Involving young people in changing their school environment to make it safer Findings from a process evaluation in English secondary schools

Purpose – The purpose of this paper is to explore the process of involving students and staff on school action groups, and staff and student experiences of reviewing local data and initiating school-level changes, to address bullying and other aggression. Design/methodology/approach – The authors draw on qualitative, process data collected at four purposively sampled pilot intervention schools in England via semi-structured interviews with school managers, action group members and facilitators (n1⁄4 33), focus groups with students (n1⁄4 16) and staff (n1⁄4 4), and observations. Findings – School staff used multiple methods to recruit a diverse range of students onto school action groups. Locally tailored data reports were an important catalyst for action groups to identify priorities and plan whole school change – both through the process of “validation” (whereby existing concerns were confirmed) and “discovery” (whereby new problems were identified). An unexpected benefit of providing schools with these data was that it triggered analyses of other data sources, including routine monitoring data. External facilitators were important in promoting student voice and ensuring the intervention retained integrity as a whole-school restorative approach. Practical implications – It was feasible to involve young people using action groups, and there was evidence of school-level actions led by students, including in disadvantaged school contexts. Future Health Promoting Schools interventions could incorporate this approach to support locally appropriate, school-level change. Originality/value – The micro-level processes that were observed, whereby action groups interrogated feedback reports and collected additional data, suggest the responsiveness of such youth-involvement interventions to local needs. Contrary to many public health interventions,


Introduction
Reducing bullying and aggression in secondary schools has been a consistent priority within recent British public health, education and criminal justice policies (Department for Children, Families andSchool, 2008, 2009;Department of Health, 2009;Home Office, 2011).In the UK, assaults against ten to 15 year olds are more likely to happen at school than anywhere else (Home Office, 2006).Secondary schools are also the "epicentre" of bullying, the site where it most commonly begins, and where young people are most concerned about being victimised (Chamberlain et al., 2010).Although cyberbullying has brought new dangers and may increase the risk of serious mental health harms and suicide (Lindfors et al., 2012;Moore et al., 2014), most bullying and aggression via the internet and mobile phones is an extension of traditional forms of peer bullying originating in school (Juvonen and Gross, 2008;Cassidy et al., 2009).Furthermore, there is a growing concern regarding the increasing "low level" aggressive and provoking behaviour among British secondary school students, which is educationally disruptive, emotionally harmful, and can lead to more overt, physical aggression and social exclusion over time (Scott et al., 2001;Steer, 2009).
While these problems exist in all schools, the prevalence and frequency of bullying and other aggressive behaviours has been found to vary markedly between different school contexts, both in the UK (Wolke et al., 2001;Stansfeld et al., 2003) and elsewhere (Battistich and Hom, 1997;Marmot, 2004).Furthermore, these school-level differences remain after adjustment for socio-demographic intake, prior behaviours and other student-level factors, which suggests that the school environment itself partly influences these aggressive behaviours at school (Bonell et al., 2013).Reviews of the evidence from trials of school-based interventions also consistently show that multi-level approaches which combine "whole school" changes with the promotion of social and emotional skills are required to reduce bullying and aggression (Hahn et al., 2007;Limbos et al., 2007;Vreeman and Carroll, 2007;Wilson and Lipsey, 2007;Park-Higgerson et al., 2008).This means that prevention activities need to address the whole school environment, particularly the institutional climate and staff-student relations (Gottfredson et al., 2005).A recent systematic review of interventions using the World Health Organization's (WHO's) Health Promoting Schools (HPS) framework also found that using such a multi-level, settings-based approach improves students' health outcomes, including through the prevention of bullying (Langford et al., 2014).
