016).

Finally, RJ advocates (e.g., Tyler 2006; Zehr 2002) argue further that traditional discipline approaches man- age student behavior rather than developing students’ so- cial and emotional capacities. It also establishes a power dynamic between teachers and students (and at times be- tween students) that is detrimental to all students having a voice and feeling empowered. Tyler (2006) argues that by giving people, particularly students, a voice in the decision-making and procedural justice process, they will view institutional power as more legitimate and just. Tyler also makes the case that empowering youth may lead to better self-regulation without the need for formal disci- pline (Tyler 2006). Zehr (2002) and others argue that RJ results in a shift in how discipline is applied, which in- creases student perceptions that educator actions are fair, thereby leading to greater compliance as students see the school order as one having legitimacy. Sherman (1993) argues that when students do not perceive the school or- der as legitimate and fair, they may be more likely to defy and misbehave.

While research and writing regarding the theory under- lying RJ in schools is rich and well developed, the re- search on the effects of restorative practices in schools is still budding, but has grown considerably in recent years. The vast majority of studies are qualitative in nature, but there are an increasing number of quantitative studies. Until recently, no study on RJ in US schools employed causal methods or methods like randomized controlled trials, which can help determine whether a given program causes (rather than is simply correlated with) certain de- sired outcomes. However, two recent randomized con- trolled trials and other quasi-experimental designs have provided a clearer lens into the impacts of RJ implementation.

Implementation Challenges and Pathways to Successful Implementation

Available literature underscores that schools and districts implementing RJ face many challenges. Initially, owing to the diverse array of RJ programs, there is confusion about what RJ is and how to implement it. RJ also requires staff time and buy-in, training, and resources and can thus be more resource intensive than discipline regimes that rely on tradi- tional sanctions such as suspension. With RJ, teachers are often required to perform duties that would traditionally be outside of their job description, such as attending RJ trainings, conducting circles during instruction time, and spending more time talking one-on-one with students. Also, research suggests educators and other stakeholders may at times perceive RJ as being “too soft” on student offenses (Evans and Lester 2013). Finally, while RJ programs will certainly vary by the size of the school and scope of the program (Sumner et al. 2010), some researchers suggest that a shift in attitudes about rela- tionships and punishment may take one to three years (Karp and Breslin 2001), and the deep shift to a restorative-oriented school climate might take three to five years (Evans and Lester 2013). This timing assumes that the program will also be sustained financially, which underscores the importance of considering what resources will be needed and for how long to introduce and sustain RJ in a school or district.

There are foundational steps that schools and educators should consider when implementing RJ in their school or district, including funding, preparation, and sustainability.2

Our review of the literature indicates that RJ is perceived to work best when it is integrated into the school’s overall phi- losophy (Ashley and Burke 2009), school culture (González 2012; Brown 2017), or ethos (Beckman et al. 2012). The most common goal in embedding RJ in the overall school culture is to create an environment that is respectful and tolerant (Hantzopoulos 2013), accepting (González 2012), and sup- portive (Mirsky and Wachtel 2007). And a key pathway to fostering such a culture is proactively nurturing relationships among students and staff that are characterized by active lis- tening and respect (Brown 2017; Cavanagh et al. 2014). In addition, regardless of whether the RJ program is of the “add- on” or “whole school” variety, educators and administrators will benefit from targeted tools and sustained resources that allow them to successfully implement, evaluate, and continu- ously improve the program. These include, but are not limited to, professional development, coaching, and school time de- voted to RJ activities and reflection.

2 While the focus of this report is on the effectiveness of RJ programs, see the WestEd Justice and Prevention Website for a list of related reports on restor- ative justice in schools, including our 2019 review of RJ literature which included a section on implementation considerations for practitioners as well as a report based on data collected from practitioners in the field. https://jprc. wested.org/project/restorative-justice-practices-in-u-s-schools/.

297Contemp School Psychol (2020) 24:295–308

Methodology of this Quantitative Research Review

This review summarizes all available quantitative research regarding the effectiveness of restorative justice (RJ) in US K-12 schools published between January 1999 and December 2019. By “quantitative research,” we mean research that em- ploys “quantative methods” (such as regression analyses, non- parametric models, t tests, and analyses of variance) to analyze student, teacher, parent, school, and/or district-level datasets. By “the effectiveness of RJ,” we mean the ability of RJ pro- grams to realize specific, measurable student and school out- comes: less student misbehavior, less bullying, less exclusion- ary discipline (suspensions, expulsions), smaller racial disci- pline disparities, improved student attendance, improved school climate, and improved academic outcomes. To locate literature, we reviewed search results using the terms “restor- ative justice” and “schools” in ProQuest Social Sciences; the University of California, Berkeley, online library of scholarly texts; and Google Scholar. We also reviewed literature sug- gested by a network of experts who were interviewed for a related report (Guckenburg et al. 2015).

Internal Validity of Available Quantitative Research

Unless otherwise indicated, each of the studies described be- low should be viewed as correlational in nature, meaning they do not use methods that can ascertain whether RJ programs cause certain desired outcomes. There are, however, notable exceptions. Two studies utilize randomized controlled trials with large samples and a few others employ quasi- experimental methods to attempt to tease out causal effects. We indicate throughout where studies utilize these more rig- orous methods. We also provide more detailed analysis of these more rigorous methods when they appear.

Effectiveness of Restorative Justice in Schools

This section summarizes quantitative research regarding the effect of RJ implementation in US schools on misbehavior and student discipline; bullying; racial disparities in discipline; attendance and absenteeism; school climate and safety; and academic outcomes. It also summarizes quantitative research regarding racial disparities in access to RJ. Where appropriate, subsections begin descriptions of the theory underlying why and how RJ might drive positive effects on these outcomes.

Student Misbehavior and School Discipline

As noted previously, the RJ theory suggests that a well- implemented program could reduce punitive disciplinary ac- tions and problem behavior over time (Tyler 2006). Nearly all

of the empirical studies we reviewed report a decrease in ex- clusionary discipline and harmful behavior (e.g., violence) after implementing an RJ program. These two phenomena (misbehavior and discipline) are related but distinct. This dis- tinction is critical because many RJ programs are “suspension diversion” programs, which take students who would have been suspended under prior discipline plans and instead send them to engage in restorative proceedings. Almost by default, such programs reduce rates of exclusionary school discipline. These reductions may or may not be related to concomitant reductions in school misconduct. However, as discussed in the introductory sections, research suggests that exclusionary dis- cipline is associated with myriad negative outcomes (e.g., dropping out of school and being incarcerated). Thus, assessing whether RJ programs are successful at reducing ex- clusionary discipline rates, overall and for particular student subpopulations, may be worthwhile regardless of whether the reductions in exclusionary discipline rates correspond with drops in misbehavior.

To determine if RJ causes a reduction in exclusionary dis- cipline (among other outcomes, discussed later), Augustine et al. (2018) conducted a randomized controlled trial (RCT). RCTs are a common mechanism for teasing out causal effects. By randomly assigning some schools to receive the “treat- ment” (here, RJ training), researchers can be sure that the only systematic difference between treated and control units is treatment status, allowing researchers to attribute differences in mean outcomes to treatment, rather than other potential confounders. Critically, though, RCTs in educational contexts commonly face two internal validity concerns. First, they may suffer from compliance issues among either treated or control units. Compliance issues take on two whereby “treatment” schools may struggle to implement the program as intended or whereby “control” schools implement the program (or something similar to the program) and thus become poor com- parators. Secondly, they may suffer from timing issues where- by RCTs are not run long enough to detect effects that take time to materialize. As noted in the prior section, some re- searchers suggest successful implementation of school climate transformation programs can take three to five years (Evans and Lester 2013). Thus, when researchers evaluate school climate programs over shorter time frames, they may not fully observe program impacts.

