What helps young people at risk of exclusion to remain in high school? Using Q methodology to hear student voices

ABSTRACT

This study explores the views of young people attending within-school ‘alternative’ provision to support their return to mainstream classes. Q-methodology was used to explore pupil views about what helps and inhibits successful on-site ‘reintegration’. Eighteen pupils aged 13–16 years with experience of attending school inclusion centres in three high schools participated. Protective factors supporting reintegration and mitigating the risk of further exclusion from school included the need to be seen and heard, and to have somewhere or someone to go to for support. The study highlights the importance of pupil voice and a need for greater clarity in terminology used to describe the shifting terrain of inclusion. Further research is needed on the context and factors driving the rise in in-school alternative provision in England, including blind spots in official data on the extent of pupil moves and subsequent opportunities within mainstream settings.

KEYWORDS:

• inclusion
• teacher–pupil relations
• reintegration
• Q-methodology

Introduction

Exclusion from school is commonly understood as, ‘a disciplinary measure imposed in reaction to students’ misbehaviour (e.g. violations of school policies or laws) by a responsible authority’ (Valdebenito et al., 2019, p. 254). In England, disciplinary exclusion from school should be treated as ‘a last resort … to ensure that other pupils and teaching staff are protected from disruption and can learn in safe, calm, and supportive environments’ (Department for Education DfE, 2022, p. 3). Pupil moves as a disciplinary sanction include formally recorded fixed-term (temporary) exclusions/suspensions (when pupils are withdrawn from school for a number of hours or days); permanent exclusion/expulsion (when pupils will not return to the excluding school); managed moves i.e. a ‘fresh start’ transfer to another school (without a formal record of exclusion); moves to Alternative Provision (AP), i.e. placements outside mainstream schools (which may include full or part-time placements in Pupil Referral Units (PRUs), further education colleges, independent schools, home tuition services and voluntary or private sector providers); and transitions to Elective Home Education (EHE) (i.e. home schooling) (IntegratEd, 2020).

Escalating rates of formal exclusion, persistent absence, and school changes have deleterious consequences for the wellbeing of young people (YP) and society. Exclusion is associated with poorer educational and social outcomes (Partridge et al., 2020), mental health challenges (Tejerina-Arreal et al., 2020), involvement with law enforcement (Arnez & Rachel Condry, 2021), and reduced employment prospects (Gill et al., 2017). Policy scholars have focused critical attention on ‘perverse incentives’ within the education system that drive exclusion upwards (Thompson et al., 2021, p. 34). The growth in test-based accountability alongside moves towards a more ‘rigorous knowledge-rich, academic curriculum’ (DfE, 2016, p.24) have increased the pressure on school and multi-academy trust (MAT) leaders who are subject to competitive comparison. Exclusion levels peak in the period preceding the national assessment period (Partridge et al., 2020). The ranking of school performance has been associated with an increase in school push-out or ‘off-rolling’ (Done & Andrews, 2020). The school inspectorate defines off-rolling as, ‘The practice of removing a pupil from the school roll without a formal, permanent exclusion […] when the removal is primarily in the interests of the school rather than in the best interests of the pupil (Ofsted, 2019, n. pag.). In England, pupils receiving support for special educational needs and disabilities (SEND), account for 45% of permanent exclusions and 43% of fixed term exclusions (DfE, 2020). Moreover, YP from low-income families, looked after children (with experience of local authority residential care), minoritised ethnic groups (notably Black Caribbean and Gypsy, Roma or Traveller children), and YP with Social, Emotional and Mental Health (SEMH) needs are overrepresented in exclusion data (Thompson et al., 2021; Timpson, 2019). Intersectional analysis reveals additional compounding factors, e.g. the exclusion rate for girls from dual heritage and Black Caribbean backgrounds was three times the rate for white female pupils in 2020/21 (Agenda, 2022). The persistent social gradient recorded in exclusion data undermines espoused commitments to the promotion of inclusive and equitable education for all.

