Abstract
Background/Context:
Restorative practices have gained traction across the United States, featuring in at least 21 state education codes and numerous district policies. The existing research on its effectiveness underscores the implementation process, particularly fidelity, as a key determinant in whether restorative practices will accomplish their goals of reducing punitive discipline and improving school community.
Purpose/Objective/Research Question/Focus of Study:
This study examines how restorative practices are implemented and scaled, drawing on data from four districts in a state that has mandated their use in discipline decisions. Using cognitive and organizational perspectives on policy implementation, we focus on understanding educator and student perceptions of implementation and identifying the organizational factors, contexts, and conditions that shape the use of the practices. By doing so, we provide novel evidence on how a statewide mandate translates into practice across schools and highlight the contextual dynamics shaping policy enactment.
Research Design:
We employ an explanatory mixed-methods approach, partnering with four districts in one county in Michigan. Data collection includes surveys from 493 educators and 4,805 students, alongside qualitative interviews with 47 staff members across nine schools during the 2022–23 school year. The explanatory design allows us to leverage interview data to explain patterns detected in the survey data.
Conclusions/Recommendations:
Survey results reveal that about half of classroom teachers, particularly at the elementary level, use proactive circles, while less than half of respondents use formal restorative conferences. Students generally report observing educators use restorative practices, with middle school students responding more positively on survey items related to implementation than high school students. However, students are less convinced about the community-building aspects of circles and whether the practices reduce conflict. Interviews indicate that implementation varies based on leadership expectations, alignment with existing practices, and site-specific discipline processes. We conclude with recommendations for policy and practice.
The negative associations of suspensions and other exclusionary discipline practices on academic, health, and criminal involvement outcomes are well-established (Ryan & Goodram, 2013; Skiba et al., 2014; Zinsser et al., 2022). These associations are particularly alarming given the disproportionate rate at which students of color and students with disabilities are subjected to suspensions (Carter et al., 2017; Losen, 2014). Given this reality, the federal government and states have taken action to reduce reliance on exclusionary discipline, with restorative practices emerging as a popular approach. Restorative practices have gained traction across the United States, featuring in 21 state education codes and numerous district policies. These codes often encourage or mandate their use, encourage professional development (PD), and allocate funding for implementation (Georgetown Law Center, 2020). Michigan is one such state that has codified the approach into law by requiring schools to consider restorative practices in disciplinary decisions (Mich. Comp. Laws § 380.1310c, 2016). Because of this state law and the growing evidence base (Darling-Hammond & Fronius, 2022), many Michigan districts are implementing these practices.
Restorative practices represent a paradigm shift from punitive approaches, emphasizing community building, repair, and nonpunitive responses to misbehavior (Fronius et al., 2016). Although specific practices may vary by context, they typically include proactive measures to foster positive school climates and strategies to resolve conflict communally and reintegrate disciplined students. Recent research, including multiple randomized controlled trials, suggests that these practices can effectively reduce reliance on exclusionary discipline, with studies showing 13–40% decreases in suspension or discipline incident rates (Acosta et al., 2019; Augustine et al., 2018; Grant et al., 2023; Gregory et al., 2022). However, the approach’s effects on racial disparities, school climate, achievement, and results by school level (e.g., middle and high schools relative to elementary schools) have been mixed, with some studies finding null effects or even negative effects in the case of achievement (see Darling-Hammond & Fronius, 2022 for review).
The implementation process likely accounts for the varying magnitude and direction of restorative practices’ effects across studies and outcomes. Research by Gregory et al. (2016) demonstrates that classrooms with higher implementation fidelity and those that use restorative practices frequently experience significantly greater benefits from these approaches. However, these approaches can face substantial implementation challenges because they necessitate fundamental cultural and behavioral shifts in educators’ community-building and disciplinary methods that might be incongruent with existing approaches. Studies examining early adopters of the approach have identified several critical factors that influence successful uptake and fidelity, including educator capacity building (Liberman & Katz, 2021), staff buy-in (Anfara et al., 2013; Liberman & Katz, 2021), and leadership (Anyon, 2016; Lustick, 2022). Although evidence for restorative practices’ enactment is growing, less is known about their implementation at scale across diverse educational contexts, particularly when mandated by state policies.
Cognitive and organizational perspectives are particularly valuable for understanding how and why implementation varies in response to state policy, because they reveal that policy interpretation is a dynamic, context-dependent process that shapes and constrains enactment (Coburn, 2005; Spillane et al., 2002; Weick, 1995). These perspectives conceptualize implementation as a behavioral change process where educators’ understanding of a policy or an approach influences action and is contingent on their previous beliefs, experiences, organizational norms and culture, and policy framing. Variation in enactment stems from differences in how implementing agents construct and make sense of the policy, which can be influenced by organizational settings and how leaders frame policy (Coburn, 2005; Spillane et al., 2002). By examining the factors that mediate cognition, researchers can better understand the complex ways educators interpret, adapt, and integrate new disciplinary approaches like restorative practices into their professional practice—particularly when these approaches contradict established paradigms like “zero tolerance.”
We use an explanatory mixed-methods design and cognitive and organizational perspectives of policy implementation to explore how restorative practices are being implemented across four districts in Michigan, with a particular focus on identifying how educators and students perceive the approach and the factors that constrain or enable their implementation. Drawing on data collected through a research–practice partnership with four districts in West Intermediate School District (ISD, 1 a pseudonym), we analyze survey data from 493 educators and 4,805 students, and qualitative interview data from 47 staff members across nine schools collected in the 2022–23 school year to answer the following research questions: (1) To what extent do educators report using restorative practices, and how does their use vary across districts and by school level? (2) What are students’ perceptions of restorative practices use, and how do these align with those of educators? (3) How do organizational factors, context, and conditions shape the interpretation and use of the approach?
Our study makes two key contributions. First, we provide novel evidence of how a statewide mandate affects local practices across districts and school levels. Second, we illuminate the critical role of cognitive and organizational perspectives in shaping policy enactment, with our results highlighting the role of congruence with existing practices, frameworks, and values (Coburn, 2004; Datnow & Castellano, 2000; Rigby, 2015). By analyzing variations in implementation, we offer insights into the complex process of restorative practices implementation.
Our survey results indicate roughly half of teachers use proactive circles in their classrooms, a core component for community building and preventing misbehavior. These practices, however, were more frequently used by elementary school teachers than middle and high school teachers. Among administrators and support staff (e.g., principals, assistant principals, counselors, social workers), slightly less than half of the respondents reported using formal restorative conferences at least once per week in the past four weeks. The vast majority of students report that their teachers help them resolve conflict in healthy ways, avoid labeling misbehaving students, and set positive norms for behavior, and middle school students are more likely to agree with these statements than high school students. However, students are less positive about whether circles build community and relationships and reduce conflict. Interviews with school staff suggest that uptake varies based on staff beliefs about the approach, as well as by district and school leaders’ expectations and commitment to implementation. Staff were more likely to describe how they use the approach in their classroom in schools and districts where leaders provided clear expectations and highlighted how the approach aligned with existing practices and school culture, including frameworks like Positive Behavioral Interventions and Support (PBIS) and other discipline processes. This alignment helped support educators’ adoption and use of specific community-building approaches like proactive circles in classrooms.
