View in EDS HTML Full Text PDF Full Text

The GCSE Attainment Gap: Assessing the Influence of Permanent School Exclusion

Saved in:
Bibliographic Details
Title: The GCSE Attainment Gap: Assessing the Influence of Permanent School Exclusion
Language: English
Authors: Stephen Hills (ORCID 0000-0001-5894-1418), Matthew Walker (ORCID 0000-0002-3080-0530), James Guinn (ORCID 0009-0005-5451-1644), Aubrey Kent (ORCID 0009-0002-8537-4409)
Source: British Educational Research Journal. 2025 51(3):1481-1497.
Availability: Wiley. Available from: John Wiley & Sons, Inc. 111 River Street, Hoboken, NJ 07030. Tel: 800-835-6770; e-mail: cs-journals@wiley.com; Web site: https://www.wiley.com/en-us
Peer Reviewed: Y
Page Count: 17
Publication Date: 2025
Document Type: Journal Articles
Reports - Research
Education Level: Secondary Education
Descriptors: Foreign Countries, Educational Attainment, Exit Examinations, Achievement Gap, At Risk Students, Disadvantaged, Scores, Secondary School Students, Inclusion, Access to Education, Barriers, Expulsion, Equal Opportunities (Jobs)
Geographic Terms: United Kingdom
DOI: 10.1002/berj.4133
ISSN: 0141-1926
1469-3518
Abstract: Permanent school exclusions continue to be a topic of keen interest to UK schools and policymakers. The debate over the practice has recently intensified owing to the perceived negative outcomes directly resulting from the exclusion event. Research has indeed shown that pupils who have been permanently excluded are at a greater risk for a variety of negative life outcomes when compared with their non-excluded peers. However, that disadvantaged groups are disproportionately represented among those excluded has not been accounted for in empirical testing. Accordingly, previous measures of the influence of permanent exclusion may have over-estimated its negative consequences because they have not controlled for disadvantageous pupil characteristics. This is a critical limitation of the research owing to the influence of confounding variables and sample selection bias. Using the National Pupil Database and a full cohort of UK pupils (N = 590,092), our analysis tracked a sample of 1490 pupils permanently excluded in year 11 of the English education system in 2018/2019. Using capped GCSE points as the academic attainment variable, we find that permanently excluded pupil scores were nearly 25 points lower than their non-permanently excluded peers. However, when controlling for disadvantageous pupil characteristics, this difference was cut roughly in half. As such, we conclude that permanent exclusion is neither the catalyst of disadvantage nor a continuation of disadvantage on the same trajectory, but rather an accentuation of existing disadvantage. In other words, the existing trajectory of disadvantage gets steeper following the permanent exclusion event. Therefore, considering that the GCSE attainment gap found is equally attributable to both permanent exclusion and disadvantageous pupil characteristics, policymakers should both limit permanent exclusion to being a last resort and provide additional support for pupils at risk of being permanently excluded. Including a permanently excluded pupil's GCSE attainment in their former school's academic league table data incentivises schools to act in the best interests of these highly disadvantaged and vulnerable pupils.
Abstractor: As Provided
Entry Date: 2025
Accession Number: EJ1475322
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwEAY_UlLwV6a7w6gvm8HQLhAAAA4zCB4AYJKoZIhvcNAQcGoIHSMIHPAgEAMIHJBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDHNmgxsGsbIArzku1wIBEICBm91RhIURfm9e7JEMQ-4tl_PyTj8A_ET8i4qE1qfw2H2a-aIhl7aifsFaW5o_Gq9Xt6M_CWdk7DElysaWxGisGS8lE4vwaLAVIEqdXedOREk0_N74jy04ryWOTrBpn56TjOEN4H48WpzplZsAW_JZFnBms_-9da39PIEAEGv3pXoZuJlaMtay1iZfH4Kq7_rdcxnxcO9M4AgdNvMs
Text:
  Availability: 1
  Value: <anid>AN0186252690;bed01jun.25;2025Jul01.05:03;v2.2.500</anid> <title id="AN0186252690-1">The GCSE attainment gap: Assessing the influence of permanent school exclusion </title> <p>Permanent school exclusions continue to be a topic of keen interest to UK schools and policymakers. The debate over the practice has recently intensified owing to the perceived negative outcomes directly resulting from the exclusion event. Research has indeed shown that pupils who have been permanently excluded are at a greater risk for a variety of negative life outcomes when compared with their non‐excluded peers. However, that disadvantaged groups are disproportionately represented among those excluded has not been accounted for in empirical testing. Accordingly, previous measures of the influence of permanent exclusion may have over‐estimated its negative consequences because they have not controlled for disadvantageous pupil characteristics. This is a critical limitation of the research owing to the influence of confounding variables and sample selection bias. Using the National Pupil Database and a full cohort of UK pupils (N = 590,092), our analysis tracked a sample of 1490 pupils permanently excluded in year 11 of the English education system in 2018/2019. Using capped GCSE points as the academic attainment variable, we find that permanently excluded pupil scores were nearly 25 points lower than their non‐permanently excluded peers. However, when controlling for disadvantageous pupil characteristics, this difference was cut roughly in half. As such, we conclude that permanent exclusion is neither the catalyst of disadvantage nor a continuation of disadvantage on the same trajectory, but rather an accentuation of existing disadvantage. In other words, the existing trajectory of disadvantage gets steeper following the permanent exclusion event. Therefore, considering that the GCSE attainment gap found is equally attributable to both permanent exclusion and disadvantageous pupil characteristics, policymakers should both limit permanent exclusion to being a last resort and provide additional support for pupils at risk of being permanently excluded. Including a permanently excluded pupil's GCSE attainment in their former school's academic league table data incentivises schools to act in the best interests of these highly disadvantaged and vulnerable pupils.