It is likely that official data under-record the extent of school exclusion due to the diverse ways in which pupil mobility between education settings is defined (Hutchinson & Crenna-Jennings, 2019; McCluskey et al., 2019). Pupil moves that fall short of a removal from school are less visible in mandatory records and consequently much less understood. Blind spots in official exclusion statistics include unofficial exclusions where children are removed from school registers while continuing with their education through a range of out-of-school providers. This may include independent and unregistered AP placements (Gill et al., 2017; Partridge et al., 2020), ‘respite’ via temporary alternative provision (Jalali & Morgan, 2018; Malcolm, 2018); or where parents feel they have little choice but to withdraw their child for EHE (Children’s Commissioner, 2019) or ‘consent’ to a managed move to avoid the damaging consequences of expulsion, or prosecution for school non-attendance (Gazeley et al., 2015; Messeter & Anita Soni, 2018). Similarly, pupils retained in school can be excluded from mainstream classes repeatedly, for extended periods of time, by being assigned to ‘alternative’ or ‘therapeutic curricula’ (Power and Taylor, 2020, p.874). In-school AP can range from seclusion to school-based behavioural interventions (Stanforth & Rose, 2020; Valdebenito et al., 2019). Current guidance in England states that, ‘schools and local authorities should work to create environments where school exclusions are not necessary because pupil behaviour does not require it’ (Department for Education DfE, 2022, p. 3).

Aim and research question

This study addresses a gap in the research on reintegration, specifically the views of young people on what supports them to remain in high school. Extant research focuses on reintegration to mainstream school following a placement in Alternative Provision (AP) in another setting (Atkinson & Rowley, 2019; Embieta, 2019). This study responds to the rapid development of in-school AP of varied types in England. It is distinctive in its focus on the perspective of young people who are supported to remain in mainstream school through in-school AP. Specifically, this small-scale exploratory study sought to deepen understanding of how in-school strategies to reduce the risk of exclusion are experienced by Young People (YP). While pupil movement data between settings reveals trends over time, there is little evidence about which approaches are most helpful within schools in supporting young people deemed ‘at risk’ of exclusion from school who are being supported by in-school AP to reintegrate to mainstream classes.

To address the research aims, the study sought a purposive sample of secondary schools that met the following criteria: a record of high and persistent rates of formal exclusion (over three or more years); development of an in-school inclusion facility to support reintegration to mainstream classes; and evidence of a subsequent reduction in exclusion from school. The research took place between May and September 2021 in three high schools in a deprived Northern town in England. The pupil intake characteristics across the schools were similar in terms of eligibility for free school meals (around 40% compared with 19% nationally), attainment (Progress 8 scores, i.e. the progress a child makes between the end of primary and end of secondary school, were among the lowest of all English local authorities) and ethnicity (predominantly white working class). Between 2014 and 2019, the town’s high schools recorded twice the rate of permanent exclusions compared to neighbouring local authorities, and four times the national rate of permanent exclusions in England. The schools received additional funding from the Department for Education and local authority over a two-year period (2019–2021) to develop inclusive practices to reduce pupil moves between settings. As part of an area-based initiative, known as the Continuum of Provision (COP) project, each school introduced an in-school ‘inclusion centre’ with graduated pathways to support reintegration to mainstream classes. When the Continuum of Provision project closed in 2022, each school had achieved a significant reduction in the number of permanent exclusions (less than 0.2 per 100 pupils in the academic year 2021–22), and a concomitant reduction in the number of fixed-term exclusions and number of days lost through fixed-term exclusions (from a peak in 2018). While caution should be exercised when making comparisons between years due to the impact of pandemic-induced school closures from 23 March 2020, in these schools year-on-year increases in exclusion rates fell from 2019. The schools participating in the COP project can be described as critical and homogenous cases that are particularly pertinent to the research aims. Within this carefully situated context, the research addressed the following question: What do young people at risk of exclusion believe supports them to remain in school?

Method

Q-Methodology

This study uses Q-methodology (hereafter Q): a mixed method approach that helps to understand subjective perspectives. As a methodology, Q is ‘inherently a mixture of qualitative and quantitative data and analyses’ (Ramlo, 2022, p. 226). In contrast to Likert scale questionnaires and attitude surveys that employ ‘objective’ measures, Q focuses on the qualitative and subjective (Ramlo, 2016). The reflective and participatory character of Q makes it a suitable approach for research that seeks to promote the inclusion of marginalised voices. It is an ‘ethical, respectful and person-centred’ approach (Hughes, 2016, p.63). Q was chosen for this study as a means of including the perspectives of YP challenged by participation in mainstream classes. Q is an effective means of supporting YP described as having Social, Emotional and Mental Health (SEMH) needs to articulate their needs and preferences (Hinkinbotham and Soni, 2021). As Hellström and Lundberg (2020, p. 419) observe, ‘Asking students to express experiences of emotionally charged situations is particularly challenging’. In contrast to face-to-face interviews, the ‘playful sorting procedure’ of Q is less confronting than direct questions, and less reliant on verbal skills (de Leeuw et al., 2019, p. 325).