Background
Restorative practices are rooted in Indigenous onto-epistemological orientations, including those from Liberia, New Zealand, Australia, and Hawai’i, to assist in peacemaking, fortifying relationships, and goal setting (Kline, 2016; Schweigert, 2002; Zehr, 2002). In the United States, restorative approaches were initially used in criminal justice (i.e., restorative justice) in response to the proliferation of mass incarceration, either as an alternative to or in addition to criminal punishment (Marsh, 2019). In recent years, restorative practices have been integrated into K–12 schools through a combination of techniques (e.g., affective statements, circles, conferences) that are used to strengthen relationships and restore community in response to harm (Kline, 2016; Zakszeski & Rutherford, 2021).
Although united by an overarching philosophy focused on community and restoration, the specific school-based practices can vary across contexts. Zakszeski and Rutherford (2021) and Lodi and colleagues (2022) conducted recent systematic literature reviews that describe the specific school-based practices that are considered “restorative.” Based on 105 articles included in these reviews (n = 71 for Zakszeski and Rutherford, and n = 34 for Lodi et al.), restorative conferences and responsive circles were the most prevalent and well-defined approaches, described across most studies (n = 48 and 43, respectively). These approaches are used in response to harm or misbehavior, either in a small (conferences) or large (circles) group setting. Additionally, proactive circles to build community (n = 25), peer mediation in response to conflict (n = 21), and impromptu restorative conversations/conferences between educators and students in response to harm or misbehavior (n = 18) were also described frequently. Practices related to logical consequences/restitution for harm caused (n = 9), circles for re-entry after exclusionary discipline (n = 1), and dialogue about issues related to equity and social justice (n = 1) were the least common.
When implementing new approaches, schools and districts often turn to vendors for curriculum and training, making adaptations as necessary to suit their local contexts. One of the largest and most widely studied vendors of the approach in the United States is the International Institute for Restorative Practices (IIRP, e.g., Augustine et al., 2018), whose model was used by the school and districts in this study. The IIRP model consists of tiered practices with tier 1 approaches used with all students to build community and conflict resolution skills, including practices like affective statements (“I” statements), restorative questions (“What happened?”), small impromptu conversations (“How can this be repaired?”), proactive circles (Prompt: “What did you do last weekend?”), and fair process. Tier 2 practices consist of large-group responsive circles in response to harm (“What would fix the harm?”) and reintegrative management of shame practice focused on reintegrating students who have been excluded due to behavior (e.g., when returning to school after suspension), which are intended for use in response to conflict with about 10–20% of students. Lastly, restorative conferences (small conferences with those most impacted by misbehavior or conflict) are a tier 3 practice used with select students (3–5%). IIRP recommends a “whole-school” approach where both proactive and responsive elements are used by all staff, including classroom teachers;the organization recommends that 80% of circles used in schools should be proactive, as opposed to responsive. The dual focus on community and restoration is critical for two main reasons. First, the proactive elements can work to prevent misbehavior from occurring in the first place as students begin to value and care for their community members. Second, restoration after harm is only possible if there is a genuine sense of community to begin with; you cannot restore a community if it has not been intentionally built and nurtured.
Significance of the Restorative Practices Implementation Process
The causal evidence on the effectiveness of restorative practices shows that they reduce suspension and discipline incident rates (Acosta et al., 2019; Augustine et al., 2018; Grant et al., 2023; Gregory et al., 2022). Despite this clear benefit, impacts on racial disparities, school climate, and student absences remain less conclusive (Augustine et al., 2018; Gregory et al., 2022; Sadler, 2021). Additionally, the effectiveness of the approach appears to vary by school level in some studies. For example, Augustine et al. (2018) found that reductions in suspension rates were driven by elementary schools where educators reported using the practices more than secondary teachers, leading the authors to conclude that restorative practices were a particularly promising approach for elementary schools.
How the approach is implemented could explain why the effects of restorative practices range across studies and outcomes. Implementation fidelity, or the extent to which an intervention aligns with its intended design, is measured across multiple dimensions, including dosage or frequency and program adherence (Abry et al., 2013), and is a common focus of implementation studies (Proctor et al., 2011). For an approach like restorative practices that requires behavioral and cultural changes, some local adaptation where the practices are altered to fit local contexts or needs might be expected (Augustine et al., 2018; Bryk, 2016). As with the implementation of education reforms more broadly, previous studies of restorative practices have shown that greater implementation fidelity results in more positive outcomes (Gregory et al., 2016; Wong et al., 2011).
The existing research on restorative practices suggests that its enactment is contingent on capacity building (Anfara et al., 2013; Guckenburg et al., 2016; Sumner et al., 2010), buy-in (Anfara et al., 2013; Liberman & Katz, 2021; Wang & Lee, 2019), and leadership (Abdou et al., 2023; Anyon, 2016; Joseph-McCatty & Hnilica, 2023). We summarize the existing implementation research related to each area. Building local actors’ capacity to enact a new program is a critical first step for implementation. Educators require continued and intensive PD to learn new approaches like restorative practices, including how to build rapport with students, create an inclusive school climate, and problem-solve with students to repair harm (Anyon, 2016; Mayworm et al., 2016). Because the approach represents a significant shift from previous disciplinary approaches, educators may struggle to deviate from a focus on punishment. For instance, Lustick (2022) documented how educators in an urban charter high school, despite explicit commitments to reducing racial discipline disparities, frequently reverted to exclusionary disciplinary practices when faced with misbehavior—illustrating how deeply ingrained punitive approaches remain and highlighting the critical need for leadership to anticipate resistance during implementation. Given the approach’s novelty, building educator knowledge, skills, and mindsets might be especially important for enactment. Prior studies have shown that effective capacity building for restorative practices goes beyond training, encompassing adequate time, personnel, and funding for schools to develop, test, and scale these practices (Anfara et al., 2013; Guckenburg et al., 2016; Sumner et al., 2010).