</p> <p>Keywords: academic attainment; disadvantage; exclusion; schools</p> <p>Key insights What Is the Main Issue that the Paper Addresses?Schools in England make much greater use of permanent exclusion than their counterparts in the rest of the United Kingdom despite considerable evidence of the negative effects of permanent exclusion. There is also evidence that permanently excluded pupils are disproportionally disadvantaged, such that prior measurements of the effect of permanent exclusion may be over‐estimations owing to not controlling for disadvantageous pupil characteristics. What Are the Main Insights that the Paper Provides?When controlling for disadvantageous pupil characteristics, the negative influence of permanent exclusion on GCSE attainment halves. This means the decision by a school to permanently exclude does not alone predict lower GCSE attainment, but also prior disadvantages. As well as reducing permanent exclusions by only using them as a last resort, greater focus should be given to alleviating the disadvantage of pupils at risk of being permanently excluded. Including an excluded pupil's GCSE attainment in their former school's academic league data will incentivise schools to act in the best interests of their disadvantaged and vulnerable pupils at risk of being excluded.</p> <p>Each year thousands of pupils are permanently excluded from schools in England at a considerable cost to British communities and UK society as a whole (Department for Education, [<reflink idref="bib13" id="ref1">13</reflink>]; Parsons & Castle, [<reflink idref="bib42" id="ref2">42</reflink>]). School exclusions are largely considered an extreme form of discipline that heightens inequalities and are regarded as the first step in exclusion from mainstream society (Blyth & Milner, [<reflink idref="bib3" id="ref3">3</reflink>]; Kulz, [<reflink idref="bib29" id="ref4">29</reflink>]). There is also consensus that permanently excluded pupils are at a greater risk of variety of negative outcomes than their non‐excluded peers (Pirrie et al., [<reflink idref="bib44" id="ref5">44</reflink>]), which include crime, drug use and other anti‐social behaviours (Berridge et al., [<reflink idref="bib2" id="ref6">2</reflink>]; Daniels & Cole, [<reflink idref="bib9" id="ref7">9</reflink>]; Hodgson & Webb, [<reflink idref="bib24" id="ref8">24</reflink>]; McCrystal et al., [<reflink idref="bib38" id="ref9">38</reflink>]; Pritchard & Cox, [<reflink idref="bib46" id="ref10">46</reflink>]). Research has found that excluded pupils have limited aspirations owing to stymied academic attainment and limited career prospects owing to being outcasts from their mainstream peers (Berridge et al., [<reflink idref="bib2" id="ref11">2</reflink>]; Daniels & Cole, [<reflink idref="bib9" id="ref12">9</reflink>]; Mainwaring & Hallam [<reflink idref="bib35" id="ref13">35</reflink>]). Despite this evidence, UK schools still support the practice and the number of those excluded continues to rise.</p> <p>Permanent exclusions are less frequent in the rest of the United Kingdom, for example Scotland, where exclusions are near zero (Duffy et al., [<reflink idref="bib15" id="ref14">15</reflink>]; McCluskey et al., [<reflink idref="bib36" id="ref15">36</reflink>]). Countries like Scotland have learned what the rest of the world is starting to realise—the negative consequences of permanent exclusion have been over‐estimated and improperly communicated. That stated, we must acknowledge that permanently excluded pupils are also disproportionally disadvantaged in terms of special education needs, disabilities (Bowman‐Perrott et al., [<reflink idref="bib4" id="ref16">4</reflink>]; Krezmien et al., [<reflink idref="bib28" id="ref17">28</reflink>]; Strand & Fletcher, [<reflink idref="bib50" id="ref18">50</reflink>]), low socio‐economic status (Strand & Fletcher [<reflink idref="bib50" id="ref19">50</reflink>]) and minority status (Achilles et al., [<reflink idref="bib1" id="ref20">1</reflink>]; Demie, [<reflink idref="bib10" id="ref21">10</reflink>]; Krezmien et al., [<reflink idref="bib28" id="ref22">28</reflink>]; Strand & Fletcher, [<reflink idref="bib50" id="ref23">50</reflink>]). These factors contribute to excluded pupils being arguably the most vulnerable in UK society and the ones most in need of better forms of intervention.</p> <p>Using General Certificate of Secondary Education (GCSE) attainment of a full cohort of pupils in England, the purpose of our study was to assess the academic attainment gap between permanently excluded pupils and non‐permanently excluded pupils. To this end, we provide a raw measurement of the GCSE attainment gap compared with an adjusted measurement of the GCSE attainment gap that controlled for disadvantageous pupil characteristics. This was done so as to assess the relative influence of permanent exclusion vs. disadvantaged characteristics, adding to the debate on whether permanently excluded pupils should be supported or blamed and punished (see MacRae et al., [<reflink idref="bib34" id="ref24">34</reflink>]; Parsons, [<reflink idref="bib41" id="ref25">41</reflink>]).</p> <hd id="AN0186252690-2">Negative consequences of permanent exclusion</hd> <p></p> <hd id="AN0186252690-3">Criminal and anti‐social Behaviours</hd> <p>Empirical research into the negative consequences school exclusion has focused on the criminal and anti‐social behaviours of excluded pupils and the subsequent education and employment outcomes achieved by these young people. For example, Pritchard and Cox ([<reflink idref="bib46" id="ref26">46</reflink>]) tracked a full cohort of 227 permanently excluded pupils in England. From their research, the authors determined that 63% of their sample went on to obtain criminal convictions between the age of 16 and 23 years old—averaging 7.4 offences each. The authors estimated this rate of criminal convictions cost UK taxpayers £4.16 million. However, the authors of the study failed to measure criminal convictions prior to permanent exclusion, which is a significant limitation not acknowledged in the research. Hodgson and Webb ([<reflink idref="bib24" id="ref27">24</reflink>]) interviewed 56 permanently excluded pupils and found that while 40 had offended, 36 of these had commenced their offending prior to exclusion. In addition, 50 of the 56 reported that they felt no more likely to offend subsequent to being permanently excluded.</p> <p>Berridge et al. ([<reflink idref="bib2" id="ref28">2</reflink>]) studied the effects of permanent exclusion on the offending careers of young people from a population of 343 permanently excluded pupils, over six different local authorities in England. The authors were able to access the complete police records of 263 excluded pupils, from which 117 (i.e. 44% of the sample) had no recorded offences prior to permanent exclusion but a record of offending following permanent exclusion. These results suggest that the permanent exclusion event was a contributing factor to these young people's offending career. In particular, 13 of these individuals began their criminal career in the same month that they were excluded. The authors found that for 132 individuals (i.e. 50% of the sample), permanent exclusion did not alter the trajectory of their offending career, as 47 young people had recorded offences before and after permanent exclusion and 85 had no recorded offences prior to or following exclusion.