Although Q was introduced in 1935 by the psychologist William Stephenson it remains ‘underutilised’ in the field of education (Rodl et al., 2020, p. 1). The application of Q in educational research is only recently emerging. Q has been used to explore diverse topics such as the subjectivity of male primary teachers (Meader & Larwin, 2021), the professional development needs of classroom teachers at historic sites and museums (Cooper et al., 2018), and client needs in accessing school counselling and mental health services (Wester et al., 2021). Q has been used in higher education as an alternative to student satisfaction surveys (Ramlo, 2017), and to explore undergraduate attitudes towards science subjects (Young & Shepardson, 2018), and doctoral candidates’ beliefs about curriculum leadership (Walker et al., 2018). In the field of educational inclusion, Q-methodology has been used to include teacher viewpoints in special education teacher evaluation (Rodl et al., 2020), the support needs of practitioners teaching students with Autistic Spectrum Disorders (Van Der Steen et al., 2020), and the perspectives of classroom teachers on inclusion (Vančíková et al., 2021; Williams-Brown & Hodkinson, 2021). In respect of school students, Q has been used to explore experiences of transition to secondary/high school (Hughes, 2016), adolescent bullying (Hellström & Lundberg, 2020), adolescent self-image (Lim et al., 2022), and YP’s awareness of SEMH classification (Hinkinbotham and Soni, 2021).

In simple terms, Q methodology involves participants (the P-set) ranking statements about a given topic (the Q-set) and placing them in a grid to form a quasi-normal distribution. Correlation and then factor analysis is carried out to group similar sorts together with the goal of revealing distinct viewpoints (McKeown & Thomas, 2013; Ramlo, 2016).

From concourse to Q-set

Concourse development is the first stage in Q-methodology. The concourse should represent ‘all that can be said about the subject matter’ (Lundberg et al., 2020, p. 2), the full range of opinions and views about the topic (Kenward, 2019). This stage is important. If the Q-set is skewed the findings will not be valid. The strategy for concourse development in this study entailed a review of extant literature on reintegration experiences, supplemented by informed professional opinion and consultation with YP with direct and recent experience of exclusion from mainstream classes and subsequent reintegration (sustained over three months). Concourse development typically takes between two and four months (Hensel et al., 2022) and was completed in this study in ten weeks. Twelve interviews were conducted with key informants with a direct role in the design and/or enactment of the Continuum of Provision project (May–July 2021). At MAT-level this included the Chief Executive Officers of the three multi-academy trusts to which the participating schools belonged. School-level inclusion strategies were explored in interviews with headteachers (3), deputy headteachers (3) and inclusion/pastoral leads (2). In addition, contextual information was sought at local authority level (LA) from Council officers (2) and leaders within the LA Pupil Referral Unit (2). An online focus group was convened with three Year 10 and Year 11 pupils (aged 15–16 years) who attended classes in an inclusion unit in one school to gather opinions about strategies they found helpful and unhelpful in supporting their continued education in school. Potential student contributors were discussed with a pastoral support lead in school, who was not present during the focus group to promote openness. While all three participants were located in one school site, they were able to draw on their experience of multiple moves between different education settings in the course of their school careers. Two of the pupils were described as having social and emotional issues and experience of peer bullying; the third was at risk of moving to an off-site Pupil Referral Unit due to conduct issues within the mainstream classroom environment.

Through an iterative process of refinement, 25 statements were included in the final Q-set. The average Q-set size is 40.3, but ranges from 14 to 78 (Lundberg et al., 2020, p. 7). Less relevant, duplicate and marginal items were removed through the review process. Statements needed to be easily accessible to pupils of varied age and levels of reading comprehension. The statements were phrased in naturalistic language and were designed to be short and unambiguous. Large numbers of statements and long sentences may lead to participants becoming bored or losing motivation (McKeown & Thomas, 2013). A behaviour manager in one of the study schools verified that the statements seemed comprehensive and easy to understand. The number of statements was limited to 25 to ensure that the task was not too demanding or overwhelming for participants, and could be completed with the necessary care required within the time available (see Table 3).

Table 1. Pupil demographics for each factor.

	Factor 1(%) N=9	Factor 2(%) N=8
Gender
Male	4 (44)	2 (25)
Female	5 (56)	6 (75)
Year
8	0 (0)	1 (13)
9	1 (11)	5 (63)
10	5 (56)	2 (25)
11	3 (33)	0 (0)
School
1	1 (11)	4 (50)
2	3 (33)	1 (13)
3	5 (56)	3 (38)

Table 2. Eigenvalues and explained variance of both factors.

