‘Off-rolling’ and Foucault’s art of visibility/ invisibility: An exploratory study of senior leaders’ views of ‘strategic’ school exclusion in southwest England

This article reports an unprecedented exploratory small-scale investigation of the views of senior school leaders in southwest England relating to ‘off-rolling’ (illegal exclusionary practices). ‘Off-rolling’ is conceptualised as a policy technology, however, the conceptual framework used in data analysis derives from Foucault’s treatment of power and pleasure, and the constitution in discourse of novel ‘permanent realities’. Relations of force that senior school leaders routinely navigate in neoliberal education cultures characterised by ‘performativity’ are highlighted. The latter describes pressures to evidence efficiency and continual improvement, including high-stakes performance monitoring and, in England, a national school inspection regime. The response rate to this online and predominantly qualitative survey was poor, which was attributed to the illegality of ‘off-rolling’ and the risk that academic research in this area could be perceived to reproduce the power – knowledge relations associated with a national school inspection regime and familiar neoliberal political or professional discourses around training of education professionals (where both are framed as addressing deficits to achieve improvements). Nevertheless, findings from a key question inviting comment on scenarios drawn from the authors’ teaching experience are presented here, as they highlight the issues arising in research of sensitive topics. Reliance on brief vignettes to explore levels of understanding was, arguably, a limitation of the study, risking participants interpreting this device as unwelcome external scrutiny. It is argued that Ofsted’s definition of ‘off-rolling’ as gaming (manipulation of academic performance data) effectively discourages recognition of exclusionary practices that are not irrefutably related to academic performance as such.


Introduction
The descriptor 'inclusive education' now features prominently in education policy discourse globally (Slee, 2011), despite considerable variation in how it is applied within specific national contexts. In Italy, for example, formal school exclusion is legally prohibited, whilst in England, the USA and Australia, it signals the legal entitlement of those with disabilities and additional educational needs to participate in a A noteworthy feature of Ofsted's attention to 'off-rolling' is the organisation's simultaneous insistence that the prevalence of such practices is minimal, a suggestion echoed by Timpson (DfE, 2019b), where 'off-rolling' is deemed a rare occurrence involving very few head teachers. Accordingly, Ofsted's recent policy of 'naming and shaming' schools where 'off-rolling' is identified has, to date, generated only several confirmed cases in inspection reports (Ofsted, 2019c). Following detailed analysis of the National Pupil Database, Machin and Sandi (2020) challenge the now taken-forgranted assumption that exclusions in Year 11 (when national GCSE examinations are taken) imply strategic gaming. The validity of this argument hinges on the timing of exclusions within Year 11, since pupils excluded after the January school census may, in fact, sit their examinations and there is no evidence of exclusions being frontloaded in this final year of compulsory education. For Machin and Sandi (2020), defining 'off-rolling' as gaming obscures the possibility that exclusion could occur much earlier in pupils' school careers, as schools strive to manipulate the pool of 27). This possibility is supported by anecdotal data in non-ministerial government reports; hence, the Office of the School's Adjudicator (OSA, 2017) reports third-party accounts of 'coerced home education' across year groups which suggest that exertion of pressure on parents by schools to remove their children is widespread or more common than supposed. In Machin and Sandi (2020), higher rates of legal exclusion were found to be associated with the pre-2010 academisation of poorly performing urban schools, such that any gains in academic performance were linked to the addressing of disciplinary issues. The later mass academisation did not, Machin and Sandi (2020) found, lead to increased formal exclusion rates because the schools in question were predominantly already higher-performing schools without significant behaviour or discipline issues. Machin and Sandi's (2020) research highlights a longstanding orthodoxy in mainstream education, one enshrined in legislation relating to legal exclusion and that Ofsted's definition of 'off-rolling' as gaming fails to challenge. Armstrong (2018, p. 997) describes this orthodoxy as the 'wicked problem of behaviour in schools' and, following Power and Taylor (2020, p. 878), it involves the removal of pupils from classrooms or schools when their behaviour is perceived to interfere with 'regular' schooling, regardless of legal requirements to include. Underfunding of the inclusion agenda (Parveen, 2019) reinforces this orthodoxy, but it is questionable whether significantly increased funding would displace such a deeply ingrained feature of English educational culture. Ironically, Ofsted (2019a) now criticises the very culture that it has been instrumental in creating, arguing that exam performance is over-emphasised in England's schools to the detriment of inclusion.