</p> <p>Daniels and Cole ([<reflink idref="bib9" id="ref29">9</reflink>]) attempted to track the trajectory of a non‐random sample of 141 permanently excluded pupils from a population of 480 drawn from 10 local education authorities in England. Interviews following permanent exclusion and follow‐up interviews 2 years later revealed that 55% had, or were believed to have, offended in the intervening time period. However, 39% of the sample were offenders prior to exclusion. This means that for the majority of the sample, there was no change in trajectory in relation to offending. McCrystal et al. ([<reflink idref="bib38" id="ref30">38</reflink>]) similarly tracked the drug use and anti‐social behaviour of permanently excluded pupils in the UK via four annual surveys from when participants were 11 years old until they were 15 years old. Over the 4 years, the authors found that permanently excluded pupils were more likely to use drugs than their non‐excluded counterparts, and that other anti‐social behaviour escalated from year two onwards.</p> <hd id="AN0186252690-4">Education and employment prospects</hd> <p>Daniels and Cole ([<reflink idref="bib9" id="ref31">9</reflink>]) tracked 141 excluded pupils and collected the academic attainment data for 91 (72%) of them. The authors found that 17 out of 91 (19%) permanently excluded pupils achieved one or more GCSEs at grade A–C. Via follow‐up interviews, the researchers noted conflicting responses. Although half of the (tracked) permanently excluded pupils viewed their exclusion as having a damaging effect on their life trajectory, an equal amount viewed the event as having had a positive effect on their lives by providing opportunities that were only made available to them as a result of exclusion. For example, Berridge et al. ([<reflink idref="bib2" id="ref32">2</reflink>]) found that nearly half of the permanently excluded young people they interviewed were either employed or in alternative education. Mainwaring and Hallam ([<reflink idref="bib35" id="ref33">35</reflink>]) asked a sample of 25 pupils (i.e. 16 excluded from mainstream school and nine current school pupils) to imagine their 'possible selves' in order to understand their aspirations and life perceptions. Not surprisingly, 100% of the currently schooled pupils were more likely to have positive 'possible selves', compared with just 69% of the permanently excluded pupils.</p> <hd id="AN0186252690-5">Disadvantageous pupil characteristics</hd> <p>Analysis of permanently excluded pupil demographics highlights that certain groups are disproportionately represented with regard to special education needs and disabilities, socio‐economic status, gender, ethnicity and prior attainment. As well as predicting permanent exclusion, these characteristics are associated with academic attainment, including GCSE attainment.</p> <hd id="AN0186252690-6">Special education needs and disabilities</hd> <p>Excluded pupils are disproportionally disadvantaged in terms of special education needs and disabilities. Krezmien et al. ([<reflink idref="bib28" id="ref34">28</reflink>]), using data collected in the state of Maryland in the United States, found that pupils with emotional disturbances and those with a learning disability were more likely to be excluded. Bowman‐Perrott et al. ([<reflink idref="bib4" id="ref35">4</reflink>]) extended the work of Achilles et al. ([<reflink idref="bib1" id="ref36">1</reflink>]), using a Special Education Elementary Longitudinal Study approach. The authors collected data from a sample of 1824 pupils with the following learning disabilities: emotional or behavioural disorder, attention‐deficit/hyperactivity disorder, learning disability or other health impairment. They found that pupils with emotional or behavioural disorder were more likely to be excluded than pupils with a learning disability. The authors also noted that pupils with attention‐deficit/hyperactivity disorder were more likely to be excluded than pupils with a learning disability. Using a nationally representative sample of 15,000 pupils with special education needs in England, Humphrey et al. ([<reflink idref="bib25" id="ref37">25</reflink>]) found that pupils with special education needs were at greater risk of poor academic attainment. Similarly, Velthuis et al. ([<reflink idref="bib53" id="ref38">53</reflink>]) found that only 24% of pupils with special education needs achieved a C or better in English and maths, which was the lowest percentage across all pupil characteristics measured. Other work follows this same track (see also Gross & McChrystal, [<reflink idref="bib22" id="ref39">22</reflink>]; Strand & Fletcher, [<reflink idref="bib50" id="ref40">50</reflink>]). The caveat here is that these studies were all limited in terms of the comparisons with those with no disabilities.</p> <hd id="AN0186252690-7">Socio‐economic status</hd> <p>Excluded pupils are disproportionally disadvantaged in terms of low socio‐economic status. Achilles et al. ([<reflink idref="bib1" id="ref41">1</reflink>]) found that socio‐economic status was negatively associated with likelihood of being excluded. Strand and Fletcher ([<reflink idref="bib50" id="ref42">50</reflink>]) found that pupils who were entitled to free school meals, which was used as a proxy for poverty, were more likely to be excluded than pupils not entitled to free school meals. Shuttleworth ([<reflink idref="bib48" id="ref43">48</reflink>]), Gorard ([<reflink idref="bib21" id="ref44">21</reflink>]) and Ilie et al. ([<reflink idref="bib26" id="ref45">26</reflink>]) found that free school meal eligibility is an indicator of deprivation and that it was associated with lower GCSE attainment. Farquharson et al. ([<reflink idref="bib17" id="ref46">17</reflink>]) showed that pupils not eligible for free school meals were three times more likely to achieve above the expected level of GCSE attainment.