Augustine et al. (2018) conducted their RCT with 44 mid- sized Pittsburgh schools serving kindergartners through 12- graders. They randomly assigned 22 schools to receive RJ training by the International Institute of Restorative Practices (IIRP) and 22 schools not to receive IIRP training, and then reviewed outcomes in the 2015–2016 and 2016–2017 school years. The authors note that their RCT may suffer from both the internal validity concerns described above as they did not verify that “control” schools did not use RJ and could serve as good comparators and they only measured outcomes two

298 Contemp School Psychol (2020) 24:295–308

years after implementation, meaning full results from the RJ trainings may not have yet materialized at the time of their evaluation.

These caveats notwithstanding, Augustine et al. (2018) es- timated that RJ implementation caused a 16% reduction in days lost to suspensions, which was statistically significant (p < .05). The reported reduction in suspension days was sta- tistically significant among certain student subgroups, includ- ing Black, low-income, female, and special needs students, as well as students in grades 2–5 and grades 10–12. In a similar vein, Armour (2014) reports a 65% drop in suspensions for 6th graders, and a 47% drop in suspensions for 7th graders, in one Texas middle school after RJ implementation. Denver (CO) schools that implemented restorative circles and confer- encing experienced a 44% reduction in out-of-school suspen- sions and an overall decrease in expulsions across the three- year post-implementation period (Baker 2009). In Oakland (CA), Cole Middle School experienced an 87% drop in sus- pensions across the first two years of implementation, com- pared to the prior three years, and expulsions were eliminated entirely after RJ was put in place (Sumner et al. 2010).

Many studies simultaneously note reductions in school discipline and reductions in student misbehavior. For example, Davis (2014) adds to research by Sumner et al. (2010), reporting more recent figures from Oakland. Davis’s (2014) research suggests continued success, with a 74% drop in suspensions and a 77% decrease in referrals for violence during a two-year follow-up. In a summary of findings from several individual reports, Lewis (2009) identifies positive results across schools that have implemented RJ. For example, the West Philadelphia High School reports that “violent acts and serious incidents” dropped 52% in the first year of RJ implementation; this initial drop was followed by an addition- al 40% drop through the first half of year two (Lewis 2009). McCold (2002) reports that RJ reduced offending by 58% for youth participants in an alternative education program in Pennsylvania during a three-month follow-up. Based on a follow-up study of the same program, McCold (2008) reports that effects were sustained through two years of implementa- tion, with reductions in offending of around 50%. In both studies, McCold (2002, 2008) reports that recidivism rates were significantly related to youth’s length of participation in RJ, with youth who completed the program showing a greater reduction compared to those who were discharged early. McCold’s (2002, 2008) analyses indicate positive in- creases in self-esteem and pro-social attitudes for “stayers” versus “leavers,” which may point to a possible mechanism for why participants who completed the alternative education program did well in terms of reduced recidivism rates.

Riestenberg (2003) notes that schools that offered intensive training and follow-up for staff demonstrated positive results across a range of discipline outcomes. For example, one ele- mentary school experienced a 57% drop in discipline referrals,

a 35% drop in average time of in-school suspensions, a 77% drop in out-of-school suspensions, and only one student was expelled during the one-year follow-up. Results from other schools in Minnesota with strong training are similar (45 to 63% decrease in suspensions, for example) (Riestenberg 2003). Importantly, Riestenberg notes that the studied schools participated in different mixes of programs and had different approaches to collecting data and defining outcomes and thus are not comparable.

McMorris et al. (2013) report similarly positive results from their study of the “Family Group Conferencing” model adopted in Minnesota. In this model, the offender and victim do not meet face-to-face in the conference (distinguishing it from most types of restorative conferencing). Instead, family members, school staff, and the offending student work togeth- er to develop a plan to ensure that the youth takes responsibil- ity for the youth’s actions, improves any harmed relationships, and takes steps to ensure that the youth does not make the same mistakes in the future. The researchers report a decrease in self-reported incidents of physical fighting and skipping school among conference participants in a six-week follow- up. In addition, participants who were referred to the program experienced a drop in suspension rates and gains in atten- dance, credit accrual, and progression toward graduation in the year following implementation of the conferencing program.

Goldys (2016) reports that at an elementary school, RJ implementation yielded a 55% decrease in office referrals. González (2015) reports that, during RJ implementation from 2006/2007 to 2012/2013, the suspension rate at Denver Public Schools dropped from 10.6 to 5.6%, with concomitant drops for Black students (17.6% to 10.4%) and Latino students (10.2% to 4.7%). More recent analysis of Denver data from 2008 to 2015 indicates a similar trend—a drop in the suspen- sion rate from 7.4 to 3.6% (Gregory et al. 2018). A 2018 analysis of Los Angeles Unified School District’s discipline records following the implementation of RJ in the 2014/2015 school year indicates that suspension rates for misconduct dropped for all measured categories of students (Hashim et al. 2018). Another research study focusing on one high school’s implementation of RJ reports a drop in suspensions as well. The out-of-school suspension rate dropped from 12 to just 7% over the five-year period of the school’s RJ imple- mentation, from 2010/2011 to 2015, and the in-school suspen- sion rate dropped from 19 to 7%. The number of repeat in- fractions fell steadily over this time period as well, from 111 to 34, and the number of repeat out-of-school suspensions dropped nearly in half, from about 50 to about 28 (Fowler et al. 2016).

Gregory et al. (2016) review thousands of school records from two diverse East Coast high schools, and surveys from 412 students and 29 teachers, to ascertain if teachers’ levels of restorative practice utilization are related to student behavior

299Contemp School Psychol (2020) 24:295–308

and school discipline outcomes. They find that students who indicated being in classrooms that utilized an above average (“high”) degree of restorative practices received fewer defi- ance and misconduct referrals than those who indicated they were in classrooms that utilized a below average (“low”) de- gree of restorative practices. Specifically, among Black and Latino students (collectively), those in low restorative practice classrooms received three times more defiance and miscon- duct referrals than those in high restorative practice class- rooms and this difference was statistically significant. Among White and Asian students (collectively), those in low restorative practice classrooms received 2.2 times more such referrals, but the difference was not statistically significant.

Jain et al. (2014) looked at students in Oakland (CA) who participated in two RJ programs: Whole School Restorative Justice (WSRJ) and Peer Restorative Justice (Peer RJ). They note that students were selected for WSRJ in part because they had higher suspension rates than average. After three years, these WSRJ students received statistically significantly fewer suspensions than students in the district overall, and fewer than students in Peer RJ.

Some recent dissertations have also reported positive re- sults for RJ on exclusionary discipline and misbehavior. McCain (2015) summarizes results from a survey of 33 school administrators from the Antioch Unified School District. In her survey, administrators generally indicated that RJ was either extremely effective (19%) or effective (50%) at address- ing student behavioral issues, with some (28%) saying it was neither effective nor ineffective and very few saying it was ineffective (3%) or very ineffective (0%). Administrators were far less sanguine about the effectiveness of detention, suspen- sions, or expulsions, with between 19 and 43% indicating these methods were ineffective or extremely ineffective. Nabors (2017) reports pre-post shifts in misbehavior via ad- ministrative data from two schools implementing RJ. She re- ports a 61% reduction in assaults on students, a 76% reduction in assaults on teachers, and a 42% reduction in disorderly conduct after West Philadelphia High School employed RJ; and a 70% reduction in incidents of disrespect to teachers and a 71% reduction in incidents of classroom disruption at Springfield Township High School after implementation. Carroll (2017) reports that, in three high schools in Merced, CA, all categories of suspensions dropped markedly after the implementation of facilitated restorative professional learning group (PLG) training. Total full-day suspension rates dropped in half (a statistically significant drop relative to trends prior to implementation), and in-school full-day equivalent suspen- sion dropped by 80% (also statistically significant relative to prior trends). Henson-Nash (2015) reports similar results from analyzing disciplinary infraction rates in a public K-8 school in Illinois from the 2006/2007 year (under zero tolerance) in comparison with rates in the 2008/2009 year (under RJ, and

after a one-year transition period). Henson-Nash reports that overall infractions during the RJ period were 83% lower, with particularly pronounced reductions for physical aggression (84% reduction), disrespect (85% reduction), and possession of a weapon or look-alike (100% reduction). Notably, the author’s decision to compare two time periods separated by a gap was a unique methodological choice that may have biased their estimates, and their results may say more about the shift away from zero tolerance than the shift to RJ. In a cleaner pre-post comparison, Katic (2017) reviewed disciplin- ary data at a middle school in San Bernardino, CA, during two time frames: a three-year period prior to implementation of RJ and a two-year period after implementation. A chi-squared analysis revealed that the suspension rate for the post- implementation period was statistically significantly lower than the rate during the pre-implementation period (p < .001). The annual per-pupil suspension rate dropped by 40% from pre- to post-implementation.