Researching 'off-rolling'
The tensions evidenced in public and political debates around 'off-rolling' imply a complex field that warrants research, particularly since the viewpoints of senior school leaders themselves are routinely omitted. This absence prompted the study reported below, which was initially conceived as a stand-alone project designed to provide a space in which senior school leaders could articulate their own perspectives on 'offrolling' and to explore the complexity that schools and their leaders must now Views of 'strategic' school exclusion in SW England 1041 navigate. An online survey was opened in April 2019 via Jisc online surveys. It had been noted that the sample of an Ofsted-commissioned survey (YouGov, 2019) on the same topic had contained only one head teacher; however, it was assumed that calls for 'naming and shaming' schools suspected of 'off-rolling' (Laws, 2019), combined with the fabrication of 'off-rolling' head teachers as gamers (Ofsted, 2019a), had deterred head teachers from participating. When it became apparent that head teachers in southwest England were similarly reluctant to participate in our research, a booklet was distributed to all secondary schools in the region underlining our attention to ethical research practice and intention to provide an opportunity for senior school leaders to express their views. The latter was premised on a rejection of a unitary concept of 'voice' (Mayes, 2019). There were only 21 responses to the senior leader survey, all from urban secondary settings, with a small proportion that had training in SEN provision coordination (SENCos). Consequently, the research then evolved into a multi-stranded project, including data collection from parents (Done et al., 2021) and SENCos and the analysis of social media posts from education professionals.

Senior leader survey
Nevertheless, it was decided to discuss our senior leader study here despite a disappointingly small sample (n = 21) in illustration of the issues which can arise when investigating highly sensitive topics where participation is likely to be perceived by potential participants as highly risky. Whereas parents readily volunteer emotive accounts of suspected discriminatory exclusionary practices and 'off-rolling' affecting children with SEND (as confirmed through legal contestation of school actions), data from this senior school leader survey primarily concerned the interpretation of guidance and the potentially profound negative effects on parents and children were largely unacknowledged or rendered invisible. This omission can be attributed to the questionnaire design and its dual objective of assessing levels of understanding of current guidance on 'off-rolling' whilst simultaneously inviting varied and authentic viewpoints. Even so, some responses were surprising on account of what was omitted rather than their explicit content. It was this opacity, in the context of 'masking' (the rendering of exclusionary practices less visible or invisible) (Power & Taylor, 2020) that prompted a conceptual analytical framework derived from Foucault (1977Foucault ( , 1978Foucault ( , 2003 and texts less usually associated with educational research. This framework, ultimately, serves to underline the affective dimension of all research analyses, from the comforting familiarity of analytical statistics to the shocks to thought provoked by theoretically informed poststructuralist analyses (Deleuze, 2004). Following Rolfe (2006, p. 304), it should not be assumed that qualitative studies are informed by a unified onto-epistemic paradigm; in this instance, data collection was by survey and frequency counts of similar responses are made transparent. The identification and abstraction of overarching themes has been avoided, however, as the limited number of responses raised questions around inference and validity, where the latter is conceived as transferability (Lincoln & Guba, 1985). Instead, we interrogate responses through a poststructuralist lens that is made explicit and that we judged to be warranted by the data set. 1042 E. J. Done and H. Knowler The emphasis in the conceptual framework outlined below is less on what was shown (or not) than on 'how things were made visible, how things were given to be seen [. . .] to knowledge or to power' (Rajchman, 1988, p. 91). The associated analytical strategy resonates with Ball's (2003, p. 221) concerns around authenticity within neoliberal 'impression management' cultures, where how information is presented matters. It cannot, of course, be known whether questionnaire responses were indicative of an authentic lack of awareness or misunderstanding of the law and guidance on 'off-rolling', or whether they imply the management of a 'dangerous truth' (Rajchman, 1988, p. 106) that such practices are routine. Comments on the questionnaire itself cut through the assumed separation of researcher and researched found in traditionally more positivistic or descriptive qualitative research.