</p> <hd id="AN0186252690-8">Looked after</hd> <p>Strand and Fletcher ([<reflink idref="bib50" id="ref47">50</reflink>]) also found that pupils who had been 'looked‐after' by their local authority, such as living in a children's home, are more likely to be excluded. Furthermore, according to Harland ([<reflink idref="bib23" id="ref48">23</reflink>]), looked‐after pupils have significantly lower academic attainment than their non‐looked‐after counterparts. Harland ([<reflink idref="bib23" id="ref49">23</reflink>]) also found the lower academic attainment of looked‐after pupils was due to stereotyping, lower expectations and the environment of being 'in care'. Luke et al. ([<reflink idref="bib32" id="ref50">32</reflink>]) found that looked‐after pupils had lower GCSE attainment than their counterparts and Fletcher et al. ([<reflink idref="bib19" id="ref51">19</reflink>]) found that looked‐after pupils had worse educational experiences, including more school changes, lower attendance and more exclusion.</p> <hd id="AN0186252690-9">Gender</hd> <p>Males are consistently more likely to be excluded than females. This is evident in both the United States as well as England (see Achilles et al., [<reflink idref="bib1" id="ref52">1</reflink>]; Bowman‐Perrott et al., [<reflink idref="bib4" id="ref53">4</reflink>]; Strand & Fletcher, [<reflink idref="bib50" id="ref54">50</reflink>]). However, it should be noted that although males remain more likely to be excluded than females in England, there have been substantial increases in the number of girls being permanently excluded (Clarke, [<reflink idref="bib7" id="ref55">7</reflink>]). On average, the GCSE attainment of females is higher than that of males while the gender gap narrowed by nearly 4% during 2021–2024 (FFT Education Datalab, [<reflink idref="bib18" id="ref56">18</reflink>]). These data perhaps help explain the historical relationships between school attainment, socio‐economic status and gender in the United Kingdom from 2000 to 2012. Early et al. ([<reflink idref="bib16" id="ref57">16</reflink>]) concluded that gender was a statistically significant indicator of attainment.</p> <hd id="AN0186252690-10">Ethnicity</hd> <p>In the United States, Krezmien et al. ([<reflink idref="bib28" id="ref58">28</reflink>]), Achilles et al. ([<reflink idref="bib1" id="ref59">1</reflink>]) and Bowman‐Perrott et al. ([<reflink idref="bib4" id="ref60">4</reflink>]) all found that Black pupils were more likely to be excluded than White pupils. In England, Strand and Fletcher ([<reflink idref="bib50" id="ref61">50</reflink>]) found that Black Caribbean pupils were more likely be excluded than White British pupils. Demie ([<reflink idref="bib10" id="ref62">10</reflink>]) reported that Black Caribbean pupils were nearly four times more likely to receive a permanent exclusion than the school population as a whole, which were attributed to institutional racism, teachers' low expectations, a lack of diversity in school staff and lack of training in multicultural education. Similarly, via an ethnography of schools in an urban area of the South of England, Carlile ([<reflink idref="bib6" id="ref63">6</reflink>]) concluded that institutionalised racism played a role in the administration of permanent exclusions.</p> <p>Strand ([<reflink idref="bib49" id="ref64">49</reflink>]) found that the academic attainment gap between White British and several ethnic minority groups was large. Li ([<reflink idref="bib31" id="ref65">31</reflink>]) used the Longitudinal Study of Young People in England to show that ethnic minorities had lower GCSE attainment. Jackson ([<reflink idref="bib27" id="ref66">27</reflink>]) similarly found gross ethnic inequalities in GCSE attainment in England and Wales.</p> <hd id="AN0186252690-11">Language</hd> <p>Clegg et al. ([<reflink idref="bib8" id="ref67">8</reflink>]) found that language abilities was associated with risk of being permanently excluded. Of 15 pupils identified as at risk of being permanently excluded from a mainstream school located in an area of socio‐economic deprivation, 10 were found to have language difficulties, five of whom had significant and severe language difficulties. Strand et al. ([<reflink idref="bib51" id="ref68">51</reflink>]) identified an academic attainment gap between pupils who had English as a first language and pupils who had English as an additional language. This gap was found to be greater in reading than maths and decreased with age. Demie and Strand ([<reflink idref="bib11" id="ref69">11</reflink>]) found a strong relationship between fluency in English and GCSE attainment, whereby pupils with a low level of fluency performed at very low levels, but fully fluent bi‐lingual pupils had higher levels of GCSE attainment then English‐only speakers.</p> <hd id="AN0186252690-12">Prior academic attainment</hd> <p>Strand and Fletcher ([<reflink idref="bib50" id="ref70">50</reflink>]) found a strong negative relationship between the Key Stage‐2 English test at the end of primary school and school exclusion rate whereby the greater the score on the English test, the lower the likelihood of exclusion. Lessof et al. ([<reflink idref="bib30" id="ref71">30</reflink>]) found that prior attainment at Key Stage‐2 was a significant predictor of subsequent GCSE attainment.</p> <p>What this literature suggests is that the concentration is on the United Kingdom and United States, despite no context restrictions being used when undertaking the literature search. Research in other contexts, in particular developing countries, focuses on drop outs, rather than exclusions, where the leaving of school is the decision of the pupil or family rather than the school itself (Wong, [<reflink idref="bib54" id="ref72">54</reflink>]). Given this difference, the drop‐out literature was not considered sufficiently relevant to inform this study's conceptual framework and was therefore omitted from inclusion or discussion.</p> <hd id="AN0186252690-13">Hypothesis development</hd> <p>Our study adds to the literature regarding the negative consequences of permanent exclusion. To do so, we specifically considered the implications of permanent exclusion, rather than fixed‐term exclusions. We took this track owing to consensus in the literature that pupils who have been permanently excluded from school are at a far greater risk of a variety of negative outcomes than pupils who have not been excluded (Pirrie et al., [<reflink idref="bib44" id="ref73">44</reflink>]). Of the studies that considered education outcomes of permanently excluded pupils, none have compared permanently excluded and non‐permanently excluded pupils in terms of academic attainment. Although Daniels and Cole ([<reflink idref="bib9" id="ref74">9</reflink>]) tracked permanently excluded pupils to their GCSEs, they failed to compare their sample in order to establish a counterfactual. And although Mainwaring and Hallam ([<reflink idref="bib35" id="ref75">35</reflink>]) compared permanently excluded and non‐permanently excluded pupils, their sample size was qualitatively and structurally limited. Accordingly, our is among the first to consider the influence of permanent exclusion on a measure of academic attainment by comparing permanently excluded pupils with the counterfactual of pupils that have not been permanently excluded.