‘Failures’ in a failing education system: comparing structural and institutional risk factors to early leaving in England and Portugal

ABSTRACT In this paper, we draw on a conceptual framework of five key categories of risk to Early Leaving (EL) to analyse and compare ‘Structural Factors’ (SF) and ‘Institutional Factors’ (IF) of risk in a region in England and in Portugal. We draw on data from an EU-funded project, which explored risks and support strategies to EL across five European countries. Our two-country comparison involved analysis of 59 semi-structured interviews and focus groups with 209 educational stakeholders across 18 educational settings. We discuss key similarities and differences across the data, reflecting on plausible factors these are attributed to. SF was the most discussed risk category in England but not in Portugal, albeit largely for the same reason: that they are difficult to address, relating to rigid and entrenched educational systems and policies, socio-economic disadvantage and inequalities. Educators were cognisant of how SF hinders educational institutions and stakeholders from addressing EL, given the lack of freedom and flexibility they engender. Young people focused more on how the IF of schooling are impacting EL. We explore the implications of our findings for EL risk and offer recommendations that will be of interest to a range of educational stakeholders, including policymakers, educators and employers.


Introduction
Early Leaving (EL) from education and training is defined by the EU as referring to young people aged 18-24 who may have left school without completing compulsory schooling or may have completed lower secondary education or pre-vocational/vocational courses but not gained formal upper secondary qualifications or equivalent (Eurostat 2020;European Commission 2011). Despite wide use of the term EL and 'Early Leavers' by and within the EU, the term is not officially recognised within the UK, where the term 'NEET' is employed to refer to a young person 'not in education, employment, or training' (Mascherini 2018). Seen as more reflective of the vulnerable status of young people in society and the challenges faced in participating in education/training as well as the labour market, the term has become increasingly utilised within several economically advanced countries, particularly as an indicator to monitor and understand factors leading to 'NEEThood' (Eurofound 2020).
In this paper, we draw on analysis of 59 interviews and focus groups with 209 educational stakeholders across 18 educational settings to compare the latter two key risk categories to EL in England and Portugal: Structural Factors and Institutional Features of School. We focus on these two risk categories in England and Portugal in order to explore how aspects of each country's education system, policies and schooling impact on EL risk.
There are some notable similarities that make these countries worthy of comparison when examining EL risk. In terms of SF, both countries have similar EL rates, just short of the EU's −10% target and are still attempting to reduce these (Eurostat 2020). They also operate very centralised education systems with standardised and largely academic curricula and have raised the age of participation in education and training to 18 (DfE, 2012;. In both countries, vocational pathways have been traditionally less valued (Chankseliani, Relly, and Laczik 2016;. We consider whether a 'one size fits all' centralised approach to learning may be detrimental to young people at risk of EL who are often blamed and shamed as 'problems' to be 'treated' so that they conform within the existing education system (Ross and Leathwood 2013, 328). Education in both countries has also suffered from a lack of funding (Bousted 2021 cited in Adams 2020; European Commission 2020, 49; OECD 2020), which is particularly problematic for young people from socio-economically disadvantaged backgrounds who are more likely to be at risk of EL.
SF interacts with and impacts on IF, for example, in the various barriers they impose on institutions and educators to address EL. Centralisation, for example, requires teachers to adhere to prescribed curricula that may not be aligned with students' interests, learning needs or aspirations. Research shows that the school environment and infrastructure play a powerful role in facilitating student engagement (Nikolaou, Papa, and Gogou 2018) and that inappropriate teaching and assessment strategies, resources, and available learning support greatly impact on student's selfesteem, contributing in many cases to school dropout (Van der Graaf et al. 2019; Lavrentsova and Valkov 2019). An overly centralised education system may also impact on students' sense of a lack of agency as they lack choice and flexibility to direct their own learning paths, which could lead those at risk of EL to become disengaged and potentially drop out.
Our findings will show that this theory (of a centralised system leading to a lack of agency and a greater risk of dropping out) applies both in England and Portugal, as two countries with otherwise different educational traditions, and in very different local contexts within these countries. We thus use the comparison of cases to argue that the theory is potentially valid across a wide variety of cases, which is a rationale of comparison that Skocpol and Somers (1980, 176) have dubbed a 'parallel demonstration of theory'. Differences between cases are then primarily contextual particularities that highlight the generality of the processes with which our theory is concerned (ibid: 178).
Within the paper, we first consider aspects of the education systems and policies in England and Portugal relevant to EL risk. We then outline key aspects of the research methodology, including the settings and stakeholders that were involved in each country's research sample. We then present the research findings, noting key issues arising within the SF and IF risk categories in each country. In our ensuing discussion, we reflect on the major similarities and differences between the two countries and the likely features these are ascribed to, noting observations regarding the convergence and divergence between our stakeholders' views. Our conclusion provides an overview of the study and unearths significant issues regarding how structural and institutional features of education and training are embedded in political, economic, social and cultural norms that set up young, disadvantaged and vulnerable people as perceived 'failures' in a failing education system. We explore the implications of our findings and provide recommendations for tackling EL that will be of particular interest to a range of educational stakeholders, including policymakers, educators and employers.

Education and NEEThood in England
In England, there have been few government policy initiatives to address NEEThood, and the majority of interventions have involved short-term charity-funded projects. This structural and cyclical dimension to tackling NEEThood and a lack of ownership are significant barriers to sustained progress (Britton et al. 2011, 4). One legislative action taken to reduce NEEThood in 2013 was raising young peoples' participation in full-time education or training from the age of 16 to 18 (Decreto-Lei 139/2012, de 5 de Julho Department for Education 2012). This includes working or volunteering fulltime while studying or training part-time to gain qualifications; or taking an apprenticeship, traineeship or supported internship (DfE, 2016). The government also pledged to expand the apprenticeship scheme but it has been criticised as 'failing the young' owing to being 'flawed from the start' and severely underfunded with 'low quality and patchy regulation' (The Guardian 2020).