Conceptual-analytical framework
The concept of 'dangerous' truths (Rajchman, 1988, p. 106) derives from Foucault's (1978) account of the fabrication of sexuality as an object of medicalised knowledge in the late nineteenth century. Medical models of deviancy and the concept of the unconscious created a 'space' of constructed visibility to illuminate the relationship between what is said and what is seen (Foucault, 1978). The enthusiastic investigation of, for example, hysteria (as a novel object of knowledge) also made this 'space' an incitement to see (Rajchman, 1988). The 'machinery of power that focused on this whole alien strain [of aberrations] did not aim to suppress it, but rather to give it an analytical, visible, and permanent reality' (Foucault, 1978, p. 44). It is tempting to identify this same process in Ofsted's efforts to identify, scrutinise, quantify and police a practice and object of knowledge ('off-rolling' as gaming) that Ofsted itself is largely responsible for fabricating (Done & Knowler, 2020b). The fabrication of 'offrolling' as a hidden practice that violates the rules of the market order (free and fair competition) is a similar incitement to see as Ofsted sets about its task of investigating disparities between what it sees and what is said in England's schools.
Determining prevalence is key in the fabrication process within medical science (Hacking, 2006), but also in education where it informs and legitimises novel political technologies (Popkewitz, 2013). A further resonance is that Ofsted's search for the deviant head teacher individualises a problem that is, in our analysis, systemic in nature. Foucault (1978, p. 44) avoids individualising the apparent pleasure or sense of anticipation produced by the investigation of hysteria as a hitherto unseen condition; instead, such pleasure is deemed a historical property of the 'space' in which psychiatry and psychoanalysis arose and a function of the wider socio-political context. A 'polyhedron' of intelligibility (Foucault, 2003, p. 249) was later mobilised as a heuristic device entailing the identification and infinite multiplication of explanatory factors that included nascent epistemological orthodoxies. In relation to 'off-rolling', a 'culture of performativity' (Ball, 2003) characterises the prevailing socio-political context, where it appears that improvement in one remit of a head teacher (academic attainment) may only be achieved at the expense of another (inclusion) (Done & Knowler, 2020b).
The art of visibility/invisibility relies on 'self-evidences' (Popkewitz, 2013) as taken-for-granted assumptions based on what is visible. Such self-evidences imply a Views of 'strategic' school exclusion in SW England 1043 strategy of power that renders specific practices tolerable to those participating in them (Rajchman, 1988, p. 94). Policymakers and bodies like Ofsted have endeavoured to frame the compatibility of the 'standards' and inclusion agendas as self-evidently achievable, yet as Power and Taylor (2020) argue, the proliferation of legal exclusionary practices would not be happening if schools were finding such compatibility unproblematic, an argument which might equally be applied to 'off-rolling'. Without radical prioritisation of the inclusion agenda, the danger is that the 'wicked problem of behaviour' (Armstrong, 2018) will be shoehorned into a familiar neoliberal narrative of amelioration or improvement (the identification of a 'gap' in training to be addressed within a target population). The government's response to Timpson (DfE, 2019a,b) illustrates this possibility, since it proposes that school staffs are educated on links between trauma, attachment styles, specific diagnosable conditions and poor or disruptive behaviours. This focus on the behavioural manifestations of some types of experiences, parent-child relationships and SEND will, again, individualise a systemic issue, with training being viewed as ineffective in some cases or some SEND being perceived as particularly resistant to behavioural management techniques.
The relation between what is thought, seen, said and done in Foucault (1978) differs from that of phenomenology (as a form of empiricism) (Moran, 2002); in Foucault, what is presented to consciousness is conditioned by the socio-political discursive context. The incitement to see is an 'incitement to discourse' (Foucault, 1978, p. 17). Discourses are 'blocks' operating within a field of force relations that continually shifts (pp. 101-102); hence, the possibility of alternative ways of thinking and seeing, that is, of resisting dominant self-evidences and practices as an ethical endeavour. Rajchman (1988, p. 108) cites Foucault at interview on these 'ethico-political choices': 'It is, in fact, because I thought I recognized something cracked, dully jarring or dysfunctioning in the things I saw, in the institutions with which I dealt [. . .] that I undertook a particular piece of work' (Foucault, 1981). The ethical dimension of this art of seeing is also temporal; being future-orientated it precludes return to bygone practices (e.g. rigid segregation as the solution in the problematisation of educating those with disabilities or SEND); similarly, an innocent return to a disciplinarian educational culture is not possible. The emphasis on economic efficiency in neoliberal 'cultures of performativity' (Ball, 2003) has superseded earlier socially prevalent moral concerns with the dangerous social and individual consequences of idleness; hence, the rationale for the goal of full inclusion as a global project was economic expediency (UNESCO, 1994). Interestingly, Power and Taylor (2020) invoke this concern with economic efficiency when arguing that the proliferation of legal exclusionary practices in schools is not cost-effective as the additional costs incurred are not underwritten by government.
Noting the gender imbalance in the teaching profession and the displacement of professional discourses of caring by an impersonal instrumentalism, Ball (2003, p. 227) argues that feminist theories of gaze are relevant to analyses of performativity, given its reliance on external scrutiny of performance. The ethics of care proposed by Tronto (2013) is similarly pertinent; it avoids biological reductionism and Tronto's (2013) concept of 'privileged irresponsibility' seems particularly apt when considering developments described by Power and Taylor (2020). The gender composition of 1044 E. J. Done and H. Knowler support staff within school-based ameliorative programmes (e.g. 'nurture' groups) is likely to evidence a similar imbalance. The concept of 'privileged irresponsibility' conveys the role of successive governments in the inclusion agenda in England; schools and teachers are charged with delivering inclusive education whilst government funding varies according to the wider economic context and governmental priorities (Done et al., 2015). The art of visibility/invisibility as the political management of what is shown and not shown is evidenced in the 'responsibilisation' (Foucault, 1982) of head teachers for reducing exclusion and eliminating 'off-rolling' through training in trauma, attachment and behavioural techniques (DfE, 2019a).
The tension between the 'standards' and inclusion agendas (where the former is prioritised and the latter underfunded) effectively incentivises head teachers to 'offroll'. Meanwhile, the historical and enduring political tendency to blame the teaching profession for failures of policy (Thrupp, 1998), combined with the illegal nature of 'off-rolling' and Ofsted's novel practice of 'naming and shaming' offending schools, will discourage senior school leaders from rendering such practices visible to researchers. It is conceivable that the findings later outlined will be interpreted as illuminating a requirement for training; however, they may equally be read as evidencing a 'cynical compliance' (Ball, 2003, p. 224) since the culture of performativity ensures familiarity with the formulaic narratives of remediation that routinely circulate within it.