Capacity building alone does not guarantee that new practices will be adopted; educators must also feel motivated to enact new practices. Because restorative practices require educators to shift their attitudes and practices, these programs are most effective when there is buy-in from school leadership and staff (Anyon, 2016; Burke & Ashley, 2008; Hall et al., 2021; Morrison et al., 2005). Educators who understand and support the approach report greater positive effects of the approach in their schools and classrooms and fewer challenges in their implementation (Dhaliwal et al., 2023). However, building buy-in is particularly challenging because restorative practices can contradict mainstream punitive approaches, with some practitioners viewing the approach as soft on students (Liberman & Katz, 2021). Leaders can play a critical role in enhancing buy-in and implementation by recognizing that exclusionary discipline practices are ineffective, by maintaining commitment to the approach when faced with resistance to changing discipline, and by allocating the necessary resources for capacity-building efforts (Anyon, 2016; Joseph et al., 2021). Because enacting restorative practices school-wide requires a sustained effort, minimizing leadership turnover can also create consistency and enhance implementation (Abdou et al., 2023).
Perspectives on Policy Implementation
Policy implementation is a complex process by which people, organizations, and policies come together to shape implementation processes and outcomes (Honig, 2006; McLaughlin, 1987). It is highly contingent on the people tasked with implementation, the organizations they are a part of, and the policy design itself. We draw on cognitive and organizational perspectives, along with the idea of policy coherence, to explain how cognition, organizational norms, and the policy environment constrain and enable implementation fidelity. Our study draws on these perspectives to motivate our investigation of how organizational factors, contexts, and conditions foster implementation.
Cognitive approaches to policy implementation underscore how interpretation and the meaning-making process guide policy enactment (Coburn, 2005; Spillane et al., 2002; Weick, 1995). Local implementers come to understand and construct meaning from policy messages based on their pre-existing beliefs, depth of knowledge, and environmental cues. Policy enactment depends on how implementers interpret and encode information, which might involve restructuring existing schemas or beliefs (Resnick, 1991; Zerubavel, 2000). Organizational theorists remind us that cognition is also situated (Powell & DiMaggio, 1991; Scott & Meyer, 1994). Institutional norms, processes, and social relations shape agency and the cognitive process, resulting in policy messages being understood differently across contexts (Ball, 1994; Coburn, 2001; Spillane & Jennings, 1997). This situated cognitive process of interpretation and meaning-making is influenced by the degree of congruence or alignment between new practices and educators’ existing beliefs and schemas. Congruence serves as a key mediator in implementation, with teachers rejecting or superficially implementing practices they perceive as incongruous with their beliefs and implicit cognitive frameworks (Bridwell-Mitchell, 2013; Rigby, 2015) or failing to make changes if practices are seen as too similar to existing methods (Coburn, 2004; Datnow & Castellano, 2000). This creates an optimal zone for implementation where practices need to feel aligned with educators’ existing practices to be credible, yet different enough to prompt change (Coburn, 2004; Spillane et al., 2002; Yoshizawa, 2025).
Educational leaders play a critical role in shaping how policy messages are understood and acted upon by staff, including what practices are emphasized and ignored in schools. Through their positional authority, leaders can shape policy messages and educator interpretation by framing and filtering specific aspects that are selectively (de)emphasized (Coburn, 2005; Diehl & Golann, 2023). For example, Coburn (2005) finds that principals influence teachers’ sensemaking of reading practices by filtering policy ideas and creating organizational conditions to support uptake. Leaders are also critical partners in helping teachers consider what parts of new practices might be congruent with existing practices and what parts might conflict with them (Datnow et al., 2023; Woulfin & Rigby, 2017). By helping teachers recognize what is new and what is familiar, leaders can help support changes in teaching practice. Beyond supporting teacher interpretation of policy messages and new practices, leaders help educators prioritize their actions and promote implementation through the strategic deployment of organizational expectations, resources, and structures. These actions can enhance enactment because educators are more likely to adopt new practices when they are woven into school expectations, norms, and routines (McLaughlin, 1987), including in the case of restorative practices where school behavioral expectations and discipline processes can foster—or inhibit—implementation (Lustick et al., 2024).
The policy environment can also shape how policy is enacted, especially in education, where schools are tasked with integrating multiple policy priorities with varying degrees of alignment. Unlike congruence, which focuses on micro-level interpretation and alignment, policy coherence addresses how organizations craft systemic alignment across reform efforts. Policy coherence was largely viewed as a problem of policy design, where incoherence stems from policies that have competing or misaligned goals—a conflict that can be resolved in policy design (Smith & O’Day, 1990). However, more recent research has viewed coherence as an ongoing process where teachers and administrators craft coherence, or a sense of alignment between new demands and concurrent efforts, as they negotiate and manage multiple demands in their daily work (Honig & Hatch, 2004).
Educators can craft coherence by reshaping external demands to fit school and district priorities. For example, leaders may create a “bridge” between a seemingly disconnected policy and an existing school priority by framing it as a practice that could enhance existing priorities, thereby creating a sense of alignment and coherence. Leaders can also “buffer” schools from external demands that are misaligned with strategic priorities by selectively engaging with policies. For example, leaders might decline a specific funding stream because the programmatic requirements are inconsistent with school priorities. This dynamic view emphasizes local implementers’ agency in managing policy demands, highlighting how restorative practices enactment may depend on educators’ capacity to reconcile competing demands or build connections between complementary policies (Honig & Hatch, 2004).
Study Context
West ISD comprises roughly 50 districts, with four of the ISD’s districts featured in this study (Ontario, Iris, New River, and Oak School Districts). 2 Districts range in size from roughly 2,000 to 10,000 students, with all districts serving significant economically disadvantaged student populations ranging from 50% to 80%. The study districts differ in racial composition. Two of the districts (Ontario and New River) are majority White with large Latino student enrollment, and two of the districts have majority students of color, with Latino (Iris) or Black students (Oak) as the largest single group. The average suspension rates range from 5% to 20% and reflect previously documented national patterns where most districts tend to suspend Black students at higher rates and most districts suspend Latino students at similar or lower rates to the average suspension rate. See Appendix Table A1 for the demographics of these districts.
Beginning in the 2021–22 school year, the ISD and the districts made considerable investments to train educators and lay the foundation to scale restorative practices using the IIRP whole-school model. Prior to this, the districts offered educators PD (typically a 1- to 4-hour training session) on the approach on an annual or bi-annual basis. However, the pandemic-related school shutdowns and changes in instructional modality disrupted the implementation of the approach. Once normal school operations resumed and students readjusted to in-person learning, leaders and educators recognized an urgent need to rebuild a sense of community and address misbehavior. In the 2021–22 school year, and with financial investment from the ISD, the districts adopted a “train-the-trainer” model to deepen their staff’s understanding, skills, and use of restorative practices, where a cohort of staff members from each district attended an in-depth 3- to 5-day IIRP training designed to equip staff members with a comprehensive understanding of the approach and tools to train others in their respective districts. In the 2022–23 school year, staff members who attended the intensive workshop trained educators in their respective districts at least once in the school year (ranging from 3 to 6 hours), including teachers, administrators, and support staff. 3 By the 2022–23 school year, 78% of educators in these districts reported that they had been trained on the approach; this rate ranged from 93% in Oak School District to 69% in Iris School District. The ISD also made restorative practices coaches available to lead training by request and offered a monthly professional learning community for teachers during the 2022–23 school year.