Ensuring young people remain in compulsory education or training from the age of 4/5 until aged 18 is the responsibility of local government, which must provide adequate education and training provision; support 13-19-year olds and those aged 20-25 with Special Educational Needs and Disabilities (SEND) to participate in education or training; and track and report on participation and NEET statistics to target those at risk for support (Brown, Douthwaite et al., 2021;Brown, Olmos et al., 2021). The September Guarantee, which was implemented nationally in 2007, is a guarantee of an offer, made by the end of September, of a suitable place in post-16 education or training for young people completing compulsory education (Parliament. House of Commons, 2010). It requires local authorities to find and report data on suitable education and training places specifically for 16-17-year olds (DfE, 2014). At this age, students are at a key transition stage as they would have sat General Certificate of Secondary Education (GCSE) exams and be moving on to further study or training. This often requires moving to a different institution given that only some schools provide education to age 18, normally through academic A levels for university entrance. Despite these government initiatives, they have not resulted in reducing EL (Brown, Douthwaite et al., 2021;Brown, Olmos et al., 2021) and the education system has been criticised for contributing to the issue.
Historically, British politicians have pushed young people to pursue a traditional, academic, knowledge-based education culminating in the implementation of a standardised national curriculum in 1990 that is centralised, prescriptive, heavily performance-oriented and competitive (Ball 1993). In critiquing the neoliberal ideology that underpins the education system, Priestland (2013) notes that 'British school students are the most tested in the industrialised world, and league tables force teachers to "teach to the test", demoralising the profession and demotivating students'. This has been exacerbated by recent curriculum reforms and revised accountability and performance measures including GCSE courses being redesigned to be more challenging and pushing increasing numbers of students to take the English Baccalaureate (EBacc). This requires students to study up to eight core academic subjects at GCSE level including English, maths, the sciences, a humanity (history or geography) and a language (DfE, 2017) resulting in less students studying creative subjects (Maguire et al. 2019). Students are now also required to gain a grade 5 or above in English and Maths GCSE. These accountability measures have been criticised for 'narrowing the curriculum', with the EBacc seen as 'an unhelpful restriction on school autonomy and another attempt to drive behaviour through exam reforms and league tables' (NAHT 2018).
Within British society, vocational routes have been viewed as low-status with few opportunities for progression, often attracting 'academically weak', low-income and disadvantaged students (Chankseliani, Relly, and Laczik 2016). In a recent attempt to raise the status of vocational education and training (VET), in 2016, the government pledged to enable young people aged 16 to choose either a traditional academic pathway or a technical pathway (streamlined into 15 routes) that would lead to achieving nationally recognised, high-quality qualifications and work experience. In terms of the technical pathway, there are two key options: 1) enrol on an apprenticeship programme 2) take a 2-year college-based programme that involves work experience (DfBIS/Department for Education 2016). The latter option can lead to the new T Level technical qualification, launched in September 2020, that is equivalent to three A levels. T Levels were developed with employers and businesses to ensure that they equip students with the knowledge and skills required for the labour market and/or enable them to pursue further training or study (DfE, 2020). Degree-level apprenticeships have also recently been introduced, which should 'provide flexibility to move between an academic and vocational pathway' (Brown, Douthwaite et al., 2021;Brown, Olmos et al., 2021). Despite these new VET routes, there are questions around their appropriateness for young people at risk of EL, many of whom have a history of low attainment (Battin-Pearson et al. 2000). Furthermore, these VET routes start from the age of 16, with younger students still generally required to complete a rigorous academic curriculum.
As a final point, it is important to note that austerity measures brought in as a result of the 2008 economic crisis have led to significant underfunding of state schools for over a decade, particularly those in deprived areas. Mary Bousted, joint general secretary of the national education union, has highlighted that this is a 'historic failure [for] the nation's children [. . .especially] children from the poorest backgrounds' who 'only get one chance to go to school' (Adams, 2020). These cuts have led to a lack of resources including key educational professionals that are particularly important in supporting vulnerable children at risk of disengaging.

Education and early leaving in Portugal
Education in Portugal has focused on implementing strategies around three central objectives: social inclusion; quality assurance; and preparing young people for the labour market (Magalhães, Araújo, Macedo and Rocha, 2015). The extension of compulsory education until the age of 18 in 2009 (Lei nº 85/2009) is viewed as an important factor in the democratisation of access to higher levels of education through provision of longer school pathways leading to higher qualifications for all. However, as noted by Mangas, Lopes, Milhano and Freire (2020) there is still a significant percentage of students who do not complete the 12 compulsory years of schooling, questioning the effectiveness of this measure in facilitating more equitable outcomes and social inclusiveness.
The education system is divided into pre-primary education (ages 3 to 5), basic education (typical ages 6 to 14), secondary education (typical ages), post-secondary non-higher education and higher education. At the age of 15 (the end of basic education), students choose to pursue either a general academic education that prepares them for further study within higher education or a vocational/ professional education that combines general, technical and work placement training in order to prepare them for working life (OECD, 2012aEurydice 2021). In total there are seven uppersecondary pathways available through a range of educational institutions: 1) science-humanities courses geared towards further study in higher education. These cover four areas (science and technologies, social and economic sciences, languages and humanities, visual arts); 2) technological courses aiming to prepare students to either enter the labour market or pursue further study, for example, through post-secondary technological specialisation courses and higher education courses; 3) specialised, non-dual vocational courses; 4) specialised artistic courses; 5) education and training courses specifically for adults who did not finish compulsory schooling; 6) vocational courses designed to provide students with an initial qualification to enable them to enter the labour market while, at the same time, allowing them to study further; and 7) apprenticeship courses (OECD 2012b(OECD , 14, 2020. Within pathways five and six, Education and Training Courses/Professional Courses (ETC/PC) can be taken, which are an opportunity to complete compulsory schooling through a flexible path adjusted to each students' interests. These can involve either pursuing studies or becoming qualified for the labour market.
All the education and training tracks lead to a school-leaving certificate, and, in theory, allow transition across pathways and into tertiary education (via national exams). However, the OECD (2020) policy outlook report found that 'the organisation of upper-secondary schooling may hinder such movement' (ibid: 11). Further research is needed to establish why movement across pathways may be problematic. Furthermore, VET was found to 'be less valued than general education' (ibid: 5).