Wavelengths or facets
The objective of the multi-stranded research project within which the senior school leader survey is now situated is not to produce a definitive overview of 'off-rolling' but, rather, to underline the complexity of the field and generate insights that might otherwise not be gained. Inclusion is an emotive topic which has generated a political rhetoric designed to discipline schools and manage parental expectations without undermining an established political discourse around academic standards. This objective precludes any presupposition that the 'voice' of stakeholder groups is unified or can readily be identified. Gaining insight into the divergent viewpoints in these groups is integral to poststructuralist analysis of relations of force that expand or limit agentic possibilities. The characterisation of this multi-stranded and evolving research project as deploying a 'wavelength methodology' (Done et al., 2021) is apt since it is the amplitude of wavelengths which determines volume and, accordingly, only purposeful listening will ensure the audibility of some wavelengths. Similarly, the electromagnetic spectrum contains visible light but also invisible components such as radio waves and microwaves; the latter may not be visible but can be detected and their effects made visible. In Foucault (1978, p. 101), 'silence and secrecy' can be a 'shelter' for power. Ball's (2003) concept of cynical compliance implies the careful management of what is made visible in high-stakes accountability procedures.
Like 'facet methodology' (Mason, 2011, p. 77), 'wavelength methodology' aims to produce insights (such that maximising data through large-scale descriptive research is unnecessary) and achieve 'different ways of seeing'. Different viewpoints within Views of 'strategic' school exclusion in SW England 1045 and between strands are foregrounded without the synthesis of findings across data sets or integration of findings into a unified whole (although connections between data sets could legitimately be drawn). Strands focusing, for example, on the views of senior school leaders and the parents of 'off-rolled' children do not imply a conventional mixed-methods strategy, as corroborative or validatory logics are replaced with a multi-dimensional logic that assumes, rather than denies, social complexity (Mason, 2006). This resonates with the anti-totalising and non-unifying import of poststructuralist thought and its relational ontology without precluding the useful interrogation of large data sets, as in Machin and Sandi (2020). Arguably, the 'gemstone' metaphor mobilised in 'facet methodology' (Mason, 2011) lacks the flexibility of Foucault's (2003, p. 249) 'polyhedron' of intelligibility (with its infinitely extending sides in the multiplication of explanatory factors). Foucault (1978) conceptualises the relationship between discourse, knowledge and power as 'a multiplicity of discursive elements that can come into play in various strategies' (p. 100), and resists providing methodological principles in favour of 'cautionary prescriptions' (pp. 92-102). The prescription relating to the 'tactical polyvalence of discourses' (p. 100) requires the researcher to 'reconstruct' the distribution of 'things said and those concealed, the enunciations required and those forbidden'. This includes 'the variants and different effects-according to who is speaking, [their] position of power, the institutional context in which [they happen] to be situated'. It also entails analysis of the 'shifts and reutilizations of identical formulas for contrary objectives' (p. 100); hence, familiarity with the law relating to illegal exclusion was invoked for very different purposes in survey responses, with very different 'effects', depending on the research strand in question. Each strand can be conceived as a 'local centre of power-knowledge' in accordance with a 'rule of immanence' (p. 98), which dictates exploration of the 'relations of force' (p. 99) at play between, in this instance, the state and schools. Power here is understood as relational and not as a property or zero-sum concept (Foucault, 1978, p. 92). Senior school leaders must contend with Ofsted's very 'specific extortion of truth' (p. 97) while, simultaneously, constituting a 'prop' or anchor point in an 'overall' governmental strategy of population control (p. 99).