Despite increasing evidence of the centrality of the school environment, the Steer Review (2009) of students' behaviour in British secondary schools concluded that current approaches to discipline, behaviour management and bullying prevention varied widely with little evidence base to support policies or practices in the UK.The Steer Review also called for English schools to consider adopting more restorative approaches to prevent aggressive behaviour and to minimise the harms associated with such problems (Steer, 2009).Restorative approaches have been developed and used widely in the criminal justice system to repair harm caused to relationships and communities rather than assign blame and punishment.Such approaches have now School environment been adapted for use in schools to improve relationships, reduce conflict and repair harm (Hopkins, 2004;Kane et al., 2007;Skinns et al., 2009).An example of a restorative practice that can be employed in schools is the use of "circle time" discussions to develop and maintain good communication and relationships between students (Hopkins, 2004).Restorative "conferencing" can also be used in schools to deal with more serious incidents (Hopkins, 2004).As well as offering a promising way forward for reducing aggressive behaviours among British youth in the secondary school system, restorative approaches have also been found to support whole-school change (Youth Justice Board for England and Wales, 2004;Skinns et al., 2009).
There is now an increasing policy interest in such whole-school restorative approaches to address increasing bullying and other aggressive behaviours reported among British youth, although the best process for implementing such an intervention is far from certain.Evidence-based school environment interventions from Australia (the Gatehouse Project) and the USA (the Aban Aya youth project) suggest how such a process of whole-school change can be achieved through using local needs assessments and staff-student action groups to support the involvement of students in enacting appropriate whole-school actions.As well as promoting the involvement of young people in decision-making about their schools' policies and practices, these two universal school environment interventions take an "ecological" (Dahlgren and Whitehead, 1991) approach to promoting health whereby aggression and violence are understood to be influenced not only by individual characteristics but also the wider social context.
The Gatehouse Project was developed and evaluated in Victoria, Australia.The aim is to improve health outcomes via changing high-school cultures to ensure that they better promote students' security, self-regard and positive communication with staff and other students (Bond et al., 2001).This intervention lasted for two school-years and those schools participating in it used baseline surveys to assess students' views on local priorities.Institutional action teams were then established in each school, comprising a range of staff and students and facilitated by an external "critical friend", to use these local survey data to review and revise existing policies in order to promote a more positive school environment.The project also included professional training for teachers and a new student curriculum to promote social and emotional skills curriculum.Evaluated using a cluster randomised controlled trial (RCT) design and compared to schools that carried on with their standard practice, participating in the Gatehouse Project was found to be associated with consistent reductions in health-risk behaviours, including anti-social behaviour (Bond et al., 2004;Patton et al., 2006).
The Aban Aya youth project also involved both school environment change and a social skills curriculum component, and it was trialled in the Chicago high-school system in the late-1990s (Flay et al., 2004).The intervention was aiming to reduce health-risk behaviours by "rebuilding the village" in disadvantaged schools to enhance students' sense of community and belonging, and increase social support, within these schools.Like the Gatehouse Project, it was strongly informed by attachment theories that postulate that increasing social ties and cultural pride in schools can reduce rates of aggression, substance use and other problem behaviours (Flay and Petraitis, 1994).The intervention involved a standardized process of school change through: convening a local, institutional task-force involving staff, students, parents and local residents to examine and amend school policies relating to young people's health, behaviour and the school ethos; developing new links with community organisations and businesses; and training teachers to develop more interactive and culturally appropriate teaching methods to improve relationships at school.An RCT evaluation found that the intervention significantly reduced violent acts, bullying, truancy and school suspension for boys (Flay et al., 2004).
These studies emphasise the importance of involving young people as part of the HPS interventions advocated by the World Health Organisation (1997).They also suggest specific key steps to support the process of changing the school environment to reduce bullying and aggression.First, a survey of students' to assess their experiences and views as all schools are different and priorities will vary in different school environments.Second, involving some young people in decision-making directly through using an action group comprising both students and staff to review and revise school policies and practices using these survey data.Third, appointing an external facilitator, or "critical friend", to ensure young people's voices are heard on this action group and to assess and monitor progress.Such interventions, which start by trying to understand communities' complex needs and use standardised processes to promote local ownership and universal adaptation, avoid the limitations of one-size-fits-all interventions (Hawe et al., 2004(Hawe et al., , 2009)).However, while the principle of involving young people in decision-making at school has been mainstreamed, there remains little guidance or evidence on what is feasible or acceptable (Children's Society, 2012;Coombes et al., 2013;Fleming, 2013).