Armour (2014), Barkley (2018), and DeAntonio (2015) report exceptions to the otherwise consistent finding that be- havioral problems drop after RJ implementation. Armour (2014) found that offense frequencies grew steadily over the course of RJ implementation in a San Antonio school. Barkley (2018) reports that office discipline referrals per student in- creased over a five-year span following RJ implementation in one middle school in Michigan. Importantly, Barkley notes that although staff at the school in the first two years received RJ training, staff received “little to no training” in subsequent years, and only 33% of the staff who were at the school in year one remained in year five, suggesting fidelity of implementa- tion issues. The school also experienced substantial changes in administrative leadership over the five-year period.

DeAntonio (2015) used a matched-pair design to attempt to ascertain whether training in restorative practices caused re- ductions in student misbehavior. Covariate matching is a com- mon tactic for attempting to teasing out causal effects from observational data. Data permitting, by matching similar schools, researchers can create treatment and control groups, which, in aggregate, are practically identical on covariates, and thus only differ in terms of treatment status. This theoret- ically allows researchers to more squarely attribute any differ- ences in mean outcome rates to treatment rather than to a measured confounder. However, all matching is only as good as the covariates used to create matches, and if there are any lingering confounders, which are not included in the model, matching can produce a biased estimate effect, the causal ef- fect. Critically, DeAntonio utilized a matching formula where- by matches were created via a composite score with five points assigned to average student income, three points to total enrollment, and one point to urbanicity. Thus, DeAntonio’s approach is subject to two major internal validity concerns: a failure to account for all relevant confounders and the utiliza- tion of a matching formula, which might create imbalance on

300 Contemp School Psychol (2020) 24:295–308

covariates, which could further bias the estimate. That said, DeAntonio reviewed data from the 2013/2014 school year from public schools in Pennsylvania and focused on 19 schools that had received restorative practice (RP) training from the International Institute of Restorative Practices prior to 2013 and compared those schools to 19 schools that had not received RP training. The resulting 19 matched pairs were then compared based on a “behavior triad” that measured the sum of incidents of fighting, incidents of disorderly con- duct, and truancy rate divided by each school’s total enroll- ment. Based on matched-pair t tests, the report notes that there was “no statistically significant difference in the frequency of behavior triad incidents between schools not utilizing RP and schools that do use RP” (p. iii).

Impacts on Bullying

Although recent data have shown decreases in the prevalence of bullying (Snyder et al. 2018), it is still a common problem affecting students. For example, some research has indicated that 30 to 45% of youth experience bullying in their peer group, either as a victim, bully, or both, and that most of this bullying occurs in schools (Kasen et al. 2004; Dinkes et al. 2009; Nansel et al. 2001). Bullying affects the perpetrator and victim, as well as the overall school climate, leading to stu- dents feeling unsafe and unsupported, which can negatively impact student learning (e.g., Limber and Nation 1998). Since RJ focuses on repairing relationships and changing the com- munity, some have suggested that it is a more viable alterna- tive to traditional peer-mediation strategies in dealing with bullying (e.g., Morrison 2006; Christensen 2009; Duncan 2011). Drawing on limited available evidence regarding RJ and bullying prevention, Molnar-Main (2014) concludes that RJ practices that incorporate meetings, or conferences, be- tween the bully and his or her victim, may help reduce bully- ing in schools. In some cases, however, victims may not be comfortable facing the bully due to fear regarding potential consequences (Amstutz and Mullet 2005). To address these and other concerns, Molnar-Main (2014) provides a number of recommendations for how to integrate RJ and bullying prevention, such as focusing on the emotional safety of the victim and ensuring that trained adult facilitators lead the conferences.

While little research has directly addressed whether RJ im- plementation reduces bullying, available evidence is mixed. Acosta et al. (2019) conducted an RCT in which seven Maine middle schools were randomly selected to receive RJ training and six were randomly selected not to. As with the research conducted by Augustine et al. (2018), Acosta et al. report evidence of compliance issues—based on student surveys, treatment schools used RJ less, and control schools used RJ more, than would be expected given the assignment scheme. Histograms depicting percentages of students receiving

various levels of RJ exposure looked remarkably similar re- gardless of student treatment assignment with “treatment” stu- dents being only a bit more likely to have a high degree of RJ exposure. Overall, assignment to treatment did not predict less bullying, suggesting that RJ training did not have a statistical- ly significant negative impact on bullying victimization. However, students who reported that their teachers used more restorative practices experienced statistically significantly less physical bullying (p < .01) and statistically significantly less cyber bullying (p < .001). More restorative practices also predicted less emotional bullying, but the effect was not sta- tistically significant. More restorative practices predicted sta- tistically significantly higher levels of peer attachment and positive peer relationships, school connectedness, and empa- thy (p < .001 in all cases), suggesting various mechanisms that might be responsible for lower levels of bullying.

Other research is mixed. Henson-Nash (2015) reviews sur- veys from students in an Illinois school district and finds that, while many types of bullying behavior diminished after RJ implementation, and female students saw a statistically signif- icant drop in bullying victimization, overall drops in bullying and victimization were not statistically significant. Augustine et al.’s (2018) randomized controlled trial of RJ in 44 Pittsburgh K-12 schools found that RJ implementation caused a small reduction in bullying (p < .1). Armour (2014) mean- while finds that a San Antonio middle school that implement- ed RJ saw an increase in bullying after implementation.

Impacts on Racial Disparities in Discipline

Research has indicated that punitive sanctions may drive students—particularly minority and poor students—out of school altogether, resulting in a “school-to-prison” pipeline (Losen 2015). As previously mentioned, research has also indicated a disparity in the rates of exclusionary punishment for racial minorities and students with disabilities in compar- ison with other students (Petrosino et al. 2017). One possible explanation is that, as Kupchik and Ward (2014) found in their nationally representative survey of schools, schools with larg- er shares of minority and low-income students are statistically significantly more likely to engage in more surveillance and law enforcement activities (e.g., armed police or security forces patrolling the grounds, metal detectors, security cam- eras, locker searches). Minority and low-income students heightened presence in schools that give more and more se- vere punishment may partially explain demographic disci- pline disparities. Another often proffered explanation is that more minority students are being disciplined because they are engaging in more serious behavior that warrants stricter pun- ishment. However, there is also considerable discretion among administrators as to what is punishable under zero- tolerance policies (Payne and Welch 2010). For example, mi- nority students may not be committing more serious offenses

301Contemp School Psychol (2020) 24:295–308

but may be more likely to receive exclusionary discipline for vaguely defined offenses such as “disrespect,” “willful defi- ance,” and “disruption.” Thus, a final explanation is that often unintentional staff biases may lead to disproportionate disci- pline for certain groups of students (Skiba et al. 2002). Notably, there is evidence to suggest that such biases may impact how teachers view student actions and whether teachers notice misconduct by students at all. For example, Gilliam et al. (2016) report that preschool teachers who were asked to monitor classroom footage for “problem behaviors” tended to more carefully track Black boys in a classroom than students of any other demographic profile. Nonetheless, we are unaware of any research demonstrating a direct causal link between teacher biases and discipline disparities. Regardless of the causes of discipline disparities, RJ has been introduced as one method for addressing this disproportionality (Gregory et al. 2016). Proponents have argued that RJ can facilitate positive student-teacher relations for students of all back- grounds by guiding both teachers and students through a pro- cess of better understanding and respecting one another.