Senior leader questionnaire
Despite the small sample, the senior leader study contained sufficient open-ended questions to generate rich data relating to understanding of the law and statutory guidance on 'off-rolling'. The central question contained several vignettes outlining scenarios which all qualify as 'off-rolling' under existing legislation; these were derived from dialogue with SENCos in our teaching practice on inclusion-related postgraduate programmes and intended to illustrate 'off-rolling' as a broader set of practices than Ofsted's (2019a) gaming definition. Participants were invited to identify each scenario as either 'off-rolling' or otherwise and comment on their choice. These scenarios are reproduced below.
Vignette 1. Mark (Y9) is frequently late to school and removed from lessons due to disruptive behaviour. He spends a lot of time in isolation from other pupils and has had 3 fixed term exclusions totalling 40 days in the last 9 months. The SENCo says 1046 E. J. Done and H. Knowler there is a place available at a nearby pupil referral unit (PRU) and arranges a meeting with Mark's parents to discuss moving Mark to this provision for Year 10.
Vignette 2. Julie (Y8) has an Education Care and Health Plan (ECHP) for Autism and Global Delay. Her parents expressed a strong desire for her to attend her local mainstream secondary school. Julie's teachers are worried that she is falling behind with her class-based assessments and she is performing well below her peers, even with Teaching Assistant support. The Head and SENCo arranged to meet with Julie's parents to discuss a move to a special school.
Vignette 3. Michael (Y10) has recently come to the attention of Social Services because his parents have separated and neighbours have expressed concerns about his father and alcohol abuse. This has triggered safeguarding concerns for the school and a feeling that Michael would receive more support in a different setting away from his family.
Vignette 4. Lesley is a 'looked after' pupil who experienced childhood sexual abuse up to the age of 12. She has been moved out of the county she was born in and was subject to a managed move from her last school placement. Teachers have serious concerns about her mental health and wellbeing and the implications for her school performance. The school feels that an alternative provision placement would be more appropriate for Lesley.
Vignette 5. Lee's (Y7) behaviour has become increasingly disruptive since his transition into secondary school. He had been supported by a teaching assistant at his primary school but this support is no longer available as he does not have an EHCP, nor is he on the school's register for SEND. The Head Teacher arranges a meeting with Lee's mother to discuss whether a formal school environment is the right one for her son and asks her to think about working with Lee at home. Lee's mother thinks this could be an option as she is constantly receiving phone calls about Lee's behaviour in school and it is causing her stress

Findings
Of the five vignettes reproduced above, only the final scenario (V.5) was almost unanimously identified as 'off-rolling' (20 of 21, 95.2%) whilst, of the remaining scenarios, affirmative responses ranged from 3 of 21 (14.3%) to 6 of 21 (28.6%) for V.2 and V.4, respectively. The vignette responses are therefore reported in reverse order and any variation in sample size is due to participants failing to provide comment.

Vignette 5
Comments included expressions of sympathy for the child's mother (4 of 16): Poor Lee's mum does not stand a chance if she does not know much about schooling.

Views of 'strategic' school exclusion in SW England 1047
The failure of the school to provide support prior to raising home education with the parent was deemed unlawful pressure (7 of 16) or premature (6 of 16); 3 of 16 participants provided detailed outlines of what the school should be doing to ensure compliance: Yes as the school is not fulfilling their obligations within the Children and Families Act 2014 and SEN Code of Practice. They should be demonstrating a graduated response rather than seeing the child as a 'problem to offload'. [School] should look to get EHCP [education, health and care plan] and other options. Mum is under stress and there is no provision suggested that could help Lee to meet his needs.
The motivation of the school's head teacher was explicitly mentioned by 5 of 16 as inappropriately exclusionary: The school is looking to take Lee out of education.
School is trying to remove the child having put no support in place. If he isn't even on the SEND register then they have done nothing to support him. This is an example of off-rolling as the school don't seem to want to deal with him.