	Factor 1	Factor 2
Eigenvalues	5.81853	1.56807
% Explained Variance	32	9

Table 3. Consensus statements.

Statement Number	Statement	Factor 1 Z-score	Factor 1 Rank	Factor 2 Z-score	Factor 2 Rank
1	Having a personalised curriculum with reduced hours or a partial timetable	0.33139 *	1	−0.87886 *	−2
2	Having flexibility in the timetable so they can choose which lessons to attend	0.51967	1	0.18376	0
3	Parents being fined for child’s non-attendance at school	−2.32026	−4	−1.85608	−4
4	A car or taxi picking the pupil up to make them go to school	−1.19567	−2	−1.4118	−3
5	Teachers taking pupil issues seriously (e.g. bullying or health/personal issues)	1.41102 *	4	0.03529 *	0
6	Teachers finding out about the pupil and showing an interest	0.07607	0	−0.31113	−1
7	Teachers being aware of pupil needs and what they need more help with	1.12534 *	3	0.5304 *	1
8	Teachers praising positive behaviours	1.07006 *	2	−0.3447 *	−1
9	School staff being trained in the underlying causes of behaviour	0.20528	0	−0.11055	0
10	Teachers staying calm	0.36399	1	0.59971	2
11	Teachers being able to tell by body language that a pupil is too stressed to learn	0.07147	0	0.16583	0
12	Out-of-class passes/time out cards/being able to leave the room	0.18265 *	0	1.97389 *	4
13	A calm and nurturing base in school	0.03325 *	−1	1.49591 *	3
14	A place in school/inclusion unit to feel safe	−0.03099	−1	0.23784	0
15	Having structure and a clear routine	−0.11282	−1	−0.25322	−1
16	Having friends in the school or unit who understand	0.26743	0	0.24601	1
17	Having a key person in school to talk to about anything	1.29839	3	1.38265	2
18	Having a mentor	−0.88675 *	−2	1.39554 *	3
19	Teachers showing pupils respect	0.81992	2	0.69405	2
20	Staff having an awareness of mental health issues	0.77354	1	0.4526	1
21	Home visits from school staff	−1.90287 *	−3	−1.1016 *	−2
22	Meetings with parents	−0.50659 *	−1	−1.65185 *	−3
23	Changing class	−0.91295 *	−2	0.36175 *	1
24	Changing school	−1.64678	−3	−1.26392	−2
25	Pupil making a lot of effort to improve behaviour/attendance	0.96622 *	2	−0.57149 *	−1

*Distinguishing statements. Consensus statements are highlighted in bold.

P-set (participants)

In Q studies, participants are not randomly chosen but should represent a diverse range of viewpoints (Watts & Stenner, 2012). In this study, purposive sampling took place. The study involved three high schools in the same town that were committed to the development of inclusive practice to reduce previous high rates of exclusion. YP who met the following criteria could participate: YP taught in internal units; being re-integrated through graduated responses, or who were in mainstream classes following AP interventions. Some YP had to withdraw due to COVID-19 affecting pupils’ attendance. The participants were chosen by school staff, so the research team was unable to ensure equal numbers by year group or gender. Eighteen pupils from Years 8 to 11, aged 13–16 years were recruited in the three schools. In Q studies, the typical participant sample size is between 12 and 40 (Wester et al., 2021). The average P-set size is just above 37, and ranges from 10 to 90 (Lundberg et al., 2020, p. 7). However, it is important to note that small sample size is not a flaw in Q-methodology. The P-set (number of participants) should not exceed the Q-set (number of statements) (Watts & Stenner, 2012). The aim is to understand the range of possible opinions, not to compare them (Wester et al., 2021).

Ethics

The research complies with the ethical guidelines of the British Educational Research Association (British Educational Research Association, 2018) and the Ethics Guidelines for Internet-mediated Research (British Psychological Society, 2021). The research protocol was approved by the Faculty Research Ethics and Governance Committee, Manchester Metropolitan University. Access permissions were obtained from the three headteachers and Trust CEOs. Time was invested in building relationships with key contacts in each school to establish a shared understanding of the purpose and parameters of the study in accordance with protocols of ethical person-centred research practice. Parents/caregivers were informed about the nature of the study and given the opportunity to withdraw their child from the study. All participants (including the education professionals who participated in interviews supporting concourse development, and the YP who participated in the focus group, Q-sort and post-sort activity) signed a consent form and were fully informed about their right to withdraw from the study at any time without penalty.