</p> <p>Specifically, like Daniels and Cole ([<reflink idref="bib9" id="ref76">9</reflink>]), our study measured the influence of permanent exclusion on GCSE attainment. And whereas the authors measured pupils involved in education or training and the range of GCSEs studied, our approach measured GCSE at a more granular level in terms of the discrete variable of capped GCSE points. Performance in these exams has been found to be 'critical to future trajectories and life chances' (Lupton et al., [<reflink idref="bib33" id="ref77">33</reflink>], p. 3) as they are the gateway to continued education. For example, Lupton et al. ([<reflink idref="bib33" id="ref78">33</reflink>]) found that lower attainers often feel like failures and feel less confident about their futures. The authors also noted the lower attainers were overlooked and underserved in terms of career advice, guidance and social support. The authors further noted that even apprenticeships, which are assumed to be a more accessible pathway for lower attainers, were not a common destination for low attainers. Parsons and Elliot Major ([<reflink idref="bib43" id="ref79">43</reflink>]) found that pupils who failed to get basic GCSE grades were less likely to be in education, employment or training, and were less likely to expect to go to university or to have professional or managerial aspirations.</p> <p>GCSE attainment has far‐reaching consequences beyond just future eductaional attainment. For example, Sabates ([<reflink idref="bib47" id="ref80">47</reflink>]) found that low GCSE attainment was associated with juvenile crime (e.g. burglary, theft, crimial damage and drug‐related offences). Parsons and Elliot Major ([<reflink idref="bib43" id="ref81">43</reflink>]) noted that pupils who did not achieve basic GCSE grades were more likely to be in poor or fair health or to have a longstanding illness. The authors also noted that lower GCSE achievers are more likely to smoke or vape, take drugs, be stopped and questioned or formally cautioned by the police, to engage in underage sex and to attempt suicide. To compare the GCSE attainment of permanently excluded and non‐permanently excluded pupils the following hypothesis was tested:</p> <hd id="AN0186252690-14">1 Hypothesis</hd> <p>GCSE attainment between permanently excluded pupils and non‐permanently excluded pupils will significantly differ.</p> <p>A causal relationship is most effectively achieved by an experimental design, whereby there is control over variables and the environment (Burtless, [<reflink idref="bib5" id="ref82">5</reflink>]). However, randomly assigning pupils to be permanently excluded from school is both unethical and impossible. Although adjustment methods cannot remove selection bias for unobserved differences, they can be used to remove selection bias for observed differences (Burtless, [<reflink idref="bib5" id="ref83">5</reflink>]). A comparison of permanently excluded and non‐permanently excluded pupils was undertaken by Mainwaring and Hallam ([<reflink idref="bib35" id="ref84">35</reflink>]) upon 'possible selves' and by McCrystal et al. ([<reflink idref="bib38" id="ref85">38</reflink>]) upon drug use and anti‐social behaviour. However, neither made use of adjustment methods to limit selection bias. This means that the influence of permanent exclusion was over‐estimated because confounding variables were not controlled for.</p> <p>Our study is the first to measure any consequence of permanent exclusion by using adjustment methods to control for selection bias. Specifically, previous research into predictors of permanent exclusion identified several disadvantageous pupil characteristics as potentially confounding variables. In order to properly examine the influence of these factors, we held them constant in a multiple linear regression model. Doing so should provide a measurement of the influence of permanent exclusion adjusted for disadvantageous pupil characteristics that predict both permanent exclusion and GCSE attainment. To achieve this end, the following hypothesis was tested:</p> <hd id="AN0186252690-15">2 Hypothesis</hd> <p>The difference in GCSE attainment between permanently excluded pupils and non‐permanently excluded pupils will decrease when disadvantageous pupil characteristics that predict both exclusion and academic attainment are held constant.</p> <p>In order to add to the debate on whether permanently excluded pupils should be supported rather than blamed and punished (see MacRae et al., [<reflink idref="bib34" id="ref86">34</reflink>]; Parsons, [<reflink idref="bib41" id="ref87">41</reflink>]), our study assessed the relative influence of permanent exclusion and disadvantageous pupil characteristics. Whereas previous research has either measured the disadvantages that predict permanent exclusion or the disadvantages that follow permanent exclusion, our study was the first to bring these two lines of research together in a single piece of research. The preceding commentary underscores the importance of two measures of the GCSE attainment gap between permanently excluded pupils and non‐permanently excluded pupils: (<reflink idref="bib1" id="ref88">1</reflink>) an unadjusted measure; and (<reflink idref="bib2" id="ref89">2</reflink>) an adjusted measure controlling for disadvantageous pupil characteristics. By calculating a ratio of these two measures, our study assessed the the relative influence of being permanently excluded vs. disadvantageous pupil characteristics.