The skills mismatch in the labour force and long-term unemployment were also emphasised along with the obstacle of a digital skills deficit and recommendations to VET providers and employers 'to better align provision with labour needs' and to improve 'information about labour market prospects, returns to education and educational tracks' (ibid: 8).
The education system is highly centralised, with 'little autonomy at local or institutional level, particularly in teacher allocation' (OECD 2020, 7). In terms of organisation and funding, securing long-term sustainability of funding for education has been a critical issue. Despite growing investment in education, it is 'still far from the levels seen before the [2008] economic crisis' and there is a need for 'large investments [. . .] in the near future to upgrade and renovate the existing school infrastructure' (European Commission 2020, 49). There have been calls for structural reforms (D'Angelo and Kaye 2018) and for funding to be better aligned with Portugal's key priorities (i.e. social inclusion; quality assurance; and the preparation of young people for the labour market). Furthermore, at the tertiary level, there have been suggestions that the student population could be more diverse, with persisting enrolment inequities across generations (OECD 2020, 5). Recommendations for additional institutional and funding support for disadvantaged students at all levels have been made in order to increase existing educational and social support to promote social inclusion and prevent EL (OECD 2020).
Portugal has had ongoing equity concerns with students from socio-economically disadvantaged backgrounds more likely to repeat grades, leave school early and perform below their more advantaged peers. In 2014, the OECD (2014) noted that EL in Portugal is a particular issue compared with other OECD countries, particularly in the last years of compulsory schooling. Since then, significant progress has been made in EL rates, which has dropped to 10.6% as of 2019 compared with 17.4% in 2014 (and 20.5% in 2012), now being much closer to the EU's 10% target (Pordata , n. d.). This impressive reduction is likely to be ascribed to the numerous regional and national initiatives implemented within Portugal's education system, which are in line with EU guidelines and targets to reduce EL (see European Commission 2020 for a comprehensive list of these). There has also been an increase in basic skills proficiency and tertiary education attainment (OECD 2020).
Despite this progress, there remain a number of challenges in education and training including the need to tackle EL. Grade repetition rates remain incredibly high (OECD 2020) with Portugal being one of the three European countries with the highest retention levels. In Portugal, grade repetition is more than an administrative measure to be 'exceptionally' applied and is ingrained in school culture. Grade repetition obliges students to remain in and repeat the same year of schooling, for reasons of failure or for having exceeded the limit of unjustified absences (Decreto-Lei 139/2012, de 5 de Julho). Research suggests that retention is an ineffective intervention in the context of poor performance (Jimerson, Anderson, and Whipple 2002;Xia and Kirby 2009). It greatly impacts on student's self-esteem, contributing in many cases to school dropout (Van der Graaf et al. 2019), leaving students feeling excluded and carrying with them a social stigma (OECD, 2012c(OECD, , 2013.
In summary, the education systems and challenges faced within them in England and Portugal hold notable similarities. In both countries, EL rates are very similar and reducing EL remains a significant issue to address. However, there are still numerous barriers relating to structural and institutional factors, including lack of sufficient funding for education and training; highly centralised education systems with prescribed curricula and little teacher autonomy; a perception that an academic education is more valuable than vocational training; a lack of labour market skills being developed amongst students; numerous issues around inequity, disadvantage and social injustice that persist in the education systems despite recent reforms leading to low attainment and selfesteem that contribute to EL risk.

Research methodology
We approached this study as a multiple comparative case-study (Yin 2003;Bartlett and Vavrus 2017) in order to gain an in-depth and critical understanding of the issue of EL (Edmonds and Kennedy 2017). Our two-country comparison of England and Portugal involved qualitative data collection and analysis of 59 semi-structured interviews and focus groups with 209 educational stakeholders working in a variety of settings and school pupils across 18 settings.
Participants were recruited through convenience and snowball sampling (Emerson 2015) with the aim of involving a range of educational stakeholders working with young people aged 12-21 who were potentially at-risk of EL or who had already disengaged. The literature on EL indicates that young people lack voice in the research process, which has contributed further to them becoming at risk of EL (Downes 2013). While it is challenging to involve those hard to reach, it is essential to hear their voices to understand their perspectives and experiences of the issues associated with EL in their contexts (Plano Clark and Creswell 2015).
Informed voluntary consent was obtained from all participants who were guaranteed anonymity, confidentiality and the right to withdraw from the research at any time (BERA 2018). For young people under the age of 16 (this applies only to England), information and consent forms were designed for parents, and only children whose parents had signed and returned these to teachers were interviewed. Interviews and focus groups were recorded with participants' consent and professionally transcribed. Participants and their settings were given pseudonyms and codes to protect their identities from the outset.
Coding of the data was undertaken via qualitative analysis software (NVivo Pro 12 for England and MAXQDA for Portugal). As noted earlier, we analysed our data by applying an ecological conceptual framework developed by the UK-based team (Brown, Douthwaite et al., 2021;Brown, Olmos et al., 2021) to code the data according to the five EL risk categories. In both countries, the significance of each risk category was determined by the frequency and length by which risk factors were discussed by participants across the datasets (Braun and Clarke 2006). In Windy County, significance was also determined by asking participants during interviews/focus groups to identify which was the most important risk factor. In this paper, we compare two of the five EL risk categories: Structural Factors and Institutional Factors.

Windy County (England)
In England, the research took place in a region in the South West of England, which we have named 'Windy County'. This context was chosen because of how its key characteristics relate to the issue of EL. The area is largely rural with agricultural work being higher than the national average. In addition, the county is situated in a large and very long geographical area meaning that the population is widely distributed, and people need to travel great distances to get to places of education, training and work. These issues together with the county's poor transport links present barriers to accessing education and training provision and employment opportunities, particularly for those at risk of EL. These issues also present challenges to supporting those at risk of EL with initiatives often being limited, disjointed and subject to project funding.