The process of planning and delivering complex interventions in secondary schools is always highly uncertain (Bonell et al., 2013).It therefore cannot be assumed that a whole-school restorative approach, using processes for involving young people that have been successful in Australia and the USA, would be feasible or acceptable in British secondary schools which are increasingly focused on educational attainment and inspection frameworks (Fletcher et al., 2010;Bonell et al., 2012).Introducing such a flexible, locally led intervention within British secondary schools may also be inappropriate if the overall logic and "integrity" of the intervention is lost when taken up in some institutional settings (Hawe et al., 2004), which may even lead to more harm than good in some contexts (Fletcher et al., 2014).It has been argued that public health improvement interventions must embrace the functions and processes of "ecological complex systems", and therefore be more "out of control", while retaining their logic and a standardised process of change (Hawe et al., 2004).However, there is little empirical evidence of whether this can be achieved in practice or not to date.
In order to explore this, in 2011 the Health Technology Assessment (HTA) research programme at the National Institute of Health Research (NIHR) funded a pilot trial of a new intervention initiating change locally in bullying and aggression through the school environment (INCLUSIVE).Informed by Markham and Aveyard's (2003) theory of human functioning and school organisation and the Gatehouse and Aban Aya projects, this intervention combines changes to the school environment with the promotion of social and emotional skills and restorative practices through: the formation of a school action group involving students and staff (supported by an external facilitator) to review, determine priorities, and develop and implement an action plan for changing the school environment to improve reduce bullying and aggression; whole-school staff training in restorative practices; and a new social and emotional skills curriculum for year-8 students (age 12-13).The action groups were coordinated by an external facilitator (with experience of secondary school leadership) and they were required to meet at least once every half-term (i.e. six times a year minimum).English secondary schools do not typically survey their students to assess their views on school and/or bullying at present so the collection and feedback of this needs-assessment data is innovative in this context.
The primary foci of this pilot trial were the issues of feasibility and acceptability (Craig et al., 2008).To consider how can we involve young people in changing their school environment to make it safer, in this paper we draw on the qualitative process evaluation data collected in this study to explore: the process of involving a range of students with staff on school action groups; how the action groups use tailored data reports to review local data and develop action plans; and what changes are made to the school environment in order to address bullying and other aggression, including whether this process retained "intervention integrity" (Hawe et al., 2004).

Research methods
A cluster randomised controlled pilot trial was undertaken in eight English secondary schools with allocation to either an intervention initiating change locally in bullying and aggression through the school environment (four schools) or continuation of normal practice (four schools).The intervention was piloted during the 2011-2012 academic year and was intended principally to augment rather than to replace existing activities (e.g.training, curricula, etc.) in intervention schools.However, the intervention process was intended to change existing non-restorative disciplinary school policies and practices where restorative approaches were deemed more appropriate.
Schools eligible to participate were mixed-sex, state secondary-schools in London and south-east England judged by the independent schools inspectorate in England (Ofsted) to be "satisfactory" or better.Independent schools, single-sex schools and schools with "unsatisfactory" Ofsted ratings and/or 6 per cent or fewer students eligible for free school meals (FSM) were not eligible for inclusion in this study.We excluded schools rated as "unsatisfactory" by Ofsted because such schools are subject to special measures interventions and were deemed unlikely to prioritise this process of involving young people at that time.We excluded schools with 6 per cent or fewer students eligible for FSM because these represent the least economically deprived 15 per cent of British schools.
Eight schools were recruited in summer 2011 using sampling and matching criteria to ensure schools varied according to their Ofsted-rating and students' (FSM) eligibility (see Table I).This ensured that the new intervention was being piloted in four diverse school contexts.This paper draws on the qualitative process data collected at the four intervention schools to explore the experiences of students, school staff and facilitators, and how these varied across the different contexts.We undertook semi-structured interviews with school management team (SMT) members, other action group members, and facilitators, focus groups with students and staff and observations of action group meetings.