Quantitative research regarding the effect of RJ at reducing discipline gaps is mixed but largely supporting the effective- ness of RJ in this domain. A randomized controlled trial (RCT) comparing outcome measures in 22 RJ schools to those in 22 control schools indicates that RJ implementation led to a reduction in the racial discipline gap between Black and White students (Augustine et al. 2018). Similarly, Gregory et al. (2016) report that after two large, diverse high schools in a small, East Coast city implemented the SaferSanerSchools program from the International Institute of Restorative Practices, teachers who implemented RJ frequently had better relationships with their students; that students felt respected by their teachers, and teachers generally issued fewer referrals; and that frequent use of RJ may have led to reductions in the racial discipline gap. Specifically, teachers who had been rat- ed by students as “highly affective” (or frequently using emo- tional communication) exhibited less of a racial discipline gap in their referrals than teachers rated lower on the “affective” scale.

A 2018 analysis of Los Angeles Unified School District’s discipline records following the implementation of RJ in the 2014/2015 school year demonstrates that suspension rates for misconduct dropped for all measured categories of students (Black, Latino, Asian, and White students; students with dis- abilities; English learner students; and students eligible for free or reduced-price lunch). The analysis also indicates that even though discipline gaps related to race and disability sta- tus persisted, those gaps had narrowed considerably (Hashim et al. 2018).

González recently reported data from Denver Public Schools demonstrating that after the schools implemented RJ, the suspension rate dropped for Black, Latino, and White students, and the discipline gaps narrowed between

Black and White students and between Latino and White stu- dents (González 2015). Based on a more recent analysis of Denver Public Schools data, Gregory et al. (2018) report that suspension rates dropped for all racial categories and that the Black-White discipline gap narrowed nearly in half, from 9 to 5%, following the introduction of RJ throughout the district. In an earlier study, Jain et al. (2014) examined 30 schools in Oakland (CA) based on the schools’ practices as of July 2014 that were categorized as “control” (no RJ), “emerging” (just starting to use RJ in a limited scope), “developing” (using RJ throughout the school but not necessarily in all functions), and “thriving” (using RJ in all aspects of school life—no schools received this designation). Looking over the period from 2011 to 2014, researchers in this study found that “developing” schools closed the Black-White discipline gap by a few per- centage points (from 12.6 to 9.2%) while the discipline gap actually grew in both “emerging” and “control” schools. They also surveyed adults connected to schools implementing RJ in Oakland and found that 11 out of 12 of the surveyed principals and assistant principals believed that RJ had helped reduce disciplinary referrals for Black and Latino boys. Although many respondents in other groups were unsure of this causal- ity, within all categories of adults (including teacher, RJ coor- dinator, staff, parent), larger percentages believed that RJ re- duced these referrals than believed that it did not. Armour (2014) found that both the Black-White and Latino-White discipline gaps narrowed after RJ implementation in a San Antonio middle school. In a review of administrative data from a large urban district, Anyon et al. (2016) found that, following an RJ intervention, discipline rates abated overall, but racial discipline gaps persisted.

Impacts on Attendance and Absenteeism

Chronic school absence and truancy have been linked to a wide range of negative childhood and adult outcomes, includ- ing low academic achievement, high dropout rates, difficulties in obtaining employment, poor health, increased chances of living in poverty, and increased risk of juvenile deviance and violent behavior (McCluskey et al. 2004; Baker et al. 2001). As mentioned previously, punitive and exclusionary ap- proaches to address absence and truancy may backfire, as such approaches may prevent youth from reengaging with school and, in turn, may increase their likelihood of engagement with the justice system. Accordingly, proponents offer RJ as an approach to addressing truancy and chronic absenteeism among students.

The research studies identified in the literature relevant to attendance vary widely in how outcomes are reported. Nonetheless, across the studies, school attendance tended to improve after RJ implementation. Baker (2009), for example, reports that students who participated in at least three RJ in- terventions over the course of a school year experienced a

302 Contemp School Psychol (2020) 24:295–308

50% reduction in absenteeism during the first year of imple- mentation and a decrease in tardiness of about 64%. McMorris et al. (2013), who studied a Family Group Conferencing pro- gram for expelled students, report that participants’ attendance increased from pre- to post-implementation periods. Reviewing data from a San Antonio middle school implementing RJ in 6th and 7th, but not 8th, grades, Armour (2014) found that students in the implementing grades had markedly fewer tardies than the 8th grade students, and that the cohort of students who experienced two years of RJ (7th graders) received less than half as many tardies as the cohort of students who experienced zero years of RJ (8th graders). A study (Jain et al. 2014) in Oakland (CA) reports that middle schools implementing RJ experienced a drop in chronic ab- senteeism by 24% while schools not implementing the pro- gram experienced an increase of 62.3% during the same peri- od. But not all schools experienced such declines. Riestenberg (2003) reports that one school that implemented RJ reported a 2% increase in absenteeism in the follow-up year. Augustine et al. (2018) did not find a statistically significant link between RJ implementation and absenteeism in their two-year RCT of 44 K-12 schools in Pittsburgh, PA.

Impacts on School Climate and Safety

Some researchers argue that educators and administrators who create a safe, supportive, and nurturing school climate help promote the social-emotional growth and positive develop- ment of students (Voight et al. 2013). One objective of ad- dressing school climate is to foster healthy, resilient students who are ready for college and careers out of school. RJ is one tool among many that educators may use to create and support a positive school climate (e.g., Health and Human Development Program 2012).

Although the evidence is limited, there are findings to sug- gest that RJ improves school climate. Based on their 44- school RCT, Augustine et al. (2018) report that RJ caused a statistically significant (p < .05) increase in teachers’ percep- tions of school climate. The authors note that this impact was driven by large and statistically significant (p < .05) positive impacts on teachers’ views about school safety and whether they understood school policies regarding student conduct. They also note statistically significant improvements in teachers’ perceptions about working conditions being condu- cive to teaching and learning, opportunities for leadership, and school leadership. Acosta et al.’s (2019) RCT painted a more mixed picture. RJ implementation did not cause statistically significant improvements in school climate, but students who experienced more RJ practices by teachers experienced statis- tically significantly higher levels of school connectedness, positive peer relations, and peer attachment (p < .001 in all cases).

In her dissertation, Featherston (2014) reviews results from an RCT of 48 Black adolescent girls attending a Mid-Atlantic high school that participated in Real Talk 4 Girls, a three-week social problem-solving program. The program uses a “restor- ative circle” format to teach cognitive strategies via lessons to help girls define social aggression, and behavioral strategies via role-playing and practicing new behaviors. Girls were guided to recognize social problems, brainstorm and select solutions, enact behaviors, and evaluate results. As noted pre- viously, RCTs in education contexts must overcome two ma- jor threats to internal validity: compliance (treatment and con- trol units follow through with their assignments) and timing (RCTs run long enough to observe anticipated results). Featherston’s intervention involved six one-hour sessions spread over four weeks and guided by a specific pedagogy to help participants develop empathy. Each of the 24 girls assigned to treatment followed through. Moreover, given the specific and intensive nature of the student-level intervention, it is hard to imagine that girls assigned to control were able to seek out and receive a similar intervention elsewhere. One lingering potential compliance issue is that girls assigned to treatment could have shared lessons with girls assigned to control, rendering them poor comparators. If this were to oc- cur, it would mean that Featherston’s estimate of the causal effect should be viewed as a conservative one. The short du- ration of Featherston’s intervention and data collection may also result in a failure to understand effects that take more time to materialize.