Despite the number of 'known' Early Leavers (for whom data are available) being lower than 2% in the region, data indicate that 9.3% of the destinations and activity of young people were categorised as 'destinations and activity is not known', bringing the possible number of NEETs to 11.3% -the fifth quintile in England. A key factor in the increase of 'unknowns' or 'hidden NEETs' was the cutting the government's Connexions Service, which had been responsible for tracking young people in England. As such, there is no longer one body responsible for mapping and tracking 16-and 17year olds. Efforts to do so by Local Authorities has been inconsistent and ineffective since 'policy changes have, in effect, fractured many LAs' established links with local schools and colleges and weakened their capacity to collect complete destinations data' (Maguire: 126). Maguire (2015, 125) highlights that although 'it cannot be assumed that all young people who have unknown destinations will, in effect, be NEET [. . .] these figures reveal that significant proportions of young people have no contact with support services, nor do they have access to intervention programmes which may improve their education, training and employment outcomes'. In addition, local government funding cuts means there is no department dealing specifically with EL in Windy County.
The research in Windy County took place in 11 educational settings, some of which were mainstream and highly centralised, and some of which provided alternative or special education to a range of learners, including those at risk of EL. We were interested to see if the latter sites might offer a different perspective to the more centralised education institutions. The settings included three academies; two SEND specialist schools; two VET settings (a Further Education College and a small community farm offering alternative provision); two educational charity programmes supporting Early Leavers and those at risk of Early Leaving with education and employment opportunities; a 'virtual school' offering educational support for children in care; and the local authority, providing employment-based training for NEET young people. Thirty semi-structured interviews and focus groups were conducted with 53 educational stakeholders including headteachers/members of school leadership teams, teachers, career advisors, educational charity workers/mentors and support service leaders. Our stakeholders helped us gain access to 39 young people identified as 'at risk' of EL who were interviewed through a total of eight focus groups. In total, 38 interviews and focus groups with 92 stakeholders were conducted across our 11 settings. Our findings are generated from the analysis of all our data, and we present quotations from a range of participants and settings that best illustrate these.

Vista Region (Portugal)
In Portugal, the research took place in 'Vista Region', which is a city and a municipality located in the Central Region of Portugal. Unlike Windy County, Vista Region's economy is mostly based on services, trade and light industries, with state-run public services including education, health and general public administration. The region also has a well-known higher education institution, whereas there are no HE institutions in Windy County.
Vista Region is highly populated with good transport links contrasting with Windy County's dispersed population and poor transport infrastructure. Upper secondary school data in the region indicate that 11.4% of students in 2018/2019 either repeated the year or dropped out of school (DGEEC, 2020). These statistics are below the national rate of 12.9%. However, in the sub-region of Vista Region, which has 10 associated municipalities, the percentage of students repeating the year or being Early Leavers was below the regional and national rate at 10.6% (Direção-Geral de Estatísticas da Educação e Ciência (DGEEC) 2020).
In the 2017/18 school year, the retention rate in the Vista region municipality was 6.2%, that is, 758 students out of 12,270 were kept back in the previous school year (CML, 2018). The data on retention rates indicate that these increase as students' progress through schooling; it is in secondary education that the highest retention rate has been recorded (n = 402 students), representing, in the last year of compulsory education, 25.1% of students' population against the national rate of 22.6% and the national 21.6% when focusing on Education and Training Courses/Professional Courses.
Our research in Vista Region took place in seven settings that were all offering Education and Training Courses/Professional Courses (ETC/PC) at the level of secondary education. All of our participants were either students or teachers on these courses. A total of 21 semi-structured interviews and focus groups were conducted with 117 stakeholders, including seven members of administrative boards, 42 teachers and 68 students.
ETC/PC courses are certified paths adapted to students' profiles and specific characteristics, providing them with opportunities to conclude compulsory schooling. Therefore, ETC/PC may also be seen as a way of EL prevention as it allows them to complete secondary education, as well as meeting identified local and regional labour needs in terms of employment and development (Law no. 26/89). One of the pedagogical characteristics of these paths is that they are conceived to have a strong technological and practical training, especially when compared to normative upper secondary schooling pathways. From 2004/2005 onwards, the professional courses became also an educational offer of secondary schools belonging to the public network, at not only until an exclusive offer of professional schools (Pinto, Delgado, and Martins 2015).

Structural risk factors in Windy County, England
In Windy County, SF was the most significant category across the dataset. However, while adult stakeholders identified numerous risks in this category, young people spoke the least about SF, discussing only three: educational policy; the neoliberal emphasis on measuring outcomes; and the lack of pathways to education and training.
Educational policy was the most significant risk area, aligning with discussions around the neoliberal emphasis on measuring outcomes. Stakeholders spoke about schools being 'forced [. . .] through the league tables' to ensure students were achieving key targets in core subjects associated with the EBacc, because these 'carry more weight on the league tables than other subjects' (Dapple Academy Lead). This resulted in a 'narrowing of the curriculum' in schools, with more time allocated to teaching core subjects 'to improve results' on league tables, leaving 'very little time' for the more 'creative subjects' and 'fun lessons', such as dance and art (Dapple Academy Lead). Pupils were therefore getting 'much more English, maths and science', which was particularly problematic for disadvantaged students who 'just can't cope with it' (Dapple Academy Lead): In addition to exam-related anxiety, students experienced a great deal of pressure from teachers to achieve acceptable targets and grades, particularly in English and maths. Students attributed this pressure to schools being concerned about 'Ofsted' noting that teachers 'want the school to get outstanding' (Sarah, Cloudy Academy). Students believed teachers were driven more by performance targets than caring about students 'learning it to learn it' (Imogen, Cloudy Academy) indicating a sense of neglect and a lack of trust for those responsible for supporting their learning. Educators also highlighted the performance pressure and stress they were under to get students through exams, feeling 'forced down a pursuit of excellence road' (NEET Project Manager).