Ofsted-rating of school effectiveness "Satisfactory" " Good"/"outstanding" At each school, one to two SMT members and a range of student and staff action group members (four to seven per school) were recruited to take part in semi-structured interviews.All the action groups facilitators (n ¼ 3) took part in individual, semi-structured interviews post-intervention.Other students and school staff who were not part of the action groups were also recruited to take part in focus groups to explore a wider range of views.At each school, four groups of year-8 students (aged 12-13) and one group of school staff were recruited to participate in focus group discussions.Students were sampled purposively, and grouped with similar peers, according to their sex and level of school engagement (as reported by staff).The size of these focus groups varied between five and ten students.School staff members were purposively sampled according to their role to include teaching and non-teaching staff (e.g.teaching assistants).The size of staff focus groups varied from four and six participants.Table II provides sample sizes by school and overall.A total of 16 action group meetings were also observed; this ranged from three to five at each school.
Semi-structured interviews with school SMT members took place during the summer term (April-July) on the school site in a private meeting room, lasting 50-70 minutes, and using detailed topic guides.These topic guides included questions and prompts addressing each of the key intervention processes in turn (recruitment to the school action group, use of local data reports and the actions taken to modify the school environment), as well as questions exploring how context might influence the process of implementation (Ozer, 2006).SMT members were interviewed either individually or in pairs.All action group members were also interviewed at the end of the 2011-2012 school-year on the school site in a private meeting room.These interviews lasted 30-60 minutes, drawing on topic guides to structure the interviews.These topic guides were also structured so as to address the key intervention processes in turn but included additional questions about action group members' experiences of meetings, training, and decision making within the group.Each intervention facilitator (n ¼ 3) participated in an in-depth interview (60-90 minutes) in August or September 2012 to explore these topics from their perspective.Focus groups with year-8 students and school staff in intervention schools were undertaken during the summer term on the school site, in private meeting rooms, and lasted 60-80 minutes, facilitated by two trained researchers using a semi-structured topic guide and participatory techniques, such as ranking exercises, to promote discussion among all participants about the intervention process and the school environment.
Interviews and focus groups were transcribed verbatim and entered into NVivo (version 10) software to aid data management and analysis.Records of observations and other field notes and documentary evidence (e.g.minutes from school action group School environment meetings) were also uploaded to support cross-checking and data triangulation.All process data were analysed together.Codes were applied to transcripts, to identify key themes and how these inter-relate in order to develop an analytical framework.Techniques associated with thematic content analysis and grounded theory were used to analyse the data within this framework (Green and Thorogood, 2004), initially using open/in-vivo coding based on the respondents' own words; with memos being used to record emerging themes and make inter-connections across schools and studies.Each transcript was coded to indicate the type of participant, school and date, allowing analytical themes to be explored in relation to different groups' experiences and to compare processes across schools.Further analyses focused on more detailed coding to interpret the meaning of, and relationships between, the initial themes and patterns within and across schools.The findings are presented below (all participants' names and school names are pseudonyms, italics are used to reflect emphasis) and if the participant being quoted was a member of a school action group this is clearly stated.Ethical approval was given by the London School of Hygiene and Tropical Medicine research ethics committee.

Findings
At all four pilot schools an action group was established and met at least six times during the 2011-2012 school year.Each action group included at least six student representatives.Meetings were always after school and typically lasted for about an hour.The establishment and maintenance of these groups appeared to be facilitated by features of the English secondary-school "market" whereby parents have some choice over schools and typically judge them based on their reputation and performance in "league tables".For example, head teachers and their SMT consistently reported that it was important to address aggressive behaviours in order to recruit and retain "the best" parents and students.SMT members also suggested that this project was prioritised as it was seen as likely to impress the national school inspectorate (Oftsed) due to its focus on student voice and behaviour.In order to consider in depth this intervention process for involving young people in changing their school environment to make it safer, the findings are structured below to address the following three sub-questions in turn: is it feasible to recruit students to join school action groups?; how do school action groups involving young people use tailored data reports?; and, how did students change their school environment?
Is it feasible to recruit students to join school action groups?The weakness of existing student voice groups, such as school councils, appeared to provide a strong source of acceptability for students, and in some cases a motivation, for joining these new action groups.At all four schools they were seen as "new" and "different" to school council: I don't see the point in having the school council, because you don't get told anything.
[This is] different […] you knew what's going on.Like, there's meetings often enough and, like, they explain it (Action group student, year-8 female, Goldstone Park).