</p> <p>In order to reduce causal inference limitations by making use of a counterfactual and controlling for certain confounding variables, our study curtailed the population inference limitations seen in existing research. For example, Hodgson and Webb ([<reflink idref="bib24" id="ref90">24</reflink>]) interviewed only 56 permanently excluded pupils when drawing their conclusions; Berridge et al. ([<reflink idref="bib2" id="ref91">2</reflink>]) tracked a non‐random sample of only 263 exlucded pupils; Daniels and Cole ([<reflink idref="bib9" id="ref92">9</reflink>]) tracked a non‐random sample of 141 permanently excluded pupils; McCrystal et al. ([<reflink idref="bib38" id="ref93">38</reflink>]) compared a non‐random sample of 51 permanently excluded pupils with approximately 4000 non‐permanently excluded pupils; Mainwaring and Hallam ([<reflink idref="bib35" id="ref94">35</reflink>]) compared a non‐random sample of only 16 permanently excluded pupils and nine non‐permanently excluded pupils. What these studies have in common is the use of small and non‐random samples, which limit the generalisability of existing knowledge on the negative consequences of permanent exclusion. In contrast, Strand and Fletcher ([<reflink idref="bib50" id="ref95">50</reflink>]) made use of the National Pupil Database to access a population of a whole cohort of over 500,000 pupils in England so to establish highly externally valid predictors of exclusion. Like Strand and Fletcher ([<reflink idref="bib50" id="ref96">50</reflink>]), our study will access a population of a whole cohort of pupils in England through the National Pupil Database in order to establish a highly externally valid consequence of permanent exclusion.</p> <hd id="AN0186252690-16">METHOD</hd> <p></p> <hd id="AN0186252690-17">Sample</hd> <p>The data for our study comes from England's National Pupil Database, which is controlled by the Department for Education and contains data on all pupils in state‐funded education. The database is sourced primarily by returns from schools that are provided three times a year by the school census. The data is then matched using names, dates of birth and other factors. In 2017/2018, 7905 pupils were permanently excluded from mainstream schools in England, the largest group being 2040 who were in year 10—the year before they were due to take their GCSE examinations (Department for Education, [<reflink idref="bib12" id="ref97">12</reflink>]). Our study tracks these students to their GCSE examinations the following year—2018/2019 when these pupils were then in year 11 of their schooling (i.e. the final year of compulsory education when pupils take their GCSE examinations)—comparing them with those non‐permanently excluded. The total population consisted of 590,092 pupils for whom data on all variables were available. The dataset included 1490 permanently excluded pupils (0.3%) who were compared with 588,602 pupils who had not been permanently excluded (99.7%). The sample was evenly distributed between females (49.1%) and males (50.9%). The majority of pupils were in the second (39.3%) and third quartiles (38.1%) of prior academic attainment. The majority of pupils (68.1%) were not of an ethnic minority (68.1%) and had English as a first language (84.9%). Less than 1% had been looked after by a local authority but nearly a quarter were eligible for free school meals (23.5%) and over one‐fifth had identified special needs (21.3%).</p> <hd id="AN0186252690-18">Variables</hd> <p></p> <hd id="AN0186252690-19">Dependent variable</hd> <p>Our study sought to establish the influence of being permanently excluded on academic attainment, for which results from GCSE examinations are being used. Pupils are required to take a minimum of five subject but most pupils take nine subjects, which must include English, maths and science. Each GCSE is marked on a scale of '1' to '9' as well as a grade of 'U', which means 'ungraded'. A pupil's capped GCSE score is the cumulative score of their English, maths and science GCSEs plus their next best six subjects, for a cumulative total GCSE score. A perfect performance of nine GCSEs graded as 9e would accrue a maximum score of 81. The capped score, rather than uncapped score, was used as a control for the number of GCSE subjects taken. In other words, this was done so that a high score is more representative of quality of educational attainment than the quantity of subjects taken. The discrete data were taken from the National Pupil Database's Key Stage Four Data Table for 2018–2019, which was restricted to the sample of year 11. The dataset does not include results from compulsory resits after year 11 such that GCSE scores in this study are based upon just the single and first attempt.</p> <hd id="AN0186252690-20">Independent variable</hd> <p>The independent variable of interest was permanent exclusion. For consistency and controlling for time, this variable was operationalised as a pupil having been permanently excluded in year 10 of the 2017–2018 school year. This variable was selected because it is the school year when the most permanent exclusions occurred (Department for Education, [<reflink idref="bib12" id="ref98">12</reflink>]). This dichotomous measure was taken from the National Pupil Database's Exclusions Data Table for 2017–2018, which was restricted to cases of year 10 pupils being permanently excluded. Given that the data were drawn from a separate data table to the National Pupil Database's Key Stage Four Data Table for 2018–2019, from which the dependent variable measure was taken, it was necessary to use the pupil matching reference number to match the data and code each pupil in terms of having been permanently excluded in year 10 (yes = 1; no = 0).</p> <hd id="AN0186252690-21">Control variables</hd> <p>Pupil characteristics available in the National Pupil Database were controlled for if established in the literature as being associated with both the independent variable of permanent exclusion and the dependent variable of academic attainment, so to avoid spurious associations. Gender is associated with both exclusions (Achilles et al., [<reflink idref="bib1" id="ref99">1</reflink>]; Bowman‐Perrott et al., [<reflink idref="bib4" id="ref100">4</reflink>]; Strand & Fletcher, [<reflink idref="bib50" id="ref101">50</reflink>]) and academic attainment (Early et al., [<reflink idref="bib16" id="ref102">16</reflink>]; FFT Education Datalab, [<reflink idref="bib18" id="ref103">18</reflink>]). Gender was available as a dichotomous variable (coded as male = 1; female = 0) in the National Pupil Database. Prior academic attainment is associated with both exclusions (Strand & Fletcher, [<reflink idref="bib50" id="ref104">50</reflink>]) and subsequent academic attainment (Lessof et al., [<reflink idref="bib30" id="ref105">30</reflink>]). The National Pupil Database recorded prior academic attainment at the end of year 6 (i.e. the end of key stage 2) based upon the SAT school‐administered tests for reading, writing, and maths, which was coded a discrete variable from 1 to 4. Ethnicity is associated with both exclusions (Achilles et al., [<reflink idref="bib1" id="ref106">1</reflink>]; Bowman‐Perrott et al., [<reflink idref="bib4" id="ref107">4</reflink>]; Carlile, [<reflink