Policy Context
These districts and the ISD decided to deepen their investment in restorative practices because of the state’s emphasis on the approach and because they wanted to reduce suspensions, improve teacher–student relationships and school climate, and increase student achievement. In Michigan, restorative practices gained significant attention after August 2017, when the state legislature required schools to consider using these practices as an alternative or supplement to exclusionary discipline (e.g., suspensions, expulsions) for a broad range of behaviors including interpersonal conflicts, vandalism, and classroom disruption (Mich. Comp. Laws § 380.1310c, 2016). The state law did not provide additional funds for the approach, nor did it require monitoring of their use. More recently, and since the pandemic, the Michigan attorney general urged superintendents to address the excessive use of suspensions and expulsions through the use of restorative practices (Nessel, 2021).
Data and Methods
We use an explanatory sequential mixed-methods design to understand the state of implementation and what factors might be shaping it. Explanatory designs enable researchers to follow up on quantitative results by using qualitative data to explain, provide context for, and enhance the interpretation of quantitative results (Creswell & Plano Clark, 2018). In this way, we leverage interview data to formulate explanations for patterns detected in quantitative data. For example, based on the quantitative differences we find in survey responses by school level, we examine our qualitative data to offer explanations for these differences and deepen our understanding of implementation across different school contexts. In what follows, we describe our quantitative and qualitative data sources and methods.
Quantitative Data Sources and Methods
Quantitative data used for this paper draws from three main sources: (1) a survey of school personnel (school-based teachers, administrators, and other staff), (2) a survey of students in grades 6–12, and (3) linked student and staff administrative data. The educator survey was administered online using Qualtrics between December 2022 and February 2023. The overall educator response rate was 41.1%, ranging from 19.0% to 75.9% in districts. The student survey was administered online via Qualtrics to students in grades 6–12 between December 2022 and January 2023. 4 In the vast majority of cases, students took the survey in class, and researchers provided their teachers with scripts for uniform administration. The overall student response rate was 50%, ranging from 37% to 62% in districts.
To address potential nonresponse bias in our survey data, we applied poststratification weights based on 2022–23 student- and staff-level administrative data. We stratified the student results by race (non-White and White), gender (male and female), and free lunch status (free or reduced priced lunch or no free or reduced priced lunch); the teacher results were stratified by gender and years of experience (whether the staff was in their first five years with the district). These factors were chosen due to their significant association with survey response probability. Unweighted results are qualitatively similar to the weighted analyses reported in this paper.
The educator survey focused on several concepts related to how staff perceive implementation and how frequently they use these practices. In this study, we use items related to the frequency of use of proactive circles (asked of classroom teachers and teacher leaders) and formal restorative conferences (asked of administrators and support staff, including counselors, behavior specialists, social workers, and other nonteacher roles), which inquired how often they use each of these practices per week on average over the past four weeks. The items included definitions of the practices, with the answer choices being 0, 1–2, 3–5, 6–9, and 10+ per week. We also administered a seven-item scale measuring perceptions of implementation based on Augustine et al.’s (2018) school staff survey, which included items related to student input in decision-making, whether school policies and procedures emphasize preventive and restorative approaches, teacher cooperation in the use of restorative practices, and the extent to which families engage with the approach. Responses were recorded on a four-point scale (strongly disagree to strongly agree), and we averaged these items together to generate an overall scale measure (Cronbach’s alpha = 0.75).
We draw on student survey items related to general perceptions of restorative practices, as well as specific perceptions about circles, conflict resolution, and labeling. These items assessed the degree to which students witness educators using approaches and have confidence in their impacts on community building, conflict, and school connectedness. All items were measured on a four-point scale, with “not at all” to “always,” “not at all” to “to a great extent,” or “disagree” to “agree” scale anchors.
For both the educator and student surveys, we created a binary indicator for each item representing whether the educator or student endorsed a favorable or unfavorable response to the question, where the indicator was coded as 1 if the survey responder endorsed a 3 or 4 on the scale (the highest response options) and coded as 0 if they responded with a 1 or 2 on the scale (the lowest response options). This procedure allows us to discuss our results as the percentage of educators or students in a school who endorse a favorable or unfavorable response to the item. See Appendices A2 and A3 for more information on the staff and student survey items.
Analytic Methods for Quantitative Data
We reported descriptive statistics of pooled survey responses to the items described above overall, by district for the educator surveys, and by school level for both the educator and student surveys. To test whether results are statistically significantly different by these subgroups, we used chi-squared tests to report statistical significance in differences across districts and schooling levels. Because we found differences in survey results by district in some cases, we also used a regression model with district fixed effects to examine whether the differences by school level hold when making within-district, across-school comparisons. We estimated variations of the following linear regression model:
where
Qualitative Data Sources and Methods
Our qualitative data collection focused on interviewing school administrators and staff at two or three schools within each district, totaling nine schools across districts, including four elementary schools, four high schools, and one middle school, in spring 2023. The schools were selected, in consultation with district administrators, to reflect a range of school levels and a range of implementation stages. We interviewed 47 participants in total, including principals, assistant principals, teachers, and other staff (e.g., counselors, social workers, behavior specialists) at each school site using a semistructured protocol. All interviews were recorded and transcribed. The interview protocol probed the cognitive and organizational aspects that shape implementation and included questions related to school culture and values, school and classroom approaches to behavior and discipline, restorative approaches used in the school and classrooms, alignment between restorative practices and other practices, and perspectives on restorative practices implementation.
Analytic Methods for Qualitative Data
For qualitative data analyses, we used a multiple case study design (Stake, 2005; Yin, 2018). Each district site serves as an embedded case, with schools embedded within each district (Yin, 2018). The primary units of analysis are the district and school. All interviews were coded both deductively and inductively using qualitative software (Dedoose), and researchers wrote detailed school case study narratives. The deductive codes were based on the broad conceptual categories of our interview protocol (e.g., factors affecting restorative practices implementation), with more granular codes that closely described the coded excerpts (e.g., staff buy-in, student support for restorative practices, and time/curricular responsibility). We used a hybrid coding approach to capture additional ideas that were not adequately captured by our existing conceptual categories but were present in our data. After coding, we documented themes in case narratives for each school where researchers synthesized emerging themes, including how perspectives converge or diverge within themes related to capacity building, the school’s behavior management system, and the coherence of PBIS and discipline systems with restorative practices. We used a cross-case matrix to compare themes by school site and district (Miles et al., 2014), allowing us to highlight similarities and differences between these contexts. Finally, we wrote narratives for each set of themes that connected findings back to key ideas related to cognitive and organizational perspectives in implementation (e.g., beliefs, understanding, structures and processes, culture, coherence), revisiting the case memos and interview transcripts as necessary to validate the linkages to key theoretical ideas.