Like they say, student council, like "oh yeah we do it every week, we've done a few things".And it's like, "what have you done?" (Action group student, year-8 male, Railside High).
The topics are good on the action group [compared to the school council.We actually talk about things that we do here, what's good for the future, stuff like that, like I said in the first place about the CCTV, the cameras and also the teachers is going to be improved and, you know, they listen (Action group student, year-8 male, Whitehorse Road).
Senior managers, including head teachers, also recognised, and were motivated by, the need to increase student voice and their limited representation on existing groups.These action groups were therefore seen as extremely useful in addressing this deficit in students' "perspective": The action group has been really important.It's the first time in the school, as far as I understand, that you've had students working with staffand properly working with staff […] It was very much a done-to climate, the ethos in the school, and staff thinking they knew best (Head teacher, Goldstone Park).
The recruitment and retention of a range of students was consistently seen as central to the success of the action group, which helped ensure diverse representation.School managers described involving a "diverse range" of students and how "positive" this process had been: I think in terms of coming onto the [action group] team, initially it was very positive and we had, in fact we had too many people probably.I got as diverse range as possible, there were quite a few students who were keen.[It wasn't just] goody-goody high-attainers" (Assistant head teacher, action group member, Railside High School).
Different methods were used to recruit students onto school action groups.Students reported that they "got chosen", "got asked" and agreed, or volunteered after finding out about the group in an assembly, newsletter or via word of mouth from a friend.One school used a theatre-based event on bullying to publicise the new project, which was popular with students.It also appeared that using multiple methods of recruitment is most acceptable to students and helps ensure diversity.For instance, at Goldstone Park the deputy head introduced the project in assembly and asked for students to contact him or their form tutors if they were interested but he also pro-actively encouraged some students who may not have volunteered to ensure a mix.A pastoral support manager at the school explained why she thought this was appropriate: It was announced obviously in the assemblies and things like that but […] we do approach students sometimes and say, "look you'd be really good on that" because I think especially with your more colourful students, they don't apply, thinking, "well, I'll never be allowed" [Then you've] just got to sort of have a chat with them and say, "you'd be really good at that, why don't you […]?" -[they say] "we're not doing that" -"no, go and apply!".And they do.And as you can see at those meetings, thoroughly enjoying it […] and you do need them involved (Action group staff member, Goldstone Park).
Students on the action group also reported that being pro-actively encouraged to participate, while still allowing them to volunteer (rather than being coerced), was appropriate: [The deputy head] he asked me [but] people, like, volunteered to do it and they weren't just [making] random people [do it]!They wanted to be there.So, they contribute instead of just sitting there 'cos they don't want to be there (Action group student, year-8 male, Goldstone Park).
Students on the action group at Williamson High School explained that their schools had involved students from different year-groups with different perspectives, which was important: [The action group's] got students from different years, like we're year eight and you've got year tens, and you've got teachers, so you've got like in a way like sort of three points of view from the younger lot and the sort of older lot and the adults, so I think yeah, it's worked well like that (Action group student, year-8 male, Williamson High School).
Students often reported that they wanted even "more involvement" via the school's action group and suggested that meetings should be once a month.
How do school action groups use locally tailored data reports?Students and staff on the action groups consistently reported that their school's student survey data had allowed them to work together to understand the "big picture" and identify priorities for action.Student action group members were supportive of the use of data to assess local needs and identify new priorities for action, which were felt to be representative of a wide range of students' views and this was a "useful" source of motivation to address aggression: It was quite useful to us because the aggression level was quite high so now we're trying to think of ideas to like to get it back down (Action group student, year-8 female, Railside High School).
One way in which they reported that this student survey data helped was by improving their understanding of the prevalence of different aggressive behaviours among male and female students at their schools.It also allowed them to identify potential institutional "problems" that may underlie conflict and aggression (e.g.poor staff-student relationships).In both cases, two key learning processes were evident: validation and discovery.