idref="bib6" id="ref108">6</reflink>]; Demie, [<reflink idref="bib10" id="ref109">10</reflink>], Krezmien et al., [<reflink idref="bib28" id="ref110">28</reflink>]; Strand & Fletcher, [<reflink idref="bib50" id="ref111">50</reflink>]) and academic attainment (Jackson, [<reflink idref="bib27" id="ref112">27</reflink>]; Strand, [<reflink idref="bib49" id="ref113">49</reflink>]). From a long list of ethnicity categories recorded in the National Pupil Database, ethnicity was recoded as a dichotomous variable with all White ethnicities (i.e. non‐ethnic minorities) coded as 0 and all other ethnicities (i.e. ethnic minorities) coded as 1. Language ability is associated with both exclusions (Clegg et al., [<reflink idref="bib8" id="ref114">8</reflink>]) and academic attainment (Demie & Strand, [<reflink idref="bib11" id="ref115">11</reflink>]; Strand et al., [<reflink idref="bib51" id="ref116">51</reflink>]). Whether English was a pupil's first language was available as a dichotomous variable (yes = 1; no = 0) in the National Pupil Database. Looked‐after status is associated with both exclusions (Strand & Fletcher, [<reflink idref="bib50" id="ref117">50</reflink>]) and academic attainment (Fletcher et al., [<reflink idref="bib19" id="ref118">19</reflink>]; Harland, [<reflink idref="bib23" id="ref119">23</reflink>]; Luke et al., [<reflink idref="bib32" id="ref120">32</reflink>]). In the context of the National Pupil Database, pupils are classified as looked‐after if they have been in the care of their local authority for 1 day or more during 2018–2019, which was available as a dichotomous variable (yes = 1; no = 0). Socio‐economic status is associated with both exclusions (Achilles et al., [<reflink idref="bib1" id="ref121">1</reflink>]; Strand & Fletcher, [<reflink idref="bib50" id="ref122">50</reflink>]) and academic attainment (Farquharson et al., [<reflink idref="bib17" id="ref123">17</reflink>]; Gorard, [<reflink idref="bib21" id="ref124">21</reflink>]; Ilie et al., [<reflink idref="bib26" id="ref125">26</reflink>]; Shuttleworth, [<reflink idref="bib48" id="ref126">48</reflink>]). A measure of socio‐economic status available in the National Pupil Database is eligibility for free school meals in the previous 6 years, which was available as a dichotomous variable (yes = 1; no = 0). Having special education needs is associated with both exclusions (Achilles et al., [<reflink idref="bib1" id="ref127">1</reflink>]; Bowman‐Perrott et al., [<reflink idref="bib4" id="ref128">4</reflink>]; Krezmien et al., [<reflink idref="bib28" id="ref129">28</reflink>]; Strand & Fletcher, [<reflink idref="bib50" id="ref130">50</reflink>]) and academic attainment (Humphrey et al., [<reflink idref="bib25" id="ref131">25</reflink>]; Velthuis et al., [<reflink idref="bib53" id="ref132">53</reflink>]). Having a special education need was available in the National Pupil Database as a dichotomous variable (coded as yes = 1; no = 0). Data for all control variables were taken from the National Pupil Database's Key Stage Four Data Table for 2018–2019 (Tables 1 and 2).</p> <p>1 TABLE Participant characteristics.</p> <p> <ephtml> <table><thead valign="bottom"><tr><th align="left" /><th align="left"><italic>n</italic> (%)</th></tr></thead><tbody valign="top"><tr><td align="left">Permanently excluded</td></tr><tr><td align="left">No</td><td align="left">588,602 (99.7%)</td></tr><tr><td align="left">Yes</td><td align="left">1490 (0.3%)</td></tr><tr><td align="left">Gender</td></tr><tr><td align="left">Female</td><td align="left">289,843 (49.1%)</td></tr><tr><td align="left">Male</td><td align="left">300,249 (50.9%)</td></tr><tr><td align="left">Prior attainment</td></tr><tr><td align="left">One</td><td align="left">55,045 (9.3%)</td></tr><tr><td align="left">Two</td><td align="left">231,845 (39.3%)</td></tr><tr><td align="left">Three</td><td align="left">224,997 (38.1%)</td></tr><tr><td align="left">Four</td><td align="left">78,205 (13.3%)</td></tr><tr><td align="left">Ethnic minority</td></tr><tr><td align="left">No</td><td align="left">402,105 (68.1%)</td></tr><tr><td align="left">Yes</td><td align="left">187,987 (31.9%)</td></tr><tr><td align="left">English as first language</td></tr><tr><td align="left">No</td><td align="left">88,941 (15.1%)</td></tr><tr><td align="left">Yes</td><td align="left">501,151 (84.9%)</td></tr><tr><td align="left">Looked after</td></tr><tr><td align="left">No</td><td align="left">584,734 (99.1%)</td></tr><tr><td align="left">Yes</td><td align="left">5358 (0.9%)</td></tr><tr><td align="left">Free school meals</td></tr><tr><td align="left">No</td><td align="left">451,514 (76.5%)</td></tr><tr><td align="left">Yes</td><td align="left">138,578 (23.5%)</td></tr><tr><td align="left">Special education needs</td></tr><tr><td align="left">No</td><td align="left">464,396 (78.7%)</td></tr><tr><td align="left">Yes</td><td align="left">125,696 (21.3%)</td></tr><tr><td align="left">Total</td><td align="left">590,092 (100%)</td></tr></tbody></table> </ephtml> </p> <p>2 TABLE Linear regression for academic attainment.</p> <p> <ephtml> <table><thead valign="bottom"><tr><th align="left" /><th align="left"><italic>β</italic></th><th align="left">SE</th><th align="left"><italic>t‐</italic>Value</th><th align="left"><italic>p‐</italic>Value</th></tr></thead><tbody valign="top"><tr><td align="left">Constant</td><td align="left">20.55</td><td align="left">0.07</td><td align="left">298.89</td><td align="left"><0.001</td></tr><tr><td align="left">Permanently excluded (vs. not)</td><td align="left">−12.49</td><td align="left">0.37</td><td align="left">−33.53</td><td align="left"><0.001</td></tr><tr><td align="left">Male (vs. female)</td><td align="left">−3.97</td><td align="left">0.04</td><td align="left">−106.06</td><td align="left"><0.001</td></tr><tr><td align="left">Prior attainment (1–4)</td><td align="left">7.67</td><td align="left">0.02</td><td align="left">318.18</td><td align="left"><0.001</td></tr><tr><td align="left">Ethnic minority (vs. not)</td><td align="left">−0.35</td><td align="left">0.05</td><td align="left">−7.28</td><td align="left"><0.001</td></tr><tr><td align="left">English as first language (vs. not)</td><td align="left">1.08</td><td align="left">0.06</td><td align="left">18.35</td><td align="left"><0.001</td></tr><tr><td align="left">Looked after (vs. not)</td><td align="left">−9.28</td><td align="left">0.2</td><td align="left">−46.98</td><td align="left"><0.001</td></tr><tr><td align="left">Free school meals (vs. not)</td><td align="left">−6.3</td><td align="left">0.05</td><td align="left">−137.19</td><td