Ethical Standards and Ensuring Trustworthiness
We took several measures to maintain the ethical integrity and trustworthiness of our study. After obtaining institutional review board approval, we secured consent and anonymized all participant data using pseudonyms to protect identities. Our research team used collaborative analysis, member checking, and triangulation across data sources to strengthen our analysis. Three researchers initially coded interview transcripts together to norm on code definitions and exclusion/inclusion criteria for each code. Regular team meetings were used to foster transparency and reflexivity in our analysis by discussing emerging findings, themes, and coding queries (Saldaña, 2013). We also conducted member checking by discussing our findings with district partners (Miles et al., 2014). For member checking, we shared results with districts during convenings where we discussed preliminary results and invited district feedback on our interpretation. These discussions largely confirmed our results, further validating our findings.
Findings
Our findings are organized around three research questions that examine the implementation and perceptions of restorative practices across multiple school districts. We explore the frequency of restorative practice use, compare educator and student perspectives on its implementation, and identify key organizational factors that influence the adoption and sustained use of these approaches.
RQ1: To What Extent Do Educators Report Using Restorative Practices, and How Does Their Use Vary Across School Characteristics?
We report the differences in the frequency of use of proactive circles (Figures 1 and 2) and formal restorative conferences (Figures 3 and 4) across districts and schooling levels. Beginning with Figure 1, we find substantial and statistically significant (p < 0.001) variation in the use of proactive circles and differences between districts. For example, teachers in Oak School District report using proactive circles more frequently than other districts, with only 18.9% reporting using zero on average, per week over the last four weeks. Meanwhile, Iris School District teachers report using proactive circles least frequently, with 71.9% of those respondents reporting using zero on average, per week over the last four weeks. Our results in Figure 2 also show statistically significant differences by school level (p < 0.001). Elementary school teachers were less likely to report conducting zero proactive circles (39.6%) than those in middle (61.7%) and high schools (60.6%). Across all schooling levels, about a quarter of respondents indicated using one or two proactive circles per week over the past four weeks. However, over a third of respondents (34.7%) in elementary schools reported leading three or more proactive circles, a much greater share than in middle schools (9.6%) and high schools (16%). These results suggest that proactive circles are being used less frequently than recommended, with greater use in some districts and elementary schools.
Frequency of Proactive Circles Used per Week by Teachers and Teacher Leaders, by District and Overall.
Frequency of Proactive Circles Used Per Week by Teachers and Teacher Leaders, by School Level and Overall
Frequency of Formal Restorative Conferences (Administrative and Support Staff), by District and Overall.
Frequency of Formal Restorative Conferences (Administrative and Support Staff), by School Level and Overall.
Similarly, we asked other roles (e.g., administrators, counselors) about the frequency of leading formal restorative conferences, on average, per week over the last four weeks. As shown in Figure 3, we see differences by districts, but we lack the statistical power to detect any statistically significant differences (n = 53, p = 0.517). Descriptively, New River School District appears to have substantially higher use compared to other districts, where the vast majority of respondents (79%) reported using these conferences one or two times a week. In Figure 4, we also have limited power to detect statistically significant differences by school level (n = 52, p = 0.469). Respondents in elementary schools appear less likely (35.8%) to report leading zero formal restorative conferences, on average, relative to respondents in middle schools (68.3%) and high schools (60.2%). Regardless of statistical significance, interpretation of higher or lower use of formal restorative conferences is challenging, because higher values may mean more restorative use (relative to other more punitive practices), or higher values may mean there are simply more conflicts or behavioral issues that require resolution.
Along with the frequency of practices, we are also interested in examining whether these practices are woven into school policies, culture, beliefs, and norms, with educator survey results related to these items shown in Table 1. There were very few differences across schooling levels, so we report the overall percentage that agreed or strongly agreed with each item. 5 We find a high level of agreement (nearly 94% or higher) with several items related to engaging students in decision-making, seeking their input, and explaining decisions to them. Some 73% of respondents agreed that policies and procedures emphasize preventative and restorative approaches, but only 26.6% reported that restorative practices have been explained to family and community members. Respondents generally indicated that teachers participate in specific practices as needed (64.6%), but many fewer (22.1%) agreed that family/community members participate as needed.
Educator Perceptions of Restorative Practices Implementation.
Note: The percentages represent the share of educators who endorsed a favorable response to each item (i.e., responded with a 3 or 4), with poststratification weights. The survey items are adapted from Augustine et al.’s (2018) school staff survey.
RQ 2: What Are Students’ Perceptions of Restorative Practices Use, and How Do These Align with Those of Educators?
Students are key stakeholders and partners in implementation, so we examine the share of students who responded favorably to items related to the approach’s use and impacts. As shown in Table 2, we see that the majority of students, and in most cases the vast majority, slightly agree or agree that adults use common restorative approaches like teaching strategies that help set positive norms for student behavior (71.9%), teaching them how to resolve conflicts in healthy ways (56.2%), using “I” statements to express their feelings (68.6%), and avoiding labeling students by separating the deed from the doer (60.1%). Students believe these things to be true even though roughly half of the educators did not report using proactive circles or formal restorative conferences in the last month. This might suggest educators rely on other strategies like affective statements, restorative questions, and impromptu conferences to teach these skills or other approaches to conflict resolution. Regardless, the majority of students feel like many of the pillars of restorative practices are reflected in their interactions with educators, even without the regular use of circles. Although most students slightly agree or agree that restorative practices build community (62.7%), they are less positive that the practices reduce conflict (37.4%) and that circles, specifically, help students feel closer to their peers (41.3%). Additionally, less than half of students (46.2%) agree that they use strategies for resolving conflicts outside of school. These results suggest students recognize adults in their school use restorative approaches, but fewer students agree that these practices reduce conflict or that circles help them feel closer to their peers—possibly reflecting educators’ inconsistent use of circles. 6
Student Perceptions of Restorative Practices Implementation.
Note: The percentages represent the share of students who endorsed a favorable response to each item (i.e., responded with a 3 or 4 on the four-point Likert scale). Poststratification weights applied.
We next look at these items disaggregated by middle and high school students to explore the degree to which there are significant differences in perceptions by school level. As shown in Table 3, we find that middle school students consistently report more favorable perceptions of adults in their schools using restorative practices compared to high school students. For instance, middle school students are 21.3 percentage points (PPs) more likely to slightly agree or agree that adults teach conflict resolution skills. This trend persists across other items where middle school students are 11 PPs more likely than high school students to slightly agree or agree that adults help students resolve conflict healthily and teach positive behavioral norms, and 9.3 PPs more likely to slightly agree or agree that adults encourage the use of circles to build community and trust. Interestingly, educators in middle and high schools tend to report using proactive and restorative conferences with similar frequency and have similar perceptions of implementation.