These caveats notwithstanding, based on MANCOVA of post-experiment student surveys, Featherston reports that the 24 girls who participated in the program exhibited statistically significant declines in social aggression (p < .001) and statis- tically significant increases in social problem-solving (p < .001) and pro-social behavior (p < .05), relative to the 24 girls in the control condition.

Similarly, in the aforementioned study of Family Group Conferencing in Minnesota, McMorris et al. (2013) report increased school connectedness and improved problem- solving among students in a six-week follow-up. The Lansing School District (2008) reported similar findings from a one-month post-survey of 289 RJ participants, finding that 91% of students said they “learned new [conflict resolution] skills in the RJ process” (p. 3) and 89% of their parents indi- cated that their children “learned new ways to resolve [dis- putes]” (p. 3). Students also generally indicated that they re- solved their initial disputes through the RJ process (92%) and had used these skills to resolve future disputes (90%), and their parents generally agreed that RJ helped their children resolve the initial dispute (85%). Based on a survey in schools implementing RJ in Oakland (CA), Jain et al. (2014) report that 69% of staff believed that RJ had improved school cli- mate and 64% believed that it helped build caring relation- ships between teachers and students. Staff were about four

303Contemp School Psychol (2020) 24:295–308

times more likely to hold each of these positive opinions than to believe RJ had had a negative impact on climate or rela- tionships. However, parents’ opinions were not as strongly positive. Whereas 100% of principals believed that RJ im- proved school climate, only 40% of parents agreed, and whereas 92% of principals believed that RJ improved teacher-student relationships, only 28% of parents did.

In more recent research, an elementary school saw a 55% decrease in physical aggression after implementing RJ, and 97.7% of students reported feeling safe in school after imple- mentation (Goldys 2016). Gregory et al.’s (2016) research in two large, diverse, East Coast high schools similarly found that students’ perceptions of their teachers’ levels of RJ im- plementation were predictive of students’ depictions of their relationships with their teachers (whether the teachers respected them), even after controlling for student race and teachers’ depictions of students’ levels of cooperativeness. Focusing on three diverse, rural, West Coast schools, Terrill (2018) reports that teachers felt that implementing the Discipline that Restores program resulted in greater respect by students for other students. And Jain et al.’s (2014) survey found that 67% of staff in schools implementing RJ indicated that RJ helped students improve their social and emotional skills.

Henson-Nash (2015), in her dissertation, compares disci- plinary infraction rates in a public K-8 school in Illinois from the 2006/2007 year (under zero tolerance) to rates in the 2008/ 2009 year (under RJ, and after a one-year transition period). She reports that infractions related to physical aggression went down by 84% (from 143 to 23 infractions) and infractions for possession of a weapon or look-alike went down by 100%, from 13 infractions to none.

Impacts on Academic Outcomes

There is limited and mixed evidence that RJ can improve academic achievement and progress. McMorris et al. (2013) note that for students in their sample who participated in Family Group Conferencing and remained enrolled in school the following academic year, participation was associated with a slight increase in the students’ grade point averages. Although there was a sizeable drop in the number of students on track to graduate in the year of their participation in RJ, this drop may have been due to the poor attendance prior to the program, and a majority of these students did get back on track in the following year. In a review of data from a San Antonio middle school, Armour (2014) found that 6th grade students exposed to RJ for a year saw 11% improvements on their statewide reading passage rates and a 13% improvement in math. These improvements were driven in part by growth by Black students (8% increase), Hispanic students (13%), and economically disadvantaged students (13%). Most notably, passage rates among special education students increased

markedly in both reading (42%) and math (50%). Jain et al. (2014) report that schools in Oakland (CA) that were implementing RJ saw reading levels increase by 128% over three years while non-RJ schools saw an increase of only 11%; four-year graduation rates increased by 60% in RJ schools, compared to 7% for schools not implementing RJ; and high school dropout rates decreased by 56% in RJ high schools compared to 17% for non-RJ high schools. Kerstetter (2016) compares outcomes at a restorative justice charter ele- mentary school to a “comparable” no-excuses charter and finds that the RJ charter school outperformed the comparison charter in the percent of third grade students scoring proficient on the state’s standard. Specifically, in the study year (2012– 2013), 60% of the RJ charter’s students were proficient while 36% of the comparison’s charter were, as were 44% students across the district. Notably, though, in the following year, 47% of students in the RJ charter were proficient, which com- pared to 41% of comparison charter and 44% of district stu- dents. The authors take this as evidence that RJ can coincide with high academic performance rather than it can improve academic performance.

Elsewhere, the results for academic outcomes are more mixed. For example, based on an RCT of RJ in 44 schools in Pittsburgh, PA, Augustine et al. (2018) report that RJ did not have a positive impact on math and reading scores, but rather led to significant (p < .05) reductions in elementary and middle school math performance and significant (p < .01) re- ductions in elementary and middle school academic performance among Black students. Similarly, Norris (2009) reports no significant change in grade point average for RJ participants (compared with non-participants). Lewis (2009) suggests that there was improvement in student test scores in one Pennsylvania school but provides no data to support this finding. Based on reviewing the student records of 80 students in a diverse, rural California high school, Terrill (2018) reports that while grade point averages of students overall fell after RJ implementation, grade point averages increased among stu- dents who had received office referrals and therefore encoun- tered the Discipline that Restores program.

Access to Restorative Justice

Given the aforementioned research suggesting that RJ might yield considerable improvements on certain outcomes, some researchers have been concerned with whether students of all backgrounds have equal access to RJ programs. To date, anal- yses on this question have yielded inconsistent answers re- garding the level of access to RJ by groups. Payne and Welch (2015) reviewed surveys of students, teachers, and principals from across the country from 1997 and 1998 to discern where restorative practices were being utilized. They report that schools with higher percentages of Black students were statistically significantly less likely to use each of four

304 Contemp School Psychol (2020) 24:295–308

restorative practices—student conferences, peer mediation, restitution, and community service—even after controlling for a range of student-level and school-level characteristics (such as percentage of students eligible for free or reduced- price lunch, percentage of students who are Hispanic and per- centage who are male; and the extent to which the school is in a disadvantaged or urban community). Based on reviewing the same data, Payne and Welch (2018) report that students who received free and reduced-price lunches were statistically significantly (p < .05) less likely to be exposed to student con- ferences. Thus, research by Payne and Welch (2015, 2018) suggests that, at least at the time when the surveys were done, RJ access was substantially constrained for Black students and somewhat constrained for low-income students. Anyon et al. (2016) approached the question of access using more recent data from Denver Public Schools. Based on multi-level re- gression modeling on the records of 9921 students with disci- plinary records in 2012/2013, they report that student groups that were overrepresented in school discipline (with the ex- ception of English learner students) generally had comparable or higher-than-average access to restorative interventions compared with student groups that were not overrepresented in discipline. In particular, Latino and Black students had higher likelihoods than White students of being exposed to restorative interventions. The distinction in findings is likely owing to a distinction in the contexts reviewed and may sug- gest that members of student groups that are overrepresented in discipline generally do not have equivalent access to RJ nationally, but have substantial access in Colorado.

Limitations of the Literature Review

The evidence presented in this literature review is limited ini- tially by what we found in searches conducted through January, 2020, and subsequently by what we chose to report from those searches. Our focus on US K-12 schools meant we did not include write-ups on the many thorough studies of RJ in schools that have been conducted in other countries, nor on the budding literature on RJ’s effectiveness in US higher ed- ucation institutions. And while we used comprehensive methods to review available literature, we will have missed any studies that both do not appear in the databases we sur- veyed and were not mentioned by our interviewed experts.