Vignette 4
Notably, while 9 of 21 (42.9%) participants did not view this scenario as 'off-rolling', 6 of 21 (28.6%) claimed not to know and a similar proportion identified it as 'offrolling' (6 of 21, 28.6%). There were declarations of sympathy for the child, combined with questioning of earlier school decision-making (5 of 17): A managed move was the wrong action to take. Lesley needs support, not judgement. I would say it is an example of systemic failure rather than just this school's failure. Consideration of the right provision of the child's needs should have been made in the first place. This approach seems reactionary rather than thoughtful and needs orientated.
Only 2 of 17 explicitly recognised that concern around the child's academic performance may imply an underlying concern with school performance data: Based on the information provided, concerns about Lesley's school performance based on her mental health are not acceptable reasons for placing her in alternative provision.
The school's attention to the child's welfare was questioned by one participant: That entirely depends on whether the school is disingenuous in their concern over her welfare in their setting.
Another suggested that schools may lack control over decision-making where multiple agencies are involved: Not if multi-agency involvement and general agreement this was best for her. Imagine this would be a short-term placement with a clear plan to return to current school after this intervention. Can she be dual-registered to remain on school's roll? 1048 E. J. Done and H. Knowler Other participants (3 of 17) also raised dual-registration as determining whether 'off-rolling' had occurred: If the school are requesting that Lesley be moved because of her mental health and wellbeing and have evidence of her needs, I wouldn't consider this as off-rolling. However, Lesley is a LAC [looked after child] and as such is bound to have significant emotional needs and would probably benefit from stability rather than moving. If she remained dualregistered an appropriate therapeutic placement may benefit her but if not this is offrolling.

Vignette 3
12 of 21 (57.1%) participants did not recognise this scenario as 'off-rolling' and those who claimed not to know (6 of 21, 28.6%) outweighed the remainder that identified 'off-rolling'. Some pointed to a lack of clarity as to whether the child was to be moved from both home and school, while others deemed the scenario unlikely (8 of 18): Not sure what you mean here by a different setting. Do you mean attending another school? It would not seem appropriate to move a student based on these concerns and it would seem better to provide some stability in this situation. I cannot offer a view as to whether this is off-rolling or not.
Probably not. If a multi-agency approach is used and agreed then this would be the best approach for Michael. I would not imagine that the school would be driving the discussion about whether another school would be better.
Notably, one participant rejected the possibility of 'off-rolling' as there was no mention of the child's academic performance: No I don't believe it is. The move, should it be considered, might be in the best interests of Michael. The scenario focuses on Michael's emotional state and not his academic one, hence the school [is] not off-rolling.
As previously, concern was expressed about the child's need for stability (3 of 18): With all that change Michael's school might be the only consistent thing in his life. Have they considered this?

Vignette 2
Responses here were unequivocal; none were neutral and 18 of 21 (85.7%) did not consider 'off-rolling' had occurred. There were, however, intimations of strategies that might be deployed in order to circumvent the legal right of a parent to select a mainstream setting for their child. Several participants insisted a move was in the child's 'best interests' (9 of 17). Others emphasised that only dialogue with parents had been initiated at this point (5 of 17). The latter seems reasonable but, as some participants acknowledged, 'discussion' can develop into pressure placed on parents to affirm the school's position:

Views of 'strategic' school exclusion in SW England 1049
No. This is only at the discussion stage and no action has been taken. Hopefully, the school will not be making any threats in the meeting such as to exclude Julie.
No, as long as it is a genuine discussion and review. It does sound like jumping the gun to go straight to discussing specialist provision. What does EP think? There may be a need to open a discussion though if everyone ultimately agrees mainstream is inappropriate then specialist setting could be needed.
School may better meet her needs with smaller class sizes and specific interventions and strategies. However, main concern seems to be her assessments and performance. Parents' wishes not taken into consideration and the words 'discuss a move' imply the decision is already made, so pressure could be applied and felt.
The issue of the funding of complex needs was introduced (2 of 17) as a rationale for the school's initiation of dialogue with the parents: This is a difficult debate as the parent and child's views should come first but balances with funding constraints also have to be made, unfortunately, which effect, at times, the quality of provision in mainstream versus the provision in specialist school.
Again it depends. Who is paying, where do results sit? She is Y8 so long time to go! The co-positioning of academic performance and funding above implies possible trade-offs between financial pressures and contribution to school performance data. As another participant suggested: If Julie was performing well with the current school they would not consider moving her.