Student Perceptions of Restorative Practices Implementation Across Schooling Levels.
Note: *p < 0.1; *** p < 0.01.
This table presents the results from a series of regression models in which each survey item was regressed on a binary indicator for middle schools. (High schools were the reference group.) Each row is the result of an ordinary least squares (OLS) regression model wherein the outcome is the share of students in the school who selected the highest and second highest responses (slightly agreed or agreed, often or always, and to a moderate extent or to a great extent) on each survey item. Coefficients represent the average difference in the outcome between students in middle and high schools. All models include poststratification weights, district-fixed effects, and robust standard errors clustered by district.
Despite these differences in the observed use of restorative practices, middle and high school students generally feel similarly about their perceived impacts. We do not see evidence that middle and high school students feel significantly differently about the degree to which circles reduce conflict and help students feel closer to their peers. However, middle school students are 8.5 PPs more likely to slightly agree or agree that the practices build community (p < 0.1) and 6.1 PPs less likely to report using them outside of school (p < 0.01).
RQ 3: How Do School Organizational Factors, Context, and Conditions Shape the Use of Restorative Practices?
Our analysis across sites reveals key factors shaping restorative practices implementation: leadership, alignment, and office discipline referral routines. Notable variations between school levels and districts help explain patterns in our survey data. Our results highlight how understanding, pre-existing knowledge and beliefs, and strategic framing informed educators’ understanding of implementation and their use of the approach.
The Role of Leadership in Promoting Restorative Practices
Leaders at the district and school levels promoted the use of the approach by modeling the practices for staff, integrating restorative practices into strategic goals, and expressing a long-term commitment to the approach. These actions were related to school leaders’ depth of understanding of the approach and prior experiences, and they ultimately shaped how educators went about using the practices. Although Michigan policy mandates the consideration of restorative practices in disciplinary decisions, it provides little guidance on whether and when the community-building aspects of the approach should be used in schools. School leaders in all districts stated they considered restorative practices when making disciplinary decisions in ways consistent with state policy. However, leaders varied in the extent to which they emphasized a whole-school approach, including using restorative approaches in the classroom, as did teachers when they described their classroom practices.
Classroom teachers in Oak School District, the district with the highest level of proactive circle use in survey data, stood apart from other districts. Nearly all educators we interviewed in the district shared examples of how they used proactive circles in their classrooms. When we asked if they were expected to use restorative practices, educators referenced a district-wide goal of conducting “a circle every week in every classroom,” and they shared that district and school administrators modeled the use of circles in staff meetings and professional development. The district-wide expectation included room for discretion, as one school administrator explained: So, there isn’t like you have to do it first thing in the morning or you have to do it midday, but we allowed teachers in our building to find that natural spot where it happens. And so now you can walk into a room, and you will see circles or something that resembles a circle going on.
Leaving room for flexibility allowed teachers to exercise autonomy in when and how to achieve the weekly circle expectations. Importantly, the district-wide expectation was accompanied by informal monitoring, with central district administrators looking for circles when walking through classrooms and talking with principals about the frequency with which educators were using circles. District leaders in Oak School District were described as “champions for restorative practices” with a deep understanding of the approach, especially the community-building aspects. As one administrator put it, “When your [district administrators] lead in that [restorative] way, it becomes a part of who you are, not just something that you do.” The depth of district leaders’ knowledge of the approach—and that a “whole-school” approach is most effective—resulted in an expectation that all district educators were to be involved in enacting approaches in their classroom, not just administrators who make disciplinary decisions (e.g., suspensions).
In the Ontario School District, where district-wide expectations were absent, school leaders played a critical role in setting expectations about the use of the approach and, in particular, proactive circles. For example, at the elementary campus, the use of proactive circles became a school improvement goal following previous challenges using restorative conferences. As an administrator explained, What we didn’t recognize is if you’re going to jump right to restorative [conferences] without doing proactive circles, the kids just don’t understand what’s going on. So, one of our school improvement goals this year is really getting into and helping teachers model what a proactive circle is going to look like, because then when we get to a restorative [conference], it’s a little bit more authentic and the kids understand it.
As a result of these previous challenges, the administrator saw value in using proactive circles in the classroom to develop students’ familiarity with and skills to participate in conferences. This school leader demonstrated a thorough grasp of restorative practices and, because of prior experiences, saw classroom use as an essential element. In another campus within the Ontario School District, a high school administrator shared that setting clear expectations and assuring staff that restorative practices would be a consistent program at the school has helped persuade staff to invest in using these practices. The administrator explained, [Staff] want to know that whatever training they’re required to attend is going to be implemented, and it’s not just going to be something we touch on today, and then we move past and jump to something else later. So, from our end, it was very important to let the staff know that we’re not moving away from this. . ..
The administrator went on to explain that emphasizing the approach will be a sustained practice, and being transparent about expectations encouraged staff buy-in. In this case, the administrator made a long-term commitment to restorative practices to signal to teachers that using the approaches is a key priority, and they emphasized that classroom teachers should “use the process, be part of the process, and make it part of what [they] do every day.” Teachers at this site shared examples of how they use restorative practices with regularity; however, survey data from the district suggests that half of the teachers did not conduct any proactive circles in the past month. In both of these examples from the Ontario School District, in the absence of clear district-wide expectations, school leaders used their positional authority to emphasize how classroom teachers should use the approach and framed its use as a sustained emphasis at the school to encourage uptake.
In the other district sites, Iris and New River School Districts, school administrators shared that although most classroom teachers had been trained on the approach, they were not expected by the school or district to use it in their classrooms. Leaders prioritized using restorative approaches for office referrals rather than emphasizing proactive classroom elements, believing administrators had more time to “sit with the kids.” Mirroring state policy framing, restorative practices were primarily viewed as reactive tools for addressing office referrals and discipline issues. Teachers internalized this message, with a high school teacher in Iris School District explaining that “[Restorative practices] is viewed as something that the counseling office does as part of their job. It’s not the teacher’s job.” This framing of the approach as an administrative tool was also reflected in survey data. In the Iris School District, where teachers were not expected to implement proactive circles, more than 70% of teachers reported using zero proactive circles per week. Meanwhile, survey data from New River School District underscores its use as an administrative tool: 79% of administrators and support staff used one or two formal restorative conferences per week on average. In comparison, only 35.2% of the district’s teachers used proactive circles with the same frequency. As one administrator noted in our study, limiting restorative practices to disciplinary issues may undermine their effectiveness, because students unfamiliar with these approaches from classroom use struggle to engage authentically during conflict resolution conferences.