A review of evidence is influenced by the quality of the studies that comprise the “sample.” For each of the outcomes mentioned in our review, there is at least some evidence that suggests a beneficial impact of RJ in schools. However, set- ting aside that results are often mixed, there are also many limitations within many of the more sanguine studies. First, there are far too few studies in each category to have confi- dence in the stability of findings. Secondly, although not re- ported in all cases, many of the studies had limited small

sample sizes. To demonstrate the statistical significance of a result, researchers must first ensure sufficient statistical “pow- er,” which is to say they must have a sufficiently large sample. For studies of RJ that focus on individual- or school-level effects, reaching an adequate sample size can be a challenge. Next, and critically, as noted above, the vast majority of these studies are best understood as correlational in nature. The most common evaluation design reported in the literature is based on pre- and post-tests. By nature, such pre/post designs only study those individuals exposed to the program (i.e., a single-group design) with no counterfactual (control or com- parison condition). As noted by Weisburd et al. (2014), such designs are very low in internal validity and thus are best understood as correlational in nature. Even where researchers utilized more rigorous methods, they often reported internal validity concerns such as challenges with compliance among treated and control cases, and intervention duration issues.

Conclusion

In general, the research evidence to support RJ in schools is still in a nascent state. Despite the exponential growth of RJ in US schools, and some evidence of its effectiveness abroad (see Sherman and Strang 2007), the evidence in the USA to date is limited, and nearly all of the research that has been published lacks the internal validity necessary to exclusively attribute outcomes to RJ. However, the prelim- inary evidence does suggest that RJ can cause improve- ments in discipline, discipline disparities, misbehavior, and school climate. While mixed, available evidence fa- vors the effectiveness of RJ at reducing bullying and stu- dent absenteeism. Finally, evidence on the impact of RJ on academic performance is thoroughly mixed, with a number of correlational studies suggesting RJ can yield impressive academic performance gains, while a large-scale school- level RCT and other correlational studies suggest RJ may harm academic performance.

Given the state of the research field, future research might accelerate the growth of knowledge about the effectiveness of RJ in schools by addressing these research challenges:

& establishing a clear, concise, and largely acceptable defi- nition of RJ;

& examining the factors associated with a school’s readiness to implement RJ and how those factors interact with the effectiveness of an school’s RJ programming;

& identifying factors that determine high levels of imple- mentation fidelity, including determining what kinds of training and professional development for school leaders enhance their ability to value and implement an RJ program;

305Contemp School Psychol (2020) 24:295–308

& examining RJ effectiveness via rigorous outcome-based research—like that of Featherston (2014), Augustine et al. (2018), and Acosta et al. (2019);

& gathering data in the places in which successful and sus- tainable RJ programs have been implemented, to uncover the conditions that lead to replicable examples; and

& examining the integration of RJ with other multi-tiered models such as Positive Behavioral Interventions and Supports (PBIS) and Response to Intervention (RTI).

Funding Information This study was originally funded by the Robert Wood Johnson Foundation (Award Number 71141).

Compliance with Ethical Standards

Conflict of Interest The authors declare that they have no conflict of interest.

Ethical Approval This article does not contain any studies with human participants or animals performed by any of the authors.

References

Acosta, J., Chinman, M., Ebener, P., Malone, P. S., Phillips, A., & Wilks, A. (2019). Evaluation of a whole-school change intervention: find- ings from a two-year cluster-randomized trial of the restorative prac- tices intervention. Journal of Youth and Adolescence, 48, 876–890. https://doi.org/10.1007/s10964-019-01013-2.

Amstutz, L. S., & Mullet, J. H. (2005). The little book of restorative discipline for schools: teaching responsibility, creating caring climates. Intercourse: Good Books.

Anyon, Y., Gregory, A., Stone, S., Farrar, J., Jenson, J. M., McQueen, J., Downing, B., Greer, E., & Simmons, J. (2016). Restorative inter- ventions and school discipline sanctions in a large urban school district. American Educational Research Journal, 53(6), 1663– 1697. https://doi.org/10.3102/0002831216675719.

Armour, M. (2014). Ed White Middle School restorative discipline eval- uation: implementation and impact, 2012/2013 sixth & seventh grade [PDF file]. The Institute for Restorative Justice and Restorative Dialogue. Retrieved from http://sites.utexas.edu/irjrd/ files/2016/01/Year-2-Final-EW-Report.pdf

Ashley, J., & Burke, K. (2009). Implementing restorative justice: a guide for schools [PDF file]. Illinois Criminal Justice Information Authority. Retrieved from http://www.icjia.state.il.us/assets/pdf/ BARJ/SCHOOL%20BARJ%20GUIDEBOOOK.pdf

Augustine, C. H., Engberg, J., Grimm, G. E., Lee, E., Wang, E. L., Christianson, K., & Joseph, A. A. (2018). Can restorative practices improve school climate and curb suspensions? An evaluation of the impact of restorative practices in a mid-sized urban school district [PDF file]. RAND Corporation. Retrieved from https://www.rand. org/pubs/research_reports/RR2840.html

Baker, M. (2009). DPS Restorative Justice Project: year three. Denver: Denver Public Schools.

Baker, M. L., Sigmon, J. N., & Nugent, M. E. (2001). Truancy reduction: keeping students in school. Juvenile Justice Bulletin. Retrieved from http://www.childtrends.org/?indicators=student-absenteeism#_ edn4

Barkley, S. (2018). A 5-year examination of data for non-minority and minority students from 2011–2016 (publication no. 10746541)

[doctoral dissertation, Northcentral University, San Diego, CA]. ProQuest Dissertations Publishing.

Bazemore, G., & Schiff, M. (2005). Juvenile justice reform and restor- ative justice: building theory and policy from practice. Willan Publishing.

Bazemore, G., & Schiff, M. (2009). Addressing the school-to-jail pipe- line: restorative justice and theory for practice in real alternatives to zero tolerance [Paper presentation]. American Society of Criminology: Criminology and Criminal Justice Policy, Philadelphia Marriott Downtown.

Beckman, K., McMorris, B., & Gower, A. (2012). Restorative interventions implementation toolkit [PDF file]. Minnesota Department of Education. Retrieved from https://education.mn.gov/mdeprod/idcplg?IdcService= GET_FILE&dDocName=048363&RevisionSelectionMethod= latestReleased&Rendition=primary.

Braithwaite, J. (2004). Restorative justice: theories and worries [Paper presentation]. United Nations Asia and Far East Institute For the Prevention of Crime and the Treatment of Offenders: 123rd International Senior Seminar, Fuchu, Tokyo. Retreived from https://www.unafei.or.jp/publications/pdf/RS_No63/No63_10VE_ Braithwaite2.pdf.

Brown, M. (2017). Being heard: how a listening culture supports the implementation of schoolwide restorative practices. Restorative Justice: An International Journal, 5(1), 53–69. https://doi.org/10. 1080/20504721.2017.1294792.

Carroll, P. G. (2017). Evaluating attempts at the implementation of re- storative justice in three alternative education high schools (Doctoral dissertation). Retrieved from https://escholarship.org/uc/ item/2t95r24f

Cavanagh, T., Vigil, P., & Garcia, E. (2014). A story legitimating the voices of Latino/Hispanic students and their parents: creating a re- storative justice response to wrongdoing and conflict in schools. Equity & Excellence in Education, 47(4), 565–579. https://doi.org/ 10.1080/10665684.2014.958966.

Christensen, L. (2009). Sticks, stones, and schoolyard bullies: restorative justice, mediation and a new approach to conflict resolution in our schools. Nevada Law Journal, 9(3), 545–579.

Davis, F. (2014). Discipline with dignity: Oakland classrooms try healing instead of punishment. Reclaiming Children and Youth, 23(1), 38– 41.