The research team adopted a processual approach to ‘ethics-in-action’ (Stokes, 2020, p. 384), including the need for vigilance in assessing YP’s active agreement through each stage of the process. Informed assent (opt-in) was sought from all prospective pupil participants, irrespective of parental consent and gatekeeper nomination. The lead facilitator was a qualified schoolteacher and post-doctoral researcher with a specialism in children’s rights and participatory methods. Pupils meeting the inclusion criteria were fully informed (verbally and via an age appropriate information sheet) about the activity involved in the Q study and how the data generated would be stored and used. Assent was sought on an ongoing basis, based on sensitive attention to verbal and non-verbal cues. For example, power dynamics may mean a YP says ‘yes’ when negotiating initial consent via a gatekeeper. However, as Flewitt (2005) notes, ‘Once initial “provisional” consent has been established, ongoing consent cannot be assumed, but is negotiated in situated contexts on a minute-by-minute basis’ (p.4).

Procedure

Q-methodology web-based software was utilised to carry out the card sort remotely due to constraints imposed by the COVID-19 pandemic (Lutfallah & Buchanan, 2019). Data collection involved three steps. First, each participant sorted the statements electronically by dragging and dropping the statements into three piles labelled ‘unhelpful’, ‘neutral’ and ‘helpful’. Second, each participant placed the statements on a 9-point grid, re-positioning them if required (Figure 1). Third, participants were invited to comment on the reasons behind their choices using an online form (for privacy) and/or in discussion with the facilitator. Q-sort distributions typically range from 7 point (−3/+3) to 12 point; the most common being 9 point (−4/+4) (Hensel et al., 2022, p. 4).

Figure 1. The Q-grid.

Participants completed the Q-sort online on school premises. The software had to be easily manageable for YP with remote support from an online facilitator and appropriate adult present in the room, and fully compliant with school e-safety protocols. No personal log-in details were required or external apps downloaded. In school 1, the researcher introduced the activity to two YP via an online video call using Microsoft Teams (July 2021) and a member of senior staff supervised others on a different day (September 2021). In school 2, the researcher introduced the activity remotely and stayed online for technical assistance and post-sort questions (September 2021). In school 3, the researcher visited the school and supervised the activity in the presence of a teaching assistant (September 2021). It was important to have real-time access to a facilitator who was able to provide clear instructions and repeat ethical protocols, e.g. reassurance about confidentiality, anonymity in reporting and the right to withdraw.

Care was taken not to introduce bias in the sorting process through the instructions offered to participants or the proximity of others in the room. Sorting was undertaken independently with support available for use of technology. Participants were advised that no view is less valid than another; no perspective is right or wrong. A trusted adult with experience of working closely with the YP was present in the room during data collection. This is also a safeguarding requirement when conducting online activity with YP in schools. The sorting process helped participants to make their points of view visible (Ramlo, 2016). Each statement is considered in relation to the others, rather than in isolation. Accordingly, ‘the sorting process is reflective and self-referent’ (Ramlo, 2022, p. 230). As participants must rank the statements against each other on a curve, there is reduced risk of YP offering the perspective they anticipate adults want to hear.

Data analysis

In Q, by-person factor analysis is used to identify groups of participants who ranked items similarly, producing a weighted ranking for each group. As Alkhateeb and Romanowski (2021, p. 4) explain, the letter Q distinguishes, ‘by-person factor analysis from Spearman’s by-variable factor analysis, which is usually conducted with R methods, i.e. surveys and questionnaires.’ In other words, Q uses ‘inverted’ factor analysis, i.e. ‘inverted because each participant (or their whole Q-sort) is treated as a variable, unlike factor analysis of surveys, where the items are variables’ (Churruca et al., 2021, p. 2). The factors (or grouping of individuals) represent shared perspectives based on the rankings.

Findings

In this study, two factors (distinct shared viewpoints) were identified as the most coherent and comprehensive using a process of abduction from the statistical outputs. A general rule is to extract one factor for every six statements in the study (Watts & Stenner, 2012). Initial factor analysis revealed three factors with eigenvalues above one (a measure of significance). However, one of these represented the views of just one participant. Extracting two factors instead led to a more balanced distribution with nine participants loading highly for Factor 1 and eight for Factor 2 (Table 1). Centroid factor analysis was chosen and the two factors rotated by the software via the varimax method. The correlation between the two factors is 0.5, which is considered to be moderate.

Both factors met Humphrey’s rule, i.e. factors should be retained if the cross-product of the two highest factor loadings exceeds twice the standard error. Factor 1 accounted for 32% of the variance and Factor 2 accounted for 9%, a total of 41% (Table 2). Only one participant did not load for either factor.

