Abstract
High teacher turnover rates and shortages of quality teachers plague many large, urban U.S. school districts disadvantaging their students who often already come from traditionally underserved populations and communities. Restorative Practices is a quickly growing whole school approach to community building and discipline, but little is known about how it impacts teachers and their career decisions. This study specifically investigates whether assignment to Restorative Practices (in combination with Diplomas Now, RP w/DN) improves school climate and increases teachers’ reported intentions to remain at their school? To answer this question, we analyzed data from a multi-site, cluster randomized control trial of in 25 schools from seven large cities across the United States. Our intent-to-treat analyses find that RP w/DN has a significantly positive effect on school climate (ES = .15, .27). Impacts on teachers’ turnover intentions were non-significant, but it significantly reduced perceptions of problematic teacher absenteeism (a related, exploratory outcome). Overall, the findings from this study suggest that Restorative Practices as a policy can improve school climate, but its relationship with teachers’ career decisions is more complicated.
Notes
1 The Real Justice program, focused on restorative conferencing, was founded in 1994. In 1999, the founders of this program established the broader International Institute of Restorative Practices (IIRP) to train professionals not only in formal restorative conferencing but also in a broad array of informal and preventative restorative practices that build community and trusting, empowered relationships. See below for a more detailed description.
2 Districts most supportive of RP may have responded to participate first, and their decisions about which blocks of schools within the district would participate could have been affected by school leader willingness to participate. Funding was not sufficient for additional blocks of schools to participate in the RP with DN sub-study.
3 Also see Supplemental Table A1.
4 Many survey items came from previously conducted school surveys by the DN developer that had been used in internal evaluation analyses.
5 In middle schools without grade 6, grade 7 was the target group. The student survey sample included an additional block of two schools (for a total of 27 schools) for which data were available.
6 A listing of these items is provided in the supplementary materials.
7 The small number of schools in the sample (limited power) and goal of parsimony suggest a model that does not include all of the adjustment variables.
8 For instance, we could not pursue a three-level model to analyze the clustering within blocks or districts because several of the blocks or districts contain only two participating schools, which is too small for accurate estimation of a model of that type.
9 We also used the Benjamini-Hochberg procedure to confirm the validity of our tests of all outcomes (Thissen et al., Citation2002).
10 The teacher survey did include a self-report measure of whether the respondent was new to the school within the last year. Analysis of survey results did not find a significant difference between treatment and control schools.
11 Greater statistical significance among student school climate factors with smaller magnitude of effects is likely due to the larger sample size and thus, greater power.
12 We did not pursue a full mediation model (e.g., Structural Equation Modeling) as the effect on teachers’ turnover intentions was not statistically significant.