DeAntonio, M. G. (2015). A comparative study of restorative practices in public schools (publication no. 3734773) [doctoral dissertation, Alvernia University, Reading, Pennsylvania]. ProQuest Dissertations Publishing.

Dinkes, R., Kemp, J., & Baum, K. (2009). Indicators of school crime and safety: 2008 (NCES 2009-022/NCJ 226343). Washington, D.C: National Center for Education Statistics.

Duncan, S. (2011). Restorative justice and bullying: a missing solution in the anti-bullying laws. New England Journal on Criminal and Civil Confinement, 37, 701.

Evans, K., & Lester, J. (2013). Restorative justice in education: what we know so far. Middle School Journal, 44(5), 57–63. https://doi.org/ 10.1080/00940771.2013.11461873.

Featherston, T. (2014). An experimental study on the effectiveness of a restorative justice intervention on the social aggression, social problem solving skills, and prosocial behaviors of African American adolescent girls (publication no. 3619936) [doctoral dis- sertation, Capella University]. ProQuest Dissertations Publishing.

Fowler, B., Rainbolt, S., & Mansfield, K. (2016). Re-envisioning disci- pline in complex contexts: an appreciative inquiry of one district’s implementation of restorative practices. Detroit, MI: Paper present- ed at University Council for Educational Administration Annual Conference.

Gilliam, W. S., Maupin, A. N., Reyes, C. R., Accavitti, M., & Shic, F. (2016). Do early educators’ implicit biases regarding sex and race relate to behavior expectations and recommendations of preschool

306 Contemp School Psychol (2020) 24:295–308

expulsions and suspensions? [PDF file]. Yale University Child Study Center. Retrieved from https://medicine.yale.edu/childstudy/ zigler/publications/Preschool%20Implicit%20Bias%20Policy% 20Brief_final_9_26_276766_5379_v1.pdf

Goldys, P. (2016). Restorative practices: from candy and punishment to celebration and problem-solving circles. Journal of Character Education, 12(1), 75–80.

González, T. (2012). Keeping kids in schools: restorative justice, punitive discipline, and the school to prison pipeline. Journal of Law and Education, 41(2), 281–335.

González, T. (2015). Socializing schools: addressing racial disparities in discipline through restorative justice. In D. Losen (Ed.), Closing the school discipline gap: equitable remedies for excessive exclusion (pp. 151–165). New York, NY: Teachers College Press.

Government Accountability Office. (2018). K-12 education: discipline disparities for Black students, boys, and students with disabilities (GAO-18-258). Retrieved from https://www.gao.gov/products/ GAO-18-258

Greer, L. S. (2018). Firm but fair: authoritative school climate as a predictor of restorative justice readiness (Doctoral dissertation). Retrieved from https://scholarworks.lib.csusb.edu/etd/644/

Gregory, A., Clawson, K., Davis, A., & Gerewitz, J. (2016). The promise of restorative practices to transform teacher-student relationships and achieve equity in school discipline. Journal of Educational and Psychological Consultation, 26(4), 325–353. https://doi.org/ 10.1080/10474412.2014.929950.

Gregory, A., Huang, F. L., Anyon, Y., Greer, E., & Downing, B. (2018). An examination of restorative interventions and racial equity in out- of-school suspensions. School Psychology Review, 47(2), 167–182. https://doi.org/10.17105/SPR-2017-0073.V47-2.

Guckenburg, S., Hurley, N., Persson, H., Fronius, T., & Petrosino, A. (2015). Restorative justice in U.S. schools: summary findings from interviews with experts [PDF file]. WestEd. Retrieved from https://jprc.wested.org/ wp-content/uploads/2015/11/1447101213resourcerestorativejustice inusschoolssummaryfindingsfrominterviewswithexperts.pdf

Hantzopoulos, M. (2013). The fairness committee: restorative justice in a small urban public high school. Prevention Researcher, 20(1), 7–10 Retrieved from https://www.fsusd.org/cms/lib/CA01001943/ Centricity/Domain/1271/TPR20-1-Hantzopoulos%20.pdf.

Hashim, A., Strunk, K., & Dhaliwal, T. (2018). Justice for all? Suspension bans and restorative justice programs in the Los Angeles Unified School District. Peabody Journal of Education, 93(2), 174–189. https://doi.org/10.1080/0161956X.2018.1435040.

Health and Human Development Program (2012). Workbook for improv- ing school climate & closing the achievement gap, 2nd edition [PDF file]. California Department of Education. Retrieved from https:// www.wested.org/online_pubs/WB_1221_allv5.pdf

Henson-Nash, S. (2015). A study of bullying: a school perspective with a restorative discipline model approach (publication no. 3728995) [doctoral dissertation, St. Francis College of Education, New York, NY]. ProQuest Dissertations Publishing.

Jain, S., Bassey, H., Brown, M., & Kalra, P. (2014). Restorative justice in Oakland schools: implementation and impacts [PDF file]. Oakland Unified School District. Retrieved from https://www.ousd.org/cms/ lib/CA01001176/Centricity/Domain/134/OUSD-RJ%20Report% 20revised%20Final.pdf

Karp, D., & Breslin, B. (2001). Restorative justice in school communities. Y o u t h a n d S o c i e t y , 3 3 ( 2 ) , 2 4 9 – 2 7 2 1 0 . 1 1 7 7 % 2F0044118X01033002006.

Kasen, S., Berenson, K., Cohen, P., & Johnson, J. G. (2004). The effects of school climate on changes in aggressive behavior and other be- haviors related to bullying. In D. L. Espelage & S. M. Swearer (Eds.), Bullying in American schools: a social-ecological perspec- tive on prevention and intervention (pp. 187–210). Mahwah: Erlbaum.

Katic, B. (2017). Restorative justice practices in education: a quantitative analysis of suspension rates at the middle school level (Master’s thesis). Retrieved from https://scholarworks.lib.csusb.edu/cgi/ viewcontent.cgi?article=1567&context=etd

Kerstetter, K. (2016). A different kind of discipline: social reproduction and the transmission of non-cognitive skills at an urban charter school. Sociological Inquiry, 86(4), 512–539. https://doi.org/10. 1111/soin.12128.

Lansing School District (2008). Lansing School District Restorative Justice Annual Report [PDF file]. Retrieved from http://www. lansingschools.net/downloads/restorative_justice_files/lansing_ school_district_restorative_justice_annual_report_07-08.pdf

Lewis, S. (2009). Improving school climate: findings from schools implementing restorative practices [PDF file]. International Institute for Restorative Practices. Retrieved from https://www.iirp. edu/pdf/IIRP-Improving-School-Climate-2009.pdf

Limber, S., & Nation, M. (1998). Bullying among children and youth. Juvenile Justice Bulletin. Retrieved from http://www.ojjdp.gov/ jjbulletin/9804/bullying2.html

Losen, D. (Ed.). (2015). Closing the school discipline gap: equitable remedies for excessive exclusion. Teachers College Press.

McCain, M. (2015). The impact of discipline policies on students of color and the inequities of suspensions and expulsions (publication no. 3738532) [doctoral dissertation, California State University, East Bay, Hayward, CA]. ProQuest Dissertations Publishing.

McCluskey, C. P., Bynum, T. S., & Patchin, J. W. (2004). Reducing chronic absenteeism: an assessment of an early truancy initiative. Crime and Delinquency, 50(2), 214–234 10.1177% 2F0011128703258942.

McCold, P. (2002). Evaluation of a restorative milieu: CSF Buxmont School/day treatment programs 1999–2001, evaluation outcome technical report [PDF file]. International Institute of Restorative Practices. Retrieved from http://www.iirp.edu/pdf/erm.pdf

McCold, P. (2008). Evaluation of a restorative milieu: restorative prac- tices in context. Sociology of Crime, Law and Deviance, 11, 99–137.