The process of validation involved the confirmation of pre-existing concerns about bullying and aspects of school-life related to aggressive behaviours.An example of this was at Goldstone Park where the student survey data strongly reinforced their existing concerns about conflict with teachers due to educational disengagement and this provided additional motivation to mobilise the school community to make "positive" changes: Did we believe the results?Yes probably.[The survey found that] they were feeling quite negative about the school and I think that's because they were getting lots of negative messages again, and it was because we were enforcing them: "you are the worst year group in the school, your behaviour is terrible, stop behaving so badly", all from a negative perspective.We're hoping that the second round of the results comes back better and shows improvement, because what we've tried to do, alongside all of this work, is to keep re-emphasising the positive (Head teacher, Goldstone Park).
This also highlights how staff-student action groups not only found these student reports as a spur to change but also planned to use them to assess evidence and champion positive changes.While, not seen as a "new issue", or great "discovery" to senior managers, the presentation of these data on this issue through a tailored report to the school nonetheless provided important impetus for the school to take action and monitor "inconsistency" and improve relationships: They knew it already […] its wasn't anything new as such [but] the data showed this inconsistency issue of children, the pupils, and staff, the data didn't show staff [views] but it turned out behaviour was inconsistently managed and I think that they really ran on that well (Vanessa, intervention manager).This suggests that this process of validation, while providing no "new" insights as such, can itself be an important catalyst to change the school environment.However, the use of these tailored reports did also provide data that were considered "new" and a "surprise".This process of discovery also provided the impetus for schools to adopt new actions to address institutional problems that may be contributing to students' aggressive behaviours.For example, at Railside High School students often reported low aspirations for the future, which informed the action group plan to organise further events aimed at addressing this: [What else came up in your school's tailored report?]It was the issue around aspiration and hope, and their futures.I think that actually was the biggest surprise, because as a school, that's something we do work on all the time so […] that's not good![…] And that's why we planned the horizon day, which was designed around looking at them, making them feel first of all better about themselves, better about school, better about the future, and that's what the whole day was about (Assistant head teacher, action group member, Railside High School).
The external facilitators also reported that this process was highly-appropriate, "powerful" and instructive for school action group's to identify new priorities: I think it was useful, I think it's always good to see data […] and I think without the data they probably would have just completely dismissed [some issues] so at least we did have something there, to say, "Well they said this you know"! (Dawn, action group facilitator).
Staff, students and facilitators suggested surveying as wide a range of student views as possible each year to monitor changes over time.The consideration of data together by students and staffrather than "behind closed doors" in staff-only forumsmay itself also be a catalyst for action, and further dissemination of the data to a wider range of students and their parents may also increase the impact of the reports.The only negative issue which was identified was that some of the SMT members' reported that the needs assessment felt too "negative" at times, especially for school managers who had been in post for several years (and who likely saw this as a reflection on their many years of work and leadership at the school).Intervention facilitators picked up on this and felt that students may also "just see the negatives" if data was not presented appropriately.However, "defensiveness" also prompted action groups to identify and analyse other data sources.For example: [If] schools were very defensive about that kind of reporting on the students and they just saw the negatives, they came up with their own data, some of them had other questionnaires and then surveys which they put into the mix (Bryn, external facilitator).
This highlights the importance of identifying schools' assets as well as their needs via student survey data.However, the ways in which these action groups' "came up with their own data" also draws attention to a further (unintended) benefit of feeding back student survey data to a school action groups and why they were so popular with students: school action plans were not only informed by the (external) student survey data reports but the process of reviewing these data encouraged schools to draw on other sources of local data and present this to their students.There were a wide range of examples of routinely collected data and existing documents that action groups identified, including attendance and exclusion statistics, parent/carer surveys, and incident reports.
The greater use of data also inspired further, larger-scale data collection methods at Goldstone Park where the head teacher commissioned new surveys specifically to monitor changes, as well as to continue to identify challenges and priorities for action: The survey [at the end of the year] that we did with students and parents, came out very positively […] We did it after Easter, so April, May time.We did the whole school.
And we had a huge response from parents […] a third of parents responded (Head teacher, Goldstone Park).
It was clear that all the schools in the pilot were very "data rich" environments and this presentation of new, external student survey data to the action group triggered them to identify, analyse and triangulate multiple sources of data which they would not typically have used.