Vignette 1
This scenario invited some neutral responses that were attributed to lack of detail (3 of 21, 14.3%); 14 of 21 (66.7%) responded negatively, while some (4 of 21, 19%) deemed 'off-rolling' to have occurred. Dual-registration (at school and the PRU) was held to invalidate charges of 'off-rolling' (5 of 18), and there were suggestions that the child's interests were being prioritised: No, because as yet there is no final decision taken. This is the beginning of a conversation to help Mark manage his behaviour.
Assuming the isolations are for issues which cause disruption or endanger/persecute fellow students. If those things are the case, no, this is not off-rolling. This is exploring a potentially better setting without insisting the child moves there.
The school is seeking appropriate alternative provision for the student. Regardless of the attendance and exclusions issues, this may be the right setting for the student in the future.
The avoidance of perceptions of pressure also appeared in response to this scenario (5 of 18): PRU may meet his needs better, this is just a preliminary discussion and no pressure. 1050 E. J. Done and H. Knowler I'm not sure as Mark is close to reaching the threshold to be permanently excluded so this could be proactive support to allow him to stay in education? Depends what the school's motives are perhaps and also if this is done with consultation with parents and Mark. Placement in a PRU was viewed as a relatively benign form of 'off-rolling' but also, potentially, as the beginning phase of a process: The response from the school is punitive through socially excluding the child within isolation which eventually leads to encouraging a move to another school. However, I recognise that there may be more 'severe' and negligent forms of off-rolling in other schools such as not allowing them on the school site with no alternative provision. I don't think this is off-rolling as Mark has a place at a nearby PRU. The SENCo is also having a discussion with Mark's parents to discuss the move. The move is being suggested rather than a permanent exclusion with nowhere to go.
Two participants suggested that different strategies were more appropriate: Should be a team around child approach involving Inclusion Officer; need to look into SEN needs; on what basis is SENCO deciding this is best? Sounds like a dramatic and rushed response to still minor/common behaviour. What do parents and Mark want and need?
It would be possible for a period of support from the PRU before Mark returns to his mainstream school for Years 10 & 11. In that case it wouldn't be off-rolling, but if it was just done to get rid of Mark then it would be.
Only one participant rejected the suggestion that the move to a PRU was in the child's interests: He has needs which need assessing. Moving him to another placement will not help Mark.
The avoidance of permanent exclusion made PRU attendance a preferable option for 5 (of 18) participants.