Alignment with Existing Practices and Culture
Alignment between existing school practices and culture was another important factor that shaped educators’ use of restorative practices. We find a particularly strong alignment in elementary schools, likely explaining why elementary teachers reported higher usage in our survey (Figures 2 and 4) and in existing studies (Augustine et al., 2018). Across districts, elementary educators drew connections between the approach and existing practices, with some noting that they were using restorative philosophy (e.g., focus on relationships and community, addressing behavioral incidents with the class in small groups or conferences) before receiving formal training. An elementary teacher described attending the training and remarking: “That’s exactly what we [already] do with our kids.” This connection resulted in the approach being more seamlessly integrated into existing cognitive frameworks about teaching. Conversely, high school teachers commented that using the practices in their classrooms felt “elementary” and was at odds with their role as a “content expert,” whose priority was to deliver the content rather than focus on the “whole child.” These deep-rooted, pre-existing beliefs and professional logics about what it meant to be an educator varied across elementary and secondary teachers and influenced how teachers saw their role in implementation.
Elementary educators also shared how they adapted pre-existing practices so that they could more seamlessly integrate the proactive approaches into their daily practice. For example, one teacher explained, “We’ve always done morning meetings. It’s just now calling it a circle.” The teacher went on to explain how they changed the purpose and process of morning meetings to better align with proactive circles. In another example, educators shared that the school also used a community-building approach of “sharing good news” via circles, which “really builds a sense of community in the classroom, as well as builds trust among peers because they have eye contact with one another.” These adaptations of pre-existing strategies were much more frequently described by elementary educators than by secondary educators, suggesting that elementary school campuses were more likely to have pre-existing organizational routines that were congruent with restorative practices.
Many of the elementary campuses we profiled also saw strong alignment between restorative practices and pre-existing systems like PBIS. Elementary campuses in the study had well-established PBIS systems, especially in comparison with the secondary campuses. Most elementary school teachers were able to describe the school’s PBIS system and how they incorporated it in their classrooms. At these schools, staff were able to articulate how restorative practices related to PBIS and existing practices. For example, an educator described how these approaches work together: PBIS has done a good job of setting consistent expectations through the building. . .. But PBIS [wasn’t] providing us with a tool [to] encourage kids to necessarily meet the expectations. . .. I feel like [restorative practices] provides more of an opportunity for the why behind the expectations.
These reflections were echoed in other elementary campuses where interviewees described these approaches as mutually beneficial, with each strengthening the other. Because elementary campuses had well-established PBIS systems, once restorative practices were introduced, educators commented on how these approaches would “be able to work hand in hand together” to support behavior in classrooms.
This understanding was also shaped by how leaders framed PBIS and restorative practices for their staff. As one elementary administrator explained, “I look at restorative practices and PBIS just being linked and leading us to the same spot. So since we’ve already been doing PBIS for a while, adding [restorative practices] is not adding more, it’s just reframing our mindset.” By building linkages between the pre-existing PBIS model and restorative practices, school leaders framed the new approach as an aligned method and highlighted how it bridged with the school’s pre-existing focus on PBIS. That educators found this alignment is notable, because the extent to which PBIS is aligned with restorative practices is debatable; there are clear philosophical distinctions between the approaches (Sugai et al., 2000; Zehr, 2002) as well as examples of how these approaches can be jointly leveraged (Sprague & Tobin, 2017). Regardless of actual alignment, educators who experienced restorative practices as aligned with pre-existing structures, including PBIS, and part of a coherent system, shared examples of how they used the practices in their classrooms.
Meanwhile, PBIS was a fairly new system in secondary school sites, with some adopting PBIS and restorative practices concurrently. Rather than seeing these systems as complementary, educators were unsure of how they were different from each other, and the dual emphasis on restorative practices and PBIS created a sense of overwhelm. A high school administrator explained that trying to enact both simultaneously “feels overwhelming because we already have so much on our plates, it just adds another thing to the plate that’s already really full. . .it’s just a lot of information overload.” Instead of seeing both approaches as interconnected, as was the case in elementary schools, administrators and educators at secondary campuses felt swamped by multiple initiatives aimed at improving behavior and school climate.
Encouraging the Use of Restorative Practices in the Office Referral Process
Districts also modified their referral process and data tracking systems to include restorative practices, thereby encouraging classroom teachers to use these practices. For example, in Oak School District, office referral forms ask educators whether they used a restorative approach before referring the student to the office. Although not mandatory, teachers reported that this prompt led them to explore alternative solutions before escalating to office referrals. In other districts, schools incorporated restorative practices into the school’s behavior matrix or code of conduct, which outlines possible responses to behavioral infractions and office referrals. Teachers viewed the inclusion of restorative practices in formal documents and organizational routines as a signal that the approach was the presumptive norm in their schools and should be used before escalating to an office referral.
Although formal policies can promote restorative practices, their effectiveness hinges on consistent use, as exemplified by the challenges in the Iris School District. Despite administrators citing policies around restorative practices in disciplinary decisions, staff members disagreed about their consistent application. One staff member noted, “I feel like our discipline is so challenging because it’s not consistent. Every discipline case is handled individually, and some students aren’t given the opportunity for restorative practices that others receive.” When asked to elaborate on these students, they explained, “maybe kids that don’t get in trouble as often, kids of a different race, straight-A students—those types that just aren’t on the [discipline] radar as much” are more likely to have an opportunity to use a restorative approach. Although several districts used comprehensive behavior tracking, few schools tracked the use of restorative practices by classroom teachers and administrative teams. Administrators acknowledged the potential in tracking and analyzing this data, particularly for “monitoring the implementation of restorative practices and observing subsequent behavioral changes.” Tracking the use of the approach might be another way to signal that these practices are valued and should be used at scale.
Discussion
We provide mixed-methods evidence on the extent to which restorative practices are used by educators, experienced by students, and the organizational factors and conditions that might shape these results. By doing so, our study makes two key contributions to the existing literature: (1) we shed light on restorative practices implementation in a novel policy environment; and (2) we highlight how cognitive and organizational perspectives, including ideas of framing, congruence, and coherence, contextualize variability in restorative practices implementation.
Our findings show roughly half of teachers use proactive circles in their classrooms, a core component for community building, preventing misbehavior, and teaching skills to engage in a restorative approach. However, our results also suggest variation in implementation by district and school level, with elementary teachers overall and teachers in Oak School District more likely to report their use. Interviews with school staff suggest that the use of proactive circles is shaped by expectations and commitment to using restorative practices by district and school leaders. Leaders played a critical role in shaping their staff’s understanding of the statewide mandate and adoption of restorative practices by reframing the state policy as a school-wide and classroom approach rather than solely a disciplinary tool. These leaders possessed a strong understanding of restorative philosophy and, in some cases, their prior challenges with implementing the approach informed their decision to pursue a school-wide emphasis. Relatedly, we find that the incorporation of restorative practices in organizational routines like the office referral process influenced educators’ understanding of the approach and encouraged its use.
