Understanding Restorative Practices: Using Implementation Science to Explore the Nuances of School Climate Intervention Effectiveness

Background and Purpose: Restorative practice has been seen as a viable alternative to exclusionary disciplinary tactics (suspensions and expulsions) in schools, which have been shown to have deleterious consequences for both the youth receiving the discipline as well as their non-disciplined peers due to spillover effects. Studies of restorative practices have almost exclusively focused on outcomes, and have demonstrated 1) consistency in the ability to reduce the use of exclusionary practices; 2) inconsistency in the ability to reduce racial disparities in exclusionary discipline usage; and 3) inconsistency in the ability to impact actual academic performance beyond behavioral perception indicators. These inconsistencies are stubborn in large part because to date, no large scale implementation science inquiries have been conducted to clearly identify facilitators and inhibitors of restorative practice across multiple sites. In response, the current study uses implementation science to identify key measures necessary for effective RP implementation across a number of implementation categories and targets, and to assess each school sites progress on implementation fidelity across more than 20 indicators in the nine schools.

Methods: The current study utilizes the ‘RE-AIM’ framework and a 20+ indicator implementation matrix to determine levels of implementation fidelity across nine Just Discipline school sites. Through direct and repeated consultation with school-based restorative practitioners, program administrators, and teacher leaders, the categories were developed, and implementation rubric scores were created to provide measurable assessments of the sites’ progress according to the tenets of RE-AIM: reach, effectiveness, adoption, implementation, and maintenance. Descriptive statistics were used to assess the level of overall and indicator specific fidelity at each site. Fidelity scores were agreed upon by 2 or more research team raters according to measurable data provided by the restorative practitioner and school officials.

Results: Findings capture the process by which the implementation matrix was developed, as well as preliminary findings on implementation fidelity across sites. Indicators with particularly high levels of variation include the frequency of teacher support interactions (providing activity scripts, direct collaboration with teachers), access to referral data, and attendance in teacher liaison meetings. More consistent implementation activities across cites include the numbers of healing circles being conducted, teacher and student survey data completion, and regular meetings with school administrators. Preliminary explanations for variations across indicators are also provided.

Conclusions and Implications: Results here demonstrate significant differences across sites in implementation effectiveness in multiple indicators. As such, results may help illuminate findings as to why there have been some mixed results in studies of restorative practice impact, despite some staunch advocacy for the approach for seasoned practitioners. It may well be that across a wide range of sites, fidelity to implementation may wain in ways that undermine results of more skilled and supported applications of the approach.