Pressure to achieve was discussed by stakeholders as crucial to pursuing future education, training and work opportunities, echoing Archer et al.'s (2016) observation that students see the EBacc and achievement of 'good' grades as a passport to future success. Nevertheless, this pressure can lead students to drop out of school, as one student noted: Stakeholders also spoke about the recent grading system reforms noting that the new 9-1 numerical grading was making it more difficult for students to achieve good grades: What used to be a C now doesn't really exist, it lies between a 5 and a 4. A 5 is a good pass, a good C, and a 4 is a low C. At the moment a 4 is being accepted as a pass, I think within the next year or so it is going to drop away, and a pass is going to be a 5, and up from there. (Dapple Academy Lead) A student expressed her frustration with recent curriculum and grading reforms noting that these were imposed on students, without consultation, by government officials out of touch with the needs, interests and realities of the younger generation: I think they should do a survey [. . .] get the students to say how they feel because with the Government, they are all like 30, 40, 50 [. . .] 60 years old, they haven't been at school for at least 20 years and we are the ones caught; and they change the grading system [. . .] because that's how they like it at school, but we are the ones that are like in this situation and no-one ever asks us [. . .] 'do you agree?' [. . .] 'do you like this?' (Imogen, Cloudy Academy) Raising the age of compulsory schooling to 18 was also seen as a risk factor causing young people to 'disengage' and 'stop attending' (LA Support Service Officer). In addition, the requirement for post-16 maths and English was viewed as a 'barrier' that 'probably puts more young people off' pursuing educational or vocational pathways (LA Support Service Lead). Educators spoke of the 'need to be looking at alternative education provision for those that are displaying signs of risk of NEET because if we carry on trying to teach our young people in the same way [. . .] we are always going to have NEET' (LA Support Service Officer). It was also noted that 'there is a distinct lack of range of apprenticeship opportunities [and . . .] nothing alternative to college or sixth form' with 'year 12s most likely [to] have dropped out [because] college is not what they're looking for' (Dapple Academy Lead). These issues led a NEET Project Tutor to note 'we're setting them up to fail'.
A lack of funding and resources was another key EL risk amongst educators. A 'rapidly decreasing budget' was leading to 'depleting resources' and cuts to youth/support services for at-risk leavers, which were 'here today and gone tomorrow' (LA Support Service Leader). In schools, there is 'offrolling of children [. . .] when it comes up to them taking their GCSEs' because 'schools [. . .] can't meet their needs [. . .] they've had funding cuts so [. . .] they want to get rid or they don't want it on their league tables' (LA Support Service Leader). Off-rolling is the practice of removing a pupil from the school roll without a formal, permanent exclusion or by pressuring a parent to remove their child from the school roll, which they often agree to in order to avoid a permanent exclusion on their child's record. It has been found that 'disadvantaged pupils, those with special educational needs, and pupils with low prior attainment are disproportionately removed from the school roll' (Owen 2019). This demonstrates how neoliberal policies and the pressure on schools to perform well and get good Ofsted ratings can result in schools operating in their best interests rather than in the best interests of pupils in need of further support. It also further highlights the inability of mainstream settings to support young people from at-risk groups.
The rurality of the region and its inadequate and expensive transport is also a significant risk factor, particularly regarding access to education, training and support services. Living in 'out of the way places' makes young people feel 'isolated' and 'cut off' because they 'can't access much' (Kite Academy Careers Advisor). This was particularly problematic for students from lower socio-economic backgrounds and families who did not drive or have cars.
Finally, adult participants highlighted the absence of effective mechanisms to track those at risk of Early Leaving, including early identification measures. They reported that there is no requirement for schools to track the home-schooled nor to inform the LA about post-16 leavers who may have dropped out at the FE stage or post-18 leavers with an Education Health and Care Plan (i.e. identified as SEND and in need of additional support than is available). This was contributing to the high level of 'unknowns' (students whose destinations and activity are not known) in the region.

Structural risk factors in Vista Region, Portugal
In the Portuguese context, SF was the fourth most significant category across the dataset. As in Windy County, stakeholders highlighted educational policy issues as the most problematic for young people at risk of EL. In particular, the Ministry of Education's centrally dictated statutory guidance for schools' implementation of educational programmes, curricula, assessment and behavioural policies was viewed as rigid and not suitable for all learners: Some [teachers] feel some difficulty in having a single type of educational model or offer for every young person. There has been some concern. (Neighbourhood School, Study Programme Coordinator) Despite both general and vocational pathways being available to students from the age of 15, a member of the school board and administration from Sun School noted that this issue of a one size fits all model of education 'also relates to the content and programmes of professional courses'. They also highlighted that the prescribed vocational and professional courses and curricula are overly theoretical and are not adjusted to the real needs of the labour market, leading to an unappealing formative offer unable to cater to the needs of young people. Those at risk of EL are left struggling to fit into the existing system and model, raising perennial questions as to 'who should adapt to whom?' and 'what knowledge is of most worth?' (Bellack 1965;Derry 2017): Often, students also don't adapt themselves to what the school has to offer. There should be other ways, other possibilities. (Downtown School, Study Programme Coordinator) Teachers reinforced their institutional leaders' perspectives stating that the curricula of vocational and professional courses prescribed by the Ministry of Education are inadequate and cannot be adapted to different learners' needs. In line with this, they noted that language studies are too general, similar if not the same to those taught within the more academic pathways, and therefore do not teach students how to use language for business or technical purposes: This notion is also shared by students who, as in Windy County, often expressed feeling a 'lack of interest in some courses' (Gus, Blue School), particularly with their largely theoretical vocational studies, which they were concerned were not preparing them well for future employment. This feeling is expressed by Amanda (Yellow School) who said: 'we have to study a lot of things that I don't think will do any good for our future'. Similar to Windy County, an overly academic and theoretical education with excessive hours required for teaching and learning abstract theoretical elements was seen to be both inappropriate and tedious by young people and adult stakeholders. It was seen as particularly inappropriate in respect of vocational and professional courses. A school board and administration teacher at Sun School mentioned that 'the kids didn't need to have so many hours of Portuguese, English, integrated area [i.e. citizenship studies], [so] maybe reduce [these] and then put a more practical element there for them to do other things'.
Another point of concern amongst both teachers and students in Vista Region was related to human resource issues, particularly in terms of 'constant inflows and outflows of teachers [and . . .] no stable teaching staff to make students feel welcomed and that they feel they are in the right place' (Blue School teacher). Teacher presence and continuity at school was deemed essential for students' sense of wellbeing and belonging. Students expressed concern around the 'lack of professionals' (Carol, Blue School) and 'few teachers in the school' (Mark, Sun School) noting that 'they should hire teachers' (Mark, Sun School). Insufficient numbers of teachers was seen to result in a lack of key support to guide students to successful completion of their education and training pathways.