McMorris, B. J., Beckman, K. J., Shea, G., Baumgartner, J., & Eggert, R. C. (2013). Applying restorative justice practices to Minneapolis Public Schools students recommended for possible expulsion [PDF file]. University of Minnesota. Retrieved from https://www. legalrightscenter.org/uploads/2/5/7/3/25735760/lrc_exec_summ- final.pdf

Mirsky, L. (2007). SaferSanerSchools: transforming school cultures with restorative practices. Reclaiming Children and Youth: The Journal of Strength-based Interventions, 16(2), 5–12.

Mirsky, L., & Wachtel, T. (2007). “The Worst School I’ve Ever Been To”: empirical evaluations of a restorative school and treatment milieu. Reclaiming Children and Youth: The Journal of Strength- based Interventions, 16(2), 13–16.

Molnar-Main, S. (2014). Integrating bullying prevention and restorative practice in schools: considerations for practitioners and policymakers [PDF file]. Highmark Foundation. Retrieved from http://www.safeschools.info/content/BPRPWhitePaper2014.pdf

Morris, E., & Perry, B. (2016). The punishment gap: school suspension and racial disparities in achievement. Social Problems, 63, 80–81 https://doi.org/10.1093/socpro/spv026.

Morrison, B. (2006). Schools and restorative justice. In G. Johnstone & D. Van Ness (Eds.), Restorative justice handbook (pp. 325–350). Cullompton, United Kingdom: Willan Publishing.

Morrison, B., & Vaandering, D. (2012). Restorative justice: pedagogy, praxis, and discipline. Journal of School Violence, 11(2), 138–155. https://doi.org/10.1080/15388220.2011.653322.

Nabors, L. (2017). The level of fidelity in restorative practices and its impact on school-level discipline data in middle Tennessee (Doctoral dissertation). Retrieved from https://www.cn.edu/ libraries/tiny_mce/tiny_mce/plugins/filemanager/files/ Dissertations/Dissertations2017/LaQuilla_Nabors.pdf

307Contemp School Psychol (2020) 24:295–308

Nansel, T. R., Overpeck, M., Pilla, R. S., Ruan, W. J., Simmons-Morton, B., & Scheidt, P. (2001). Bullying behaviors among U.S. youth: prevalence and association with psychosocial adjustment. Journal of the American Medical Association, 285(16), 2094–2100 Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/ PMC2435211/pdf/nihms53619.pdf.

Norris, A. (2009). Gender and race effects of a restorative justice inter- vention on school success. Philadelphia, PA: Paper presented at American Society of Criminology Annual Conference.

Payne, A., & Welch, K. (2010). Modeling the effects of racial threat on punitive and restorative school discipline practices. Criminology, 48(4), 1019–1062. https://doi.org/10.1111/j.1745-9125.2010. 00211.x.

Payne, A., & Welch, K. (2015). Restorative justice in schools: the influ- ence of race on restorative justice. Youth and Society, 47(4), 539– 564 10.1177%2F0044118X12473125.

Payne, A., & Welch, K. (2018). The effects of school conditions on the use of restorative justice in schools. Youth Violence and Juvenile Justice, 16(2), 224–240 10.1177%2F1541204016681414.

Petrosino, A., Guckenburg, S., & Fronius, T. (2012). “Policing schools” strategies: a review of the evaluation evidence. Journal of MultiDisciplinary Evaluation, 8(17), 80–101. Retrieved from https://journals.sfu.ca/jmde/index.php/jmde_1/article/view/337.

Petrosino, A., Fronius, T., Goold, C. C., Losen, D. J., & Turner, H. B. (2017). Analyzing student-level disciplinary data: a guide for dis- tricts (issues & answers report, REL 2017–no. 263) [PDF File]. U.S. Department of Education, Regional Educational Laboratory Northeast and Islands. Retrieved from https://ies.ed.gov/ncee/ edlabs/regions/northeast/pdf/REL_2017263.pdf

Riestenberg, N. (2003). Restorative schools grants final report, January 2002–June 2003: a summary of the grantees’ evaluation [PDF file]. Minnesota Department of Education. Retrieved from http:// crisisresponse.promoteprevent.org/webfm_send/1200

Sherman, L. W. (1993). Defiance, deterrence, and irrelevance: a theory of the criminal sanction. Journal of Research in Crime and Delinquency, 30, 445–473 10.1177%2F0022427893030004006.

Sherman, L. W., & Strang, H. (2007). Restorative justice: the evidence [PDF file]. Smith Institute. Retrieved from http://www.iirp.edu/pdf/ RJ_full_report.pdf

Skiba, R. J., Michael, R. S., Nardo, A. C., & Paterson, R. L. (2002). The color of discipline: sources of racial and gender disproportionality in school punishment. The Urban Review, 34(4), 317–342. https://doi. org/10.1023/A:1021320817372.

Snyder, T. D., Brey, C., & Dillow, S. (2018). Digest of education statis- tics 2016 [PDF file]. National Center for Education Statistics. Retrieved from https://nces.ed.gov/pubs2017/2017094.pdf

Stinchcomb, J. B., Bazemore, G., & Riestenberg, N. (2006). Beyond zero tolerance. Youth Violence and Juvenile Justice, 4(2), 123–147. https://doi.org/10.1177/2F1541204006286287.

Sumner, D., Silverman, C., & Frampton, M. (2010). School-based restor- ative justice as an alternative to zero-tolerance policies: lessons from West Oakland [PDF file]. University of California, Berkeley, School of Law. Retrieved from https://www.law.berkeley.edu/files/ thcsj/10-2010_School-based_Restorative_Justice_As_an_ Alternative_to_Zero-Tolerance_Policies.pdf

Terrill, S. (2018). Discipline that restores: an examination of restorative justice in the school setting. Olathe, KS: Paper presented at MidAmerica Nazarene University colloquium.

The Advancement Project. (2014). Restorative practices: fostering healthy relationships and promoting positive discipline in schools: a guide for educators [PDF file]. Retrieved from http:// schottfoundation.org/sites/default/files/restorative-practices-guide. pdf

Tyler, T. (2006). Restorative justice and procedural justice: dealing with rule breaking. Journal of Social Issues, 62(2), 307–326 https:// psycnet.apa.org/doi/10.1111/j.1540-4560.2006.00452.x.

Voight, A., Austin, G., & Hanson, T. (2013). A climate for academic success: how school climate distinguishes schools that are beating the achievement odds [PDF file]. WestEd. Retrieved from https:// www.wested.org/online_pubs/hd-13-10.pdf

Weisburd, D., Petrosino, A., & Fronius, T. (2014). Randomized experi- ments. In G. Bruinsma & D. Weisburd (Eds.), Encyclopedia of criminology and criminal justice (pp. 4283–4291). New York: Springer Press.

Zehr, H. (2002). The little book of restorative justice. Intercourse: Good Books.

Publisher’s Note Springer Nature remains neutral with regard to jurisdic- tional claims in published maps and institutional affiliations.

Sean Darling-Hammond is a PhD student at UC Berkeley’s Goldman School of Public Policy. He also holds a JD from UC Berkeley Law School and a BA in Sociology from Harvard University.

308 Contemp School Psychol (2020) 24:295–308

  • Effectiveness of Restorative Justice in US K-12 Schools: a Review of Quantitative Research
    • Abstract
    • Overview of Restorative Justice
      • Implementation Challenges and Pathways to Successful Implementation
    • Methodology of this Quantitative Research Review
      • Internal Validity of Available Quantitative Research
    • Effectiveness of Restorative Justice in Schools
      • Student Misbehavior and School Discipline
      • Impacts on Bullying
      • Impacts on Racial Disparities in Discipline
      • Impacts on Attendance and Absenteeism
      • Impacts on School Climate and Safety
      • Impacts on Academic Outcomes
      • Access to Restorative Justice
    • Limitations of the Literature Review
    • Conclusion
    • References