Discussion
The near unanimous identification of a single vignette (V.5) as 'off-rolling' suggests that Ofsted's definition of 'off-rolling' as the gaming of academic performance monitoring procedures has acquired a hegemonic status which obscures or, indeed, facilitates non-recognition of other types of 'off-rolling' as such. Alternatively, a 'local cynicism of power' (Foucault, 1978, p. 95) implies tactics that are 'often quite explicit at the restricted level where they are inscribed'. Strict adherence to Ofsted's ( , 2019a linking of 'off-rolling' to academic performance data effectively enables senior school leaders who are so inclined to engage in illegal exclusionary practices. Repeated reference to dialogue with parents suggests that senior leaders are aware of charges that schools can pressurise parents into agreeing a course of action that benefits the school not the child in question, and that parental consensus can determine whether 'off-rolling' has occurred. Hence, managed moves, as a legal mechanism for transferring a child to another school, require the agreement of all parties, Views of 'strategic' school exclusion in SW England 1051 while home education to which a parent consents is deemed 'elective' (OSA, 2017). The suggestion of one participant that a school may be 'disingenuous' (V.4) in its concern for the child's welfare (given her poor academic performance) confirms this possibility of 'cynical impression management' (Ball, 2003, p. 221) by school leaders. In YouGov's (2019) survey of teachers, verbatims refer to routine exaggeration of accounts of poor or disruptive behaviour to justify permanent exclusion (legal grounds for such exclusion under current legislation).
There was some evidence of power relations as an 'interplay of powers and pleasures' (Foucault, 1978, p. 46): 'The power that lets itself be invaded by the pleasure it is pursuing; and opposite it, power asserting itself in the pleasure of showing off, scandalizing, or resisting' (p. 45). The first could be exemplified by participants who were clearly keen to demonstrate their extensive knowledge of the law and/or statutory guidance relating to 'off-rolling'. The remark 'it depends. Who is paying, where do results sit? She is Y8 [Year 8] so long time to go!' can be read in at least two ways, that is, as indicative of a perverse pleasure in scandalising through the implicit unveiling of a professional 'secret' not usually seen beyond the profession. It is an example of data which arrests, creates intrigue and demands attention, given its obvious ambiguity and the tantalising possibility that the researcher is accessing an otherwise hidden 'truth' (Foucault, 1978, p. 97). A more reasonable interpretation is that the participant is emphasising the need to avoid over-hasty decision-making, but the ambiguity remains, conveying the pleasure of evasion as a form of resistance (p. 45).
Similarly, the phrase 'offload' (V.5), although used by a single participant, can also be read as the glimpse of an otherwise concealed reality, that schools have ways of evading their legal responsibilities and remove children from classrooms or the school roll. The recurring mention of dual-registration (V.1, V.4) suggests that senior school leaders are aware of how accusations of 'off-rolling' can be avoided.
A small number of perverse responses indicating that 'off-rolling' had not yet occurred as the child was still in school were, perhaps, evidence of resistance to researcher interest in the topic. The use of vignettes to ascertain levels of awareness and understanding of legal and statutory guidance could easily have been interpreted as reproducing the 'extorted admissions' of Ofsted or the prelude to a discourse of amelioration through training. Foucault notes the 'pleasure that comes of exercising a power that questions, monitors, watches, spies, searches out, palpates, brings to light; and on the other hand, the pleasure that kindles at having to evade this power, flee from it, fool it, or travesty it' (Foucault, 1978, p. 45). This statement conveys a potential aspect of the power relations between schools and Ofsted, but also between academic researchers and senior school leaders.

Limitations
A much higher response rate was anticipated that would, for example, have permitted some grouping of data by participant type and contextualisation of responses. Details of the study were circulated through a BERA blog site and a university database of regional schools. The poor response was, we would argue, indicative of the difficulties that can be encountered in much survey-based research around highly sensitive (and 1052 E. J. Done and H. Knowler politicised) topics and planned depth face-to-face interviewing could not be actioned as only one head teacher volunteered to be interviewed. An alternative method of sampling and data collection will be adopted in future studies to increase the likelihood of participation.
The choice of conceptual framework reflects Deleuze's (2004) insistence that thinking beyond orthodoxies or normative assumptions requires a degree of provocation, and the resultant deductive analytical strategy, combined with staying close to the data, afforded a transparency that is not always evidenced in, for example, the move from manifest and latent meaning in thematic qualitative research (Bengtsson, 2016).

Conclusions
Data from an, albeit small, sample of senior school leaders serves to challenge a familiar neoliberal discourse in which policy objectives are deemed achievable through refinements of existing policy and further training of school leaders and staffs. The 'discourses of truth' (Foucault, 1978, p. 97) generated by Ofsted ( , 2019a individualise a systemic failure to address tensions between policy agendas (raising educational standards and inclusion) and such tension was apparent in senior leader responses. Prior to the Covid-19 pandemic, a 'discursive ferment' (Foucault, 1978, p. 17) around 'off-rolling' was evidenced in England (Done & Knowler, 2020a,b) which would, undoubtedly, have continued had governmental and media priorities not shifted. Scrutiny of legal and illegal exclusionary practices (DfE, 2019b;Ofsted, 2019c) culminated in the latter's 'name and shame' policy and negative media coverage. Fabrication of deviant professional identities, however, does little to address the socio-political prioritisation that facilitates exclusionary practices in schools, and it has been suggested here that a radical shift in socio-political priorities is required if legal and illegal exclusionary practices are to be minimised. Whilst we recognise the pressures that schools are under to deliver academic progress in a 'culture of performativity' (Ball, 2003), our analytical strategy of staying close to the data was not solely dictated by sample size but also reflected an interest in conceptual frameworks that might capture the experiences of senior school leaders in neoliberal educational systems. Foucault's (1978) theorising around an 'art' of visibility/invisibility (as the management of what is revealed) can be usefully mobilised in research into sensitive topics, particularly where deviation from normative discourses carries significant risks for participants or is perceived to do so. The objective of the reported study was not generalisability or, as in qualitative research, transferability but, rather, following Mason (2011), insight generation which will influence future research.