Out of the 25 statements, 14 consensus statements were identified which implies an overlap in viewpoints between the two factors (Table 3).

Factor 1: being seen and heard

Factor 1 was a viewpoint held by five female pupils and four males. All of the female pupils were in Year 10 (age 14–15) at the same school so had likely encountered similar types of provision. They were all transitioning back into mainstream classes after a period of support in an in-school inclusion unit. As shown in Table 3, the 11 distinguishing statements for Factor 1 cover a variety of pastoral interventions. The three most highly ranked statements for Factor 1 are listed in Table 4.

Table 4. Most highly ranked statements for Factor 1.

No.	Statement
5	Teachers taking pupil issues seriously (e.g. bullying or health/personal issues)
17	Having a key person in school to talk to about anything
7	Teachers being aware of pupil needs and what they need more help with

These statements suggest that pupils loading highly on Factor 1 valued personalised support and attention. The focus of these statements is being seen and heard rather than having to cope alone or being subject to generic interventions. Analysis of the qualitative data provided at the end of the Q-sort further illuminates the factor. For example:

I think that teachers taking things seriously would help students a lot so they feel like they can talk to a teacher without being worried that they aren’t gonna be listened to and judged.

Because it would be good to get respect from teachers

(‘Teachers showing pupils respect’ ranked highest)

In answer to the final written question ‘is there anything else you would like to add about what helped or did not help in your case?’ pupils loading highly on Factor 1 added:

Talking about my issues and having time out of class when I needed it helped me a lot.

Teachers helped me with my problems and understood what I needed and gave me things that would help.

Out-of-class passes really helped me out because it allowed me to have a break from everything going on around me when it got too much.

The final statement was written by a female pupil who had placed other statements (5, 20, 1) as more helpful on the grid but felt she wished to mention out-of-class passes as being useful. This demonstrates the flexibility of Q methodology in that it enables participants to express themselves freely despite the constraints of the Q-sorting process.

A minor discrepancy is the ranking of ‘having a mentor’. For pupils with the viewpoint represented by Factor 1, having a mentor is deemed to be somewhat unhelpful (−2). At first glance, this seems to contradict their high placement of ‘having a key person in school to talk to about anything’ (+3). A suggested reason for this is that not all schools use the word ‘mentor’ to describe the staff member (including teaching assistants) allocated to support pupils and who works with them most closely on a day-to-day basis. In some schools, a ‘mentor’ may be an external volunteer recruited to offer targeted and episodic support in other areas of activity, e.g. career aspirations or literacy support, or the term might not be familiar to YP at all. Schools vary in formal and informal mentoring arrangements for a range of purposes.

Factor 2: having a safe place or person to go to

Factor 2 seems to be a viewpoint favoured more by younger pupils with around three quarters of them being in Year 9 (aged 13–14 years) or below.

The most highly ranked statements for Factor 2 (Table 5) were more difficult to categorise but seem to suggest that pupils valued having somewhere or somebody to go to outside the mainstream classroom. The out-of-class passes may give the pupil a sense of agency by enabling them to leave a stressful environment if they feel they cannot cope, without incurring sanctions.

Table 5. Most highly ranked statements for Factor 2.

No.	Statement
12	Out of class passes/time out cards/being able to leave the room
13	A calm and nurturing base in school
18	Having a mentor

So then if they’re in school and having a bad day they can go somewhere to calm down.

Having a calm and nurturing base in school and/or a mentor may evoke a sense of security among YP deemed at risk of exclusion. The statement below suggests that the pupils value self-determination and a sense of personal agency. When involved in making decisions that affect them, such as the choice of ‘mentor’, YP expressed a preference to have somebody they already know and whose support is accessible to them at the point of need.

If you need someone to talk to at school then you should be able to have someone to talk to and you should be able to choose the person.

Fifty per cent of the pupils with this viewpoint attended school one, where being allocated a mentor was an important part of the school’s inclusion strategy. One pupil from this school described how their mentor acted as a mediator supporting a progressive planned reintegration into mainstream classes.

Having a mentor is great because my mentor helps me get through lessons.

What do pupils perceive to be unhelpful?

The statements pupils ranked most negatively were very similar for both factors, so are presented together. Although different interventions are valued by different pupils, YP generally agree on what they feel does not help. The lowest ranking statement for both factors was ‘parents being fined for child’s non-attendance at school’.

Parents shouldn’t be fined for us being late because it’s not them that’s late

(sic)

Some people like to have mental health days.