Institutional risk factors in Windy County, England
In Windy County, the third most discussed risk category across the dataset was institutional features of schooling. For young people, IF received more attention than the other risk categories. Stakeholders noted that traditional schooling and the classroom environment are unsuitable for at-risk learners who 'don't like it in there' and find it 'so boring just sitting at a desk' (Dapple Academy Lead). Students noted feeling physically and psychologically uncomfortable in the classroom, which made them not want to be there: today in the class, it was really hot [. . .] and I was getting so annoyed because it felt like I was burning [. . .] I literally wanted to walk out of class. (Alex, Misty SEND School) Large class sizes are a particular challenge as at-risk learners 'generally don't function in classroom settings in groups over about ten (Misty SEND School Leader) and 'could really do with a more personalised approach, smaller classes' (Dapple Academy Curriculum Lead). A NEET Project Manager noted that when '[y]ou've got a very big class [. . .] you can't dedicate too much time to a young person [. . . so] it's quite difficult to provide some of the young people the levels of attention that they need'. Students noted feeling overwhelmed by 'too many people in class' and 'too much going on' (Juliet, Dusky SEND School) and that 'when there are 30 other kids and you are asking for help [. . . teachers] have to go around all the other students and then go to you' (Adam, Dusky SEND School). A teacher at Cloudy School noted that 'it's very difficult with some students [. . .] particularly an ASD student who cannot cope in mainstream'. SEND children were seen to be better placed in special schools that are more equipped to meet their individual needs.
School management of student behaviour was the most significant issue in the IF category and the second most discussed area across the whole dataset. Some students felt unsafe in school as there were not enough staff present to deal with highly disruptive students, particularly outside the classroom: Despite the lack of school staff to fully monitor and respond to such behaviour, school rules and behavioural sanctions were viewed as overtly rigid and unsupportive of at-risk learners who were frequently removed from the classroom and 'sent straight away to an isolation room' often for the entire day (Dapple Academy Pastoral Lead). These 'internal exclusions' were seen by educators as preferable to school exclusions, which could lead to 'kids being stabbed' and further 'offending' and 'antisocial behaviour' (LA Support Service Officer). Isolation was also viewed as a reasonable way to 'concentrate on the ones that do want to or can learn' (LA Support Service Officer) while providing those removed with their own work to do. However, students who had experienced isolation complained that it was 'insane!' [. . . as we] don't learn anything in there because [. . .] every time you go in there, you're doing the same work and it's the same book that you'll get given' (Daniel, Dapple Academy). Jeff (Dapple Academy) expressed frustration that in isolation 'it's really silent and it's really annoying because [. . .] you're sat down for six lessons' and 'I feel like the room closes down', suggesting he felt confined and restricted. There was also a sense of injustice and confusion amongst young people regarding behavioural penalties being regularly and haphazardly enacted by teachers. Some students tried to discuss the perceived problematic behaviour with their teachers but felt that they were not listened to, prompting Daniel (Dapple School) to note that 'some teachers give you warnings for no reason' when 'it's not your fault and it's not fair'.
Another significant issue was the limited time and resources available to schools and teachers to support the needs of all learners. Cuts to finances and services have resulted in 'pastoral systems, counsellors [. . .] being pulled and stripped back' so that 'schools aren't resourced for mental health' and are 'not set up and supported and financed to deal with everything we should be' (Dusky SEND School Leader). This was impacting on 'teachers' workloads and teachers leaving the profession' (ibid). It also meant that students were not receiving enough guidance during key transitions from school to further education or training leading to at-risk students who 'don't have a destination for September or can't progress onto the next level of course' (Wispy College NEET Project Manager). A lack of guidance also led to at-risk learners starting courses that are not within their interests and 'end up dropping out anyway because it wasn't the right match' and they were 'unhappy with the course or feeling jaded' (Wispy College Tutor). An officer at the Violet School noted that at-risk learners need support to reflect on the options available to them and think through these, which needs to involve educators 'just sitting down with the young person and saying 'oh, why have you chosen hair and beauty? You've never showed any interest before, is this something new? Is this something you really are keen on?' and if not then saying 'so actually this isn't the course for you, let's look at something else that you might be good at'.
Educators noted that at-risk learners also often become cut off from future learning opportunities, 'particularly in the rural areas, because the education establishments are so far away [. . .] and the buses are so limited' (NEET Project Manager). What 'keeps then NEET is the rurality of the county' and that 'transport systems are poor' and 'expensive' (LA Support Service Officer). This was seen to be even more challenging for children who were previously in 'special educational environments' because 'when it comes to going to college [. . .] it's just all very overwhelming [and] there doesn't seem to be a warm-up to that [. . .] so we find that some of them, if it's not planned properly -the transition -they can really regress and not want to go' (Dusky Academy Lead). It was noted that '[t] he school aren't putting anything in place' to support 'vulnerable' children, particularly those who are 'fairly high functioning [. . .] academically' but have 'additional needs' and 'don't have the skills to be unsupervised [. . . needing] support with transport' (NEET Project Manager). As such, it was suggested that education for those at-risk of EL should include 'travel training to get them from home to the college independently' to try to minimise the impact on attendance (Dusky Academy Lead).

Institutional risk factors in Vista Region, Portugal
In Portugal, institutional features of schooling was the third most significant risk category across the dataset. The requirement for and emphasis on theoretical learning means that teachers feel obliged to implement largely traditional teacher-centred pedagogical approaches. These aspects were seen to be unmotivating and inappropriate for the current generation of students who value more interactive and active approaches to learning. One member of the school board noted that 'the issue of [. . .] sitting in classes, being "lectured" -this is not beneficial for them' (Yellow School). This was viewed as particularly inadequate for those who are less 'academic', more vocationally oriented and/or at risk of EL: we are talking about our vocational education students, and if we have a theoretical, expository pedagogical approach, this does not resonate with them, this generation. (School board member, Green School) Educational stakeholders explained that students on vocational and professional courses experience very long days in the confined space of the traditional classroom, which is counterintuitive to equipping students with the practical experience necessary to enter the labour market: They spend many hours in classrooms and that is not the philosophy of vocational education, it is much more practical. They come in at half past eight and often leave at half past six. They are closed in for a whole day. (Julia, Downtown School teacher) They spoke about the need for reviewing and reforming the curriculum and syllabuses of such courses to get students out of the classroom and into professional contexts to gain practical experience in real-world settings: It would be desirable to come here only one or two days a week and have practical classes in integrated teaching in companies or in places open to the social context as seen in other countries. (Louise, Downtown School Teacher) This is what was recently undertaken in the context of England, where new vocational routes, such as the T Level course and qualification comprising work experience, were launched in September 2020.