How did students change their school environment?
The four pilot intervention schools all used the new action group to promote student involvement in decision making with the ultimate aim of promoting a more studentcentred, restorative school environment.Relatively few staff or students reported any concerns that involving students had (or would potentially have) any unintended, harmful consequences.Observations of action group meetings did include some examples of students' suggestions that may have done more harm than good, such as relaxing the requirements and training for peer mentors or changing school food and drink policies, but external facilitators' oversight role meant that they were able to limit any actions that may have done more harm than good in terms of bullying and aggression.
The most significant student-led changes appeared to occur at Goldstone Park where the action group enacted significant school-level policy changes, including the behaviour policy: They changed their behaviour policy, as in their rewards policy was completely overhauled […] and that was as a result of the fact that a lot of, well, most of the pupils in the action group felt unrewarded and didn't value the rewards that they were given, and they certainly didn't perceive them to be motivational.And also, the staff, a real issue of inconsistency came up and a lot of staff felt that people were rewarded inappropriately, so that was why that part of the behaviour policy was altered.So that was the key thing, that was the key policy that was changed (Vanessa, intervention manager).
As well as reviewing and revising school policies, students on the action group at Goldstone Park were also involved in developing innovative new initiatives, such as a new student-led school blog that was being rolled out across the school years: [What have been the main successes for you?]My blog, with Katie, our blog […] I met with the marketing manager that the school just hired and she's pretty keen, so we're going to meet with her sometime and we're just going to have a blog meeting with her, and then she's going to see if she can to advertise it more too (Action group student, year-8 male, Goldstone Park).
Two factors appeared to facilitate student-led changes to the school environment at Goldstone Park.First, students on the action group reported that they were "doing stuff" straight away, which appeared to engage them in the project, stimulate other ideas, and built trust between students and staff on the action group.For example, after the second meeting, when the needs-assessment data were presented, the deputy head on the action group immediately enacted students' suggestions to access a wider range of views and ideas from other year-groups about the school's , "inconsistent" practices and "unfairness" via a suggestion box and students focus groups.Students on the action group reported that they therefore quickly realised this was "different" to school council and a wide-ranging action plan, informed by a range of data sources, was initiated by the group.
The second factor that appeared to be important at Goldstone Park was that the school had a new head teacher, which supported broader, structural changes to school environment to ensure that restorative practices could be implemented across the whole school.For example, the students on the Goldstone Park action group suggested that ten-minute tutor periods were too short to promote a restorative culture and that rushed, unsociable registrations did little to improve relationships between students as it prevented any active "checking in" and "checking out" (or other restorative practices).As he was in his first year in the role and reviewing the whole school day, the head teacher supported the students' suggestion and re-organised the timetable to extend "tutor time" to try and ensure it would have a much greater focus on the social and emotional aspects of students' learning and allow for greater pastoral support: Now tutor time is very short here.Now this is one of the things that's come out of the action group.[…] So it looks like we're going to have a much bigger tutor time where we're going to be able to implement a programme through the year, and I would definitely envisage that SEAL [social and emotional aspects of learning] would be a part of that […] If it's not SEAL itself it would be SEAL related, because we're looking at their emotional development within tutor time and checking in and checking out (Action group staff member, Goldstone Park).
There were also examples of school-level actions at the other pilot schools, including in the most economically disadvantaged schools contexts.At Railside High School, where all the action group students were involved in planning an "away day" for all year 8 students to improve relationships and raise aspirations (a priority identified via their student data).At Whitehorse Road, the students reviewed and the revised existing policies and practices in order help to create a safer educational environment, including through changes to disciplinary policies, pastoral support and more effective use of CCTV.
Relatively few actions were observed or reported that were student-led at Williamson High School and the main barrier appeared to be the lack of a very senior manager (e.g. head or deputy) alongside students on the action group.External facilitators reported that the involvement of a head or deputy head teacher with the "power" to change wider policies was likely to be essential for encouraging some students' participation and enacting student-led actions.This was not the case at Williamson High School, where a relatively new, assistant head teacher represented the SMT and changes to the school policies and ethos were much harder to achieve: [At Williamson] they had an assistant head teacher who was part of the senior management team but what I think it's important [is that] you have somebody onboard who has responsibility for behaviour policy in there […] you really need to have a person who can lead as a power (Bryn, external facilitator).