Home visits from school staff and changing school had the same rankings for both factors. Pupils who placed home visits from school staff as the least helpful statement suggested:

I feel like home visits wouldn’t help much because if anything is going on with family or home situations its (sic) most likely to be hidden when the teachers arrive.

At home they will probably feel safe and having a teacher at home would probably make them stressed.

Two of the focus group pupils spoke passionately about home visits:

I had someone who came into my room and basically threatened that if I didn’t come in my parents would get like a fine and go to jail. And that just made my problems ten times worse.

They are invading your personal space. It’s your safe space.

Arguably, the statements placed at the most unhelpful end of the grid do not meet the pupils’ needs of being seen and heard (Factor 1), or having a safe place or person to go to (Factor 2). School staff visiting the home, sending a vehicle to collect them or fining their parents/carers for ‘allowing’ them to be absent may exacerbate issues. Active and intensive outreach can be experienced as punitive surveillance that may inhibit the development of trusting relations and school connectedness.

Discussion and conclusion

This study used Q-methodology to help YP with varied experience of in-school support, designed to reduce the risk of exclusion, to share their views on what helps to keep them in school. Following a ten-week preparatory period for concourse development, a Q-set of 25 statements was generated. Eighteen YP aged 13–16 years in three high schools that can be considered critical cases participated in the Q-sort and post-sort activity. Factor analysis supported the identification of protective factors that YP felt supported successful re-integration to mainstream classes. In this final section, we acknowledge some limitations, return to the literature to establish the contribution to knowledge and consider implications for future research, policy and practice.

The logistics of the study were influenced by the constraints to education of the COVID-19 pandemic. The participants were chosen by school professionals, using explicit criteria, who acted as gatekeeper in regard to access and safeguarding (Kay, 2019). The participation of YP and school professionals may have been influenced by the largely remote mode of engagement. Remote interviews during concourse development were unavoidable due to social distancing mandates. However, there is some evidence to suggest that the physical distance and sense of privacy engendered in remote interviews compares favourably with in-person interviews in terms of disclosure and rapport (Lobe et al., 2020). Where possible, post-sort questions were completed immediately after the Q-sort. However, only a minority of participants provided detailed additional qualitative data (i.e. more than 3–4 sentences). Some YP may have been insecure in communicating their views on this topic in the presence of an authorised adult in a school setting, or unused to articulating their preferences, or lacked confidence in their narrative capability. Despite careful attention to ‘ethics-in-action’ (Stokes, 2020), the limited post-sort contributions may indicate that some YP acquiesced rather than gave affirmative assent. Logistical constraints meant that there was no opportunity to return to participants to involve them in the interpretation of the factors (Lundberg et al., 2020). The study would be strengthened by enlarging the participation of YP across the stages, especially in co-creating the Q-set. Additional insights would be generated by replicating the study with a wider participant pool, and refining the concourse to include primary age pupils. In addition, further research would benefit from inclusion of the parent/caregiver perspective. General criticisms of Q in regard to validity have been discussed by Brown et al. (2015), and defended on the grounds of the misapplication of quantitative tenets to a method designed to elicit participants’ emic perspectives. Statistical generalisability is not a concern of Q methodology, where ‘substantive inference’ is more appropriate (Alanazi et al., 2021, p. 2).

This small-scale exploratory study adds to the body of work advocating a rights-based approach to pupil participation in research that informs school policy and practice (United Nations, 1989). The opportunity for young people to voice their views and concerns is a fundamental tenet of inclusion. As de Leeuw et al. (2019, p. 235) argue, ‘The inclusion of the voices of children is not only rights based, but a necessity in realising an inclusive educational system’. The Special Educational Needs and Disability Code of Practice (Department for Education DfE, 2015, p. 19) emphasises, ‘the importance of the child or young person, and the child’s parents, participating as fully as possible in decisions’ and a need for ‘greater choice and control for young people and parents over support’. In contrast to earlier studies that report education practitioners’ perspectives on the factors that facilitate or inhibit successful reintegration from AP (Thomas, 2015), this study attends closely to the views of young people. Other studies have explored pupil perspectives on reintegration between settings following permanent exclusion (Atkinson & Rowley, 2019; Jalali & Morgan, 2018; Owen et al., 2021) or combine the experiences of YP receiving support via in-school Learning Support Units and Pupil Referral Units (Pillay et al., 2013).