Discussions amongst the young people interviewed reinforce the fact that their classes have a strong theoretical component that does not align with their needs, interests or aspirations and does not help prepare them for the labour market in the context in which they live or further afield. One student noted that having more interactive classes that involve discussion and collaboration 'would be a good idea' (Danny, Downtown School). Another student made the following comment: teachers could adapt this program to do practical activities, not just theoretical. So maybe the classes would be a bit more productive, the students would be more interested, and it wouldn't always be the same routine -sit down, open the book, write, and close the book, leave, repeat -going in, sitting down, opening the book over, and over, it gets tiring. A practical part of class would make it a little more appealing. (Travis, Neighbourhood School) Young people also highlighted elements they felt could be improved in schools to reduce the likelihood of EL, including the need for greater financial investment and a broader range and number of material resources. This was seen to be particularly important for professional courses where the traditional classroom setting and arrangement around tables and chairs is not conducive towards learning how to carry out a range of necessary practical tasks. As such, spaces need to be created that are suitable for the number of students in the classroom and that enable students to engage in concrete hands-on activities, for example, having functional sinks where students can wash hair as part of a hairdressing course.
Another key issue that was mentioned by teachers and students was the importance of supporting young people emotionally and psychologically, particularly when behavioural issues are identified. It was noted that clearer and more effective measures need to be in place both to deal with disruptive behaviour at the moment it occurs and to have longer-term support in place. Educators noted the importance of open channels of communication and the involvement of families and educational professionals when necessary: talking is very important. To talk to them [the students], the head teacher essentially, [. . .] and when I think it is important, to call the family, the parents, and the psychologist afterwards. (Hill School, Study Programme Coordinator) However, a significant barrier in this respect was the lack of trust students have for educational professionals, such as school psychologists, that results in them not feeling comfortable to express themselves or confide in them. This severely impacts on students with emotional and mental health needs gaining the support they need: Here we have a psychologist at school [. . .] people go there, talk, [. . .] vent with the person and everything that comes out of there will be told to parents [. . .] in the middle of it, it all gets shuffled and they don't want to talk to anyone, they prefer to keep everything to themselves. (Susan, Neighbourhood School) Students also spoke about poor organisation and decision-making in relation to school rules and behavioural policies, describing them as overly rigid and inappropriate for a variety of circumstances, particularly in terms of how behaviour is monitored throughout the school year. There was a sense of injustice and dissatisfaction in regard to the implementation of behavioural rules and sanctions, with students expressing concern that these are not having a positive impact on their learning experience and need to change.
In terms of planning their education and training pathways and learning experiences, young people talked about the lack of information provided to help them make informed decisions around the range of formal curriculum and extra-curricular options on offer at various educational institutions. Some students reported they may have made different choices that were more appealing to them if they had been provided with key information that was appropriately presented to them and if they had been given opportunities to reflect on their interests and aspirations and talk through their options. Guidance with selecting extra-curricular activities was seen to be as important as more formal course choices as they promote the development of positive learner identities and often have a more practical component that is appealing.

Discussion
We now compare our key findings across the two regional contexts, discussing major similarities and differences. Issues highlighted within and across the two risk categories include overly academic educational provision; lack of teacher autonomy and time; insufficient school resources and funding, including a lack of adequate career guidance; and ineffective behavioural policies leading to a lack of trust by students in adults.
Several issues related to SF emerged in the data, which often reflected young people's wider disadvantages and the various challenges posed by the national education system. Educators in both contexts perceived a top-down influence in the way that SF of EL risk impact on institutional features of schooling. This echoes Kandel's (1933), Bronfenbrenner's (1979 and Sobe and Kowalczyk's (2013) observations that macro-level systems, factors and forces intersect, overlap and influence micro-level systems, such as the school and educational setting.
In both regions, stakeholders were dissatisfied with the centralised approach to educational policymaking, which resulted in a prescribed, broad, academic curriculum determined by government officials out of touch with students' current requirements. Educational provision was seen as a 'one size fits all' model that fails to adequately prepare young people for the labour market. The focus on theoretical knowledge over practical skills reflects traditional theories that every child should learn a central body of valuable knowledge (e.g. Hirsh, 1988). However, the prevailing socio-cultural view that academic education is more valuable than a vocational one was viewed as problematic, as was the lack of alternative and vocational options in England (Chankseliani, Relly, and Laczik 2016). The traditional classroom setting was also deemed outdated with the school environment comprising inappropriate learning spaces.
Teachers expressed lacking autonomy and agency, particularly in organisational and curriculum aspects. They felt restrained in terms of delivery, with the emphasis on theoretical learning requiring them to adopt largely teacher-centred rather than learning-centred pedagogical approaches (Hubball and Burt 2004). They also noted that spending long hours in the classroom delivering largely academic subjects detracted from having time for pastoral elements that are particularly important in supporting vulnerable young people at risk of EL (Tucker, 2013;Kyriacou 2015).
Insufficient funding and resources, including staff shortages, were also seen as a risk to EL. In Windy County, educators highlighted funding cuts, depleting resources and off-rolling of those deemed least likely to do well in exams. In Vista Region, students emphasised the lack of financial investment in providing a suitable learning environment with adequate materials and resources. This also interfered with educators' capacity to adequately support at-risk learners in terms of providing guidance around suitable future education and training pathways. As such, many of those at risk took unappealing courses, leading them to drift and drop out (Van der Graaf et al. 2019;Lavrentsova and Valkov 2019;D'Angelo and Kaye 2018). This echoes the literature that a lack of or misdirected 'orientation' (i.e. personal, educational or career guidance) is a significant EL risk and therefore an important intervention point (Olmos and Gairín, 2021).