Additionally, like other studies, we find that elementary educators reported using proactive circles with greater frequency than middle and high school educators. Our qualitative results suggest that elementary educators’ construction of their professional identities and existing practices facilitated uptake. Elementary educators adapted their existing routines related to community building to support classroom implementation, whereas secondary educators less frequently shared examples of this and, instead, saw the approach as conflicting with their professional identities as “content experts.” Elementary campuses also had well-established PBIS systems before adopting restorative practices, and leaders at these sites strategically bridged restorative practices with PBIS by framing the approach as a complementary set of practices. As a result, educators thought of restorative practices as part of a coherent school behavior system, a connection absent in secondary schools where PBIS was either nonexistent or in its infancy.
Slightly less than half of respondents reported using formal restorative conferences at least once per week in the past four weeks, which suggests that restorative practices are being used by administrators and support staff outside of the classroom (e.g., principals, assistant principals, counselors, social workers). In some districts, administrators did not engage classroom teachers as critical partners in enactment. Instead, restorative practices were a tool for administrators to use in discipline decisions, mirroring the language of the state policy. However, overlooking the approach’s classroom utility and neglecting proactive components might limit their positive effect on school community-building and diminish their effectiveness in disciplinary contexts. When students have not been exposed to restorative practices in the classroom, they may lack the requisite skills to meaningfully engage in these processes when serious conflict arises.
Finally, the vast majority of students report that educators help them resolve conflict in healthy ways, avoid labeling misbehaving students, and set positive norms for behavior. Middle school students are more likely than high school students to agree that this is the case. Although most students agree that restorative approaches help build community, they are less convinced that specific practices—such as restorative circles—effectively foster community or reduce conflict. This skepticism may reflect the limited use of proactive classroom circles, as reported by middle and high school teachers, making it less likely that students have experienced their potential benefits firsthand.
Our study is not without limitations, however. We rely on self-reported rather than observational measures of implementation, so our measure of implementation reflects the perceived use of the practices and social desirability bias. This type of measure tells us little about the quality of implementation in districts. Another limitation is that our districts and their schools have been implementing restorative practices for a relatively short time frame, with most only implementing the approach for 1–2 years. Thus, our results are most generalizable to contexts that are early in their implementation, and we might expect to see measures of implementation increase over time or look different in schools with longer implementation time horizons. Lastly, we did not explore how educators’ or students’ perceptions of restorative practices varied by key background characteristics like race and gender. Although we explore heterogeneity by district and school level, there is likely more granular variability that we did not investigate, which is an area for future research because educators and students of color might experience the approach in distinct ways.
Connections with Existing Research and Theory
As suggested by cognitive and organizational perspectives on policy implementation, our results underscore how framing, meaning-making, and organizational conditions shape enactment. In particular, our results underscore how congruence, coherence, and framing support implementation. Consistent with existing studies, we find that congruence with existing practices, beliefs, and logics facilitates the uptake of proactive circles—a core restorative practices component (Bridwell-Mitchell, 2013; Coburn, 2004; Datnow & Castellano, 2000; Rigby, 2015). Teachers in elementary schools experienced heightened congruence and were able to adapt existing classroom routines to align with restorative practices, whereas secondary school teachers had few aligned practices and held professional logics that conflicted with the approach. Similarly, teachers responded positively when leaders could frame restorative practices as part of a coherent approach, which was easier to accomplish in elementary schools that had already established nonpunitive behavioral systems. Additionally, whereas existing research on framing and filtering often focuses on how leaders emphasize or de-emphasize policy requirements, our results show that leaders can broaden policy messages by filtering “in” related foci. In our study, leaders’ prior understanding of the approach shaped their decision to emphasize aspects like community building, even though this was absent in the state policy. By broadening their conceptualization of the state policy, leaders shaped educators’ understanding and actions.
Implications for Policy, Practice, and Future Research
Restorative practices have gained popularity as an alternative approach to punitive discipline, with (Georgetown Law Center, 2020). In many ways, the districts in our study reflect the typical implementation process, where districts offer professional development without additional state funding on a yearly or semiannual basis. Published studies, particularly based on randomized controlled trials, tend to profile districts that have much more robust training and implementation plans (e.g., coaching, regular training, facilitators onsite). However, a key strength of our study is that it provides a glimpse into restorative practices uptake under more typical circumstances. Our findings suggest that key proactive practices are not being implemented as frequently as suggested by the model, particularly in secondary schools. Negating the proactive elements of restorative practices in state policies could undermine effectiveness as articulated by the theory of action, which recommends a whole-school approach where all staff use the practices. Policymakers and leaders should consider how to draw on all elements of the approach that are reasoned to work better when implemented as a whole.
Our results also suggest that leadership is a key factor in ensuring uptake, at both the district and school levels. Previous studies have emphasized the role of school leaders in catalyzing adoption and alleviating implementation challenges (Hall et al., 2021; Lustick et al., 2024). We extend this focus on leadership by showing how district and school leadership can support educators’ understanding of the approach and encourage its use by embedding it within or drawing connections to the school’s existing practices and culture. Drawing connections with existing practices might help teachers bridge perceived disconnects. Moreover, when the approach reflects a significant cultural change, leaders should signal the depth of change required by restorative practices by aligning routines and processes with the approach. It is heartening that elementary educators felt that restorative practices are aligned with the conceptions of their roles; however, given that the use and racial disparities in exclusionary discipline grow in secondary schools, this raises questions about how the practices can be implemented with fidelity in middle and high school. Our findings also provide suggestive evidence that organizational context may explain why causal studies have found varying effectiveness of restorative practices across different school levels. Secondary schools might face distinct challenges in institutionalizing a restorative approach due to deeply held beliefs related to professional identity, and a lack of pre-existing behavioral practices or systems that align with the approach.
Our results suggest a few areas for future research. First, we encourage more investigations that compare secondary and elementary restorative practices implementation. Examining differences in norms, culture, and processes that facilitate implementation fidelity might yield insights into the conditions that support secondary school implementation. We also encourage studies to examine implementation at scale, defined broadly to include sustainability, widespread use, and a focus on deep change that alters educator beliefs and how schools operate (Coburn, 2003; Gregory & Evans, 2020). Finally, we encourage more studies that use cognitive approaches to examine restorative practices and related school discipline reforms. These reforms can be particularly complex to implement because they often require fundamental changes in beliefs about punishment. Cognitive theories can be useful for understanding how and to what extent educators restructure existing beliefs to support implementation rather than assimilate the approach into existing practices that lead to little change (Yoshizawa, 2025).
Footnotes
Appendices
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: From the Steelcase Foundation and Michigan State University’s Institute for Public Policy and Social Research.