Discussion
The limitations of existing student voice groups and the mix of methods used to recruit students ensured that a diverse range of young people were included at each school, including more and less academic students.Local data reports helped students work with staff to identify problems and priorities for action at their schools.The presence of students on school action groups effectively makes these data more "public" and this may itself partly help ensure that these data are a catalyst to action for school School environment managers, however, it was apparent that students were also actively involved in discussing the findings and designing actions.Learning from these reports involved both the validation of pre-existing concerns and the discovery of new ones, which also mobilised them to review other sources of routinely available local data.The identification, analysis and triangulation of multiple local data sources was an unintended benefit, although this may be problematic for taking actions where these sources conflict or routine data are used to ignore "outsiders" data (Leicester, 1999).However, there was no evidence that this was the case at these schools.While data is ever only likely to be one factor in any institutional decision making (Nutley et al., 2003), the micro-level processes observed suggest that tailored feedback reports were used as evidence to inform student-led groups actions.
Traditional conceptions of schools as sites for the delivery of standardised health improvement intervention have increasingly been replaced by a recognition that such settings represent a dynamic system that shapes, and is shaped by, those within it (Hawe et al., 2004;Shiell et al., 2008;Naaldenberg et al., 2009).In this way, intervention implementation represents an "event" within an existing, dynamic system (Hawe et al., 2009).Interventions such as this, which use a standardised process (rather than a "one-size-fits all" approach) in order to address local needs, can promote intervention ownership and universal adaptation.However, there is a danger that increasingly nonstandardised HPS interventions may lose their "intervention integrity" when implemented and adapted locally (Hawe et al., 2004).Although only carried out over one school year, this study suggests that locally-adaptable, youth-led interventions retain their integrity when introduced in British secondary schools (Hawe et al., 2004).In this study, the involvement of external facilitators limited the potential for involving young people to lead to too many "out of control" actions and maintained this integrity of intervention logic.This study also suggested the importance of school management cycles for intervention implementation, which has been largely ignored in the implementation literature to date.
In keeping with previous studies of whole-school interventions (e.g.Bond et al., 2004;Bonell et al., 2010), the use of multiple different intervention components functioned synergistically to deliver changes to the school environment.The action group was an innovative and powerful mechanism for supporting student-led change to address key school-level "risk" and "protective" factors for aggressive behaviour.There were no reports of teachers or students feeling uncomfortable about this innovation.The locally tailored reports were an acceptable and powerful external input, which helped action groups identify priorities and should remain integral to this intervention approach, although the process of feeding back data could be improved through: an approach which identifies both the positive and negative features of the school environment, including the "protective" factors for aggression and bullying and school "assets"; and, to maintain momentum, regular surveys of the whole school to monitor change, help celebrate success, and identify new/on-going priorities.Schools should also ensure that either the head teacher or a deputy head teacher is a member of the group to ensure it has sufficient "power" to change school policies.
This approach to involving young people could be integrated within future HPS interventions to support inclusive, student-led, school-level change.A recent systematic review found strong evidence to support the use of HPS interventions as a holistic, settings-based approach to promoting young people's health at a population level (Langford et al., 2014).However, the review also found that few HPS interventions have a clear focus on equity and inclusion, with very little research to date examining their impact and processes according to relevant equity criteria such as socioeconomic status, gender and ethnicity (Langford et al., 2014).
These qualitative data have been valuable but this research is not without its limitations.Data were collected at only four pilot schools and our data may not reflect the experiences and perspectives of all students, even within our sample of schools.However, the large number of young people and staff involved, allied with the purposive sampling approach, helped ensure a diversity of views and settings.It is also worth re-affirming that this pilot trial cannot assess effectiveness or sustainability but can explore the feasibility and acceptability of involving young people in this process across a range of school contexts.Informed by this pilot study and process evaluation, further evaluation has now been commissioned by the NIHR via a cluster RCT involving young people across forty secondary schools in England.