Institutional rigidity around school rules and behavioural regulations was a significant issue amongst at-risk learners who noted that these were inappropriate, inconsistently applied and impacted negatively on their learning. In Windy County, being removed from the classroom and sent to isolation was seen to be particularly harmful, and in Vista Region, there was a sense of injustice and dissatisfaction with the disorganised and inconsistent manner in which behaviour was monitored and sanctioned. This resulted in a lack of trust in educational professionals and a breakdown in communication (Hattam and Smyth 2003;Blue and Cook 2004), leaving students disconnected from those who were supposed to be supporting them (Hodgson 2007).
There were also noticeable disparities between the two regions and groups of stakeholders. In Windy County, SF was the most discussed category, particularly amongst educators, whereas in Vista Region, it was the fourth most significant category. This difference could be related to the varying sociocultural, political and economic contexts of the two regions. For example, in Portugal, there is less expectation around governmental and institutional intervention and a view that the family has a central role to play in the reduction of EL (Mangas, Lopes, Milhano and Freire, 2020). In England, on the other hand, governmental and institutional responsibility to support those disadvantaged in society has been emphasised, with a 'blame perspective' attributed to parents leading to increasing intrusion of the state in childrearing policies (Levitas 2005).
This finding presents comparative researchers with a dilemma because while structural factors are rated as the fourth most significant category in Portugal, perhaps due to cultural differences, the data from the teachers and students suggest that over-centralisation, a structural factor, is the key issue in both countries. While it is arguable as to how we should interpret this difference, we are tempted to claim that our overall perspective on both education systems should be considered by policymakers.
We also found differences between the risk categories and issues emphasised by adults compared with young people. Educators were more inclined to highlight various structural challenges imposed by governmental policies and the national education system that were impacting on IF, including their ability to support those at risk of EL. Young people were more cognisant of how IF were directly impacting their everyday experiences of schooling and learning, including preparing them for future life. These different foci might be explained by the likelihood that stakeholders are more conscious of how factors just outside their jurisdiction impact on their daily realities (Bronfenbrenner 1979(Bronfenbrenner , 2005Brown, Douthwaite et al., 2021;Brown, Olmos et al., 2021).
Another noteworthy variation was the issue of exam pressure. In England, neoliberal marketisation policies mean that teachers are under considerable pressure to reach certain targets in league tables (Harlen and Deaken, 2002). This led to at-risk learners being left largely unsupported so that teachers can focus on the students more likely to perform well, echoing Gilbourn and Youdell's (2000) model of educational triage. In Portugal, exam pressure has been problematic in the past, but was not mentioned by stakeholders, perhaps due to recent assessment reforms (D'Angelo and Kaye 2018). Nevertheless, performance in terms of students needing to pass the year to avoid grade repetition was a significant issue that left those at risk of EL trapped in a system that fails to provide them with a valuable learning experience and delays or prevents them from being able to move into more appropriate education, training or employment pathways.
Finally, the urban versus rural geographical contexts of the two regions as well as the socioeconomic status of families resulted in some key differences. In Windy County, young people encountered significant challenges in accessing education and training because of poor transport links, huge distances to travel and the associated high cost of public transport. In Vista Region, however, the area is highly populated with good transport links, so this was not raised as an issue. These differences reflect previous research that has found that students living in rural areas are more at risk of EL (Nikolaou, Papa, and Gogou 2018).

Conclusion: implications and recommendations for policy and practice
In this article, we drew on Brown et al.'s (Brown, Douthwaite et al., 2021;Brown, Olmos et al., 2021) ecological framework of five risk categories to EL to analyse and compare structural and institutional factors of risk across two geographical regions in Europe: 'Windy County' in England and 'Vista Region' in Portugal. We drew on interview and focus group data with 209 educational stakeholders identifying similarities and differences in the issues raised within the two contexts, as well as between educators and young peoples' views. We noted linkages in the way that the identified issues within the two risk categories significantly overlap, interact and influence each other, particularly in terms of how SF hinder schools and educational professionals from effectively supporting those at risk of EL. Our analysis highlights the complexity of addressing EL, which comprises multifarious factors that when merged, further impact on the challenges young people face, contributing to a cycle of disadvantage, a danger of marginalisation from future learning and work opportunities and bleak life chances.
Our analysis indicates that the schooling system in both countries is outdated and unfit for purpose (Brown, Douthwaite et al., 2021;Brown, Olmos et al., 2021). Our findings suggest that reform of centralised educational policies is needed, including elimination of the requirement for grade repetition in Portugal and of not being able to move forward in education or training without level 4 English and maths, and until aged 18, in England. The education systems need to move away from providing a 'one size fits all' academic model of education that results in schools essentially failing to cater to diverse students' interests and abilities and to prepare them with the skills needed to enter the labour market. There is a need to provide further vocational provision at the school level, which should be flexible and enable young people to adapt their learning pathway to their needs and interests. This should include options to spend less time in a traditional classroom setting, with opportunities for work experience and placements. Forging links with employers who would support and facilitate such opportunities would be beneficial.
In line with the above, our findings reveal that while educators spoke about lacking agency in terms of curriculum and pedagogy, which was impacting negatively on students' learning experience, students felt that they had to fit into an educational mould, lacking choice, freedom and flexibility to pursue paths that suit them. Learners in England mentioned feeling they lacked voice and wanted to be consulted in matters affecting them, suggesting young people would benefit from being given greater agency and involvement in shaping their own educational experiences. As such, and echoing previous recommendations, we suggest that young people are offered more opportunities to voice their preferences, concerns and views in school but are also consulted on governmental education policies that directly affect their lives (European Commission 2013).
In conclusion, our two-country comparison has revealed significant structural and institutional issues that need to be addressed in both contexts to reduce EL. Our analysis has highlighted the importance stakeholders place on the need for further preventative measures that are holistic, coherent and consistent, inclusive and ethical, and that enable all stakeholders, including young people, to be agents in determining and directing learning pathways that are personalised to learners' needs.

Disclosure statement
No potential conflict of interest was reported by the author(s).