Table 1 Changes to the original protocol approved in Amendment 1

Staff telephone interviews

The protocol originally included conducting interviews with 1 member of the school senior leadership team (SLT) and 2 teaching staff annually (years 1–3) across 40 schools (intervention and control). These were completed as per the protocol for year 1. We do not intend to conduct staff telephone interviews in year 2. We will conduct interviews with 1 SLT member in each of the 40 schools (intervention and control) in year 3. Control schools will be interviewed in term 1, and intervention schools will be interviewed in term 3

Interviews in year 2 were considered unnecessary since we are already collecting other data (e.g. via interviews with action team members, curriculum surveys, focus groups) on how the intervention is progressing in intervention schools. Interviews in years 3 and 1 are sufficient to assess provision in control schools. Some control schools have also reported overburden following year 1 interviews, so we have reduced the number of interviews for year 3. Resources are being re-directed to in-depth case studies of intervention schools (and away from superficial data collection across all schools)

Researcher observations of curriculum delivery

We originally intended to observe n = 1 curriculum session in each school but are now using a curriculum survey circulated to the intervention curriculum co-ordinator in each school to assess what was delivered, how and when. Interviews with curriculum leads will also be conducted

The lead intervention facilitator advised us that observations would create an excessive administrative burden for schools, and our modified approach provides fuller data on implementation of this component

Action group meeting observations

This will be done in n = 10 schools per year rather than n = 20 schools

We are collecting substantial amounts of other data on action groups via facilitator diaries and collection of all action group documentation. The observations act as a check on the validity of diary data provided by facilitators and do not need to be done across all 20 schools each year. We will re-direct the researcher time that would have been spent on this to more in-depth data from case study schools

Case study schools

The protocol originally specified case studies in n = 4 control schools and n = 4 intervention schools. We now plan to conduct case studies in n = 6 intervention schools only

Control schools have complained about being overburdened with fieldwork requests, and we think that asking too much of them may threaten follow-up rates in the trial. The main purpose of the case studies is to capture data on intervention mechanisms. Case studies of control schools will not be informative about mechanisms, but will only inform us about what activities constitute the control condition in the trial, which we are already collecting across all control schools. We have re-directed resources so that we are doing more work in intervention schools (n = 6 schools as case study sites; conducting 1 focus group with staff, 2 focus groups with students and 2 interviews with students who were involved in restorative practices in each school)