The differences in ranking recorded by participants in this study signals that ‘one-size fits all’ approaches will not address diverse individual needs. This confirms previous findings by de Leeuw et al. (2019, 2018) that appropriate approaches will be contingent on the situational context and child-related context. As recent guidance from the Department for Education (DfE, 2022, p. 3) implies, imposition of a ‘no exclusion’ policy per se will not address the complex needs that give rise to young people’s emotional and behavioural disengagement from school (process) and a school’s decision to exclude (outcome).

The findings of this study lend further support to an established body of research that emphasises the importance of dialogue and positive relationships in supporting integration. Positive teacher–pupil relationships are a protective factor that helps YP to cope in mainstream classes, especially those with identified SEMH needs (Sheffield & Morgan, 2017). A systematic review by Messeter and Anita Soni (2018, p.180) of factors that support reintegration following managed moves highlights the imperative of ‘positive relationships … open lines of communication and a pastoral and personalised support plan’. It is unsurprising that the connectedness and ‘affective engagement’ of YP influences behavioural engagement and outcomes (Pinzone & Reschly, 2021). Caring social interactions with teachers and peers fosters a sense of belonging among YP with SEMH needs and those receiving additional support for learning (Bouchard & Berg, 2017; Dimitrellou and Hurry, 2019). Young people who do not form connections with school, peers or teachers are more likely to experience persistent serious disengagement from school (Hancock and Zubrick, 2015).

The findings presented here highlight the value of non-deficit approaches. In this study the school that had made the most progress in reducing exclusions from school between 2018 and 2021 adopted a school-wide approach to inclusion, at the centre of which was a planned transition from a confrontational student-deficit approach towards a social model of inclusion. All three schools had undertaken sustained development to integrate (previously separate) approaches to behaviour, additional support and pupil wellbeing. Research suggests that re-engagement is more likely via positive behaviour support than ‘zero tolerance’ approaches. In their review of research on support for secondary pupils with SEMH needs, Carroll and Hurry (2018, p. 319) found that approaches that ‘embrace techniques which encourage pupils to feel secure and that foster good relations with teachers, result in pupils who were more motivated to learn and are therefore at less risk of exclusion’. Similarly, a research review by Owen et al. (2021, p. 332) found post-reintegration regression was associated with punitive approaches, staffing instability or lack of training, and relational difficulties with peers and teachers.

There is clear value in actively involving YP, and their parents and caregivers – in addition to school professionals and administrators – in research on strategies to promote inclusion. A better understanding of experiences of reintegration is imperative in appraising strategies to promote inclusivity within mainstream schools. In-school AP that is premised on personalised pastoral support can minimise the disruption to learning and social relationships that are an outcome of off-site fixed-term or permanent exclusion. Graduated responses to reintegration according to individual need may reduce the risk of reintegration breaking down. However, intensive support for culture change requires resource. The schools in this study received £40k per annum for two years to strengthen systems for the identification of unmet need and timely early intervention. Drawing on an ecological approach to system change, the schools undertook whole-school staff development in targeted areas such as restorative practice, trauma-informed practice and de-escalation strategies, and protected resources to support the redeployment of trained staff (e.g. for bespoke personalised support) and repurposing areas of the school estate (e.g. for nurture rooms, alternative curricula).

Through the application of Q-methodology, this study has shown the importance of ‘being seen and heard’ and ‘having a safe person or place to go to’ in helping YP at risk of exclusion to remain in school. The involvement of secondary schools that represent critical cases in their commitment to turn around embedded cultures of exclusion adds credibility to the findings. Awareness of the situatedness of the study also aids appreciation of the infrastructural, material and cultural resources deployed to promote inclusion (Atkinson & Rowley, 2019). It is important to acknowledge that truly inclusive settings support the inclusion of young people with diverse needs in the same classroom, not simply within the organisational boundaries of the same institution. In-school AP is not intended as isolation from mainstream opportunities. The schools in this study sought a graduated full return to mainstream education, where appropriate. Further research is needed on the context driving the rise in in-school alternative provision in England and the purposes and character of such provision. Although currently an underused approach to educational evaluation, Q-methodology can make a valuable contribution to formative mixed-method evaluations of interventions to support reintegration (Churruca et al., 2021; Harris et al., 2019). Future evaluations need to address the blind spots in exclusion data pertaining to inclusion centres in mainstream schools. This will entail putting together the pieces that make up pupil experience of staggered reintegration including the frequency, mode, duration of interventions, the intervention- and child-context, and pupil trajectories following reintegration. The inclusion of pupil views is of fundamental importance in community deliberation on the (re-)design and monitoring of ‘alternative’ learning environments within mainstream settings.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The authors confirm that the data supporting the findings of this study are available within the article.