align="left"><0.001</td></tr><tr><td align="left">Special education needs (vs. not)</td><td align="left">−14.26</td><td align="left">0.05</td><td align="left">−293.12</td><td align="left"><0.001</td></tr></tbody></table> </ephtml> </p> <hd id="AN0186252690-22">Analytic approach</hd> <p>Using Statistical Package for the Social Sciences, our analytical procedure began with a comparison of mean averages of capped GCSE points between non‐permanently excluded pupils and permanently excluded pupils. To accomplish this, we used an independent‐samples <emph>t</emph>‐test to provide an unadjusted measurement of the GCSE attainment gap between permanently excluded pupils and non‐permanently excluded pupils. Next, a multiple linear regression model was constructed to measure the influence of the independent variable of permanent exclusion on the dependent variable of capped GCSE points (i.e. holding constant the control variables of prior academic attainment, gender, ethnicity, first language, looked‐after status, free school meal eligibility and having a special education need). Holding these measures constant provided an adjusted measurement of the GCSE attainment gap between permanently excluded pupils and non‐permanently excluded pupils. The multiple linear regression formula is expressed as follows: <ephtml> <math altimg="urn:x-wiley:01411926:media:berj4133:berj4133-math-0001" display="block" overflow="scroll" xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow><msub><mi>Y</mi><mi>i</mi></msub><mo linebreak="goodbreak">=</mo><mi>α</mi><mo linebreak="goodbreak">+</mo><msub><mi>β</mi><mn>1</mn></msub><msub><mi>X</mi><mrow><mi>i</mi><mn>1</mn></mrow></msub><mo linebreak="goodbreak">+</mo><mo>...</mo><mo linebreak="goodbreak">+</mo><msub><mi>β</mi><mi>k</mi></msub><msub><mi>X</mi><mi mathvariant="italic">ik</mi></msub><mo linebreak="goodbreak">+</mo><msub><mi>ε</mi><mi>i</mi></msub></mrow></semantics></math> </ephtml> where <emph>Y</emph> is the outcome variable, <ephtml> <math altimg="urn:x-wiley:01411926:media:berj4133:berj4133-math-0002" display="inline" overflow="scroll" xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow><msub><mi>X</mi><mi>k</mi></msub></mrow></semantics></math> </ephtml> refers to the explanatory variables, <emph>α</emph> is the intercept, <ephtml> <math altimg="urn:x-wiley:01411926:media:berj4133:berj4133-math-0003" display="inline" overflow="scroll" xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow><msub><mi>β</mi><mi>k</mi></msub></mrow></semantics></math> </ephtml> is the regression coefficient for the variable <emph>k</emph> and <emph>ε</emph> accounts for the residual. Lastly, the unadjusted and adjusted GCSE attainment gaps were compared in terms of a ratio (i.e. the coefficient of the independent variable of permanent exclusion from the multiple linear regression was divided by the mean difference in the independent‐samples <emph>t</emph>‐test analysis). This was done in order to determine the extent of the GCSE attainment gap that could be attributed to being permanently excluded and the extent to which ot could be attributed to prior academic attainment and disadvantageous pupil characteristics.</p> <hd id="AN0186252690-23">RESULTS</hd> <p>The results of the analyses indicate that permanently excluded pupils had statistically significantly lower capped GCSE points (8.94 ± 9.84) when compared with non‐permanently excluded pupils (33.58 ± 17.24), a mean difference of 24.64, <emph>t</emph>(1512.21) = 96.22, <emph>p</emph> < 0.001. In other words, non‐permanently excluded pupils scored 3.76 times more GCSE points than permanently excluded pupils. Therefore, Hypothesis 1 that GCSE attainment between permanently excluded pupils and non‐permanently excluded pupils will significantly differ was accepted. When holding constant disadvantageous pupil characteristics this difference approximately halved—on average, permanently excluded pupils achieved 12.49 fewer capped GCSE points than non‐permanently excluded pupils (<emph>β</emph> = −12.49, <emph>p</emph> ≤ 0.001). In other words, when holding constant the <emph>a priori</emph> factors in this study (i.e. prior attainment, gender, ethnicity, English as a first language, looked‐after status, free school meals and special needs), the GCSE attainment gap between permanently excluded pupils and non‐permanently excluded decreased from 26.64 to 12.49. Therefore, Hypothesis 2, that the difference in GCSE attainment between permanently excluded pupils and non‐permanently excluded pupils will decrease when disadvantageous pupil characteristics that predict both exclusion and academic attainment are held constant, was also accepted.</p> <p>The control variables were all associated with GCSE attainment. Prior attainment was significantly positively associated with capped GCSE points (<emph>β</emph> = 17.67, <emph>p</emph> ≤ 0.001). Being registered as having special needs (vs. not) was significantly negatively associated with capped GCSE points (<emph>β</emph> = −14.26, <emph>p</emph> ≤ 0.001). Being of looked‐after status (vs. not) was significantly negatively associated with capped GCSE points (<emph>β</emph> = −9.28, <emph>p</emph> ≤ 0.001). Being eligible for free school meals (vs. not) was significantly negatively associated with capped GCSE points (<emph>β</emph> = −6.3, <emph>p</emph> ≤ 0.001). Being male (vs. being female) was significantly negatively associated with capped GCSE points (<emph>β</emph> = −3.97, <emph>p</emph> ≤ 0.001). English being a pupil's first language (vs. not) was significantly positively associated with capped GCSE points (<emph>β</emph> = 1.08, <emph>p</emph> ≤ 0.001). Being of an ethnic minority (vs. not) was significantly negatively associated with capped GCSE points (<emph>β</emph> = −0.35, <emph>p</emph> ≤ 0.001). As well as all variables being individually predictive of capped GCSE points, the model significantly predicted capped GCSE points, <emph>F</emph>(<reflink idref="bib8" id="ref133">8</reflink>, 590,083) = 33462.64.393, <emph>p</emph> < 0.001, <emph>R</emph><sups>2</sups> = 0.312. Approximately 32% of variance in capped GCSE points is explained by the model.</p> <hd id="AN0186252690-24">DISCUSSION</hd> <p>Prior research has over‐estimated the influence of the event of permanent exclusion on subsequent negative consequences by failing to control for existing disadvantages (Berridge et al., [<reflink idref="bib2" id="ref134">2</reflink>]; Daniels & Cole, [<reflink idref="bib9" id="ref135">9</reflink>]; Mainwaring & Hallam [<reflink idref="bib35" id="ref136">35</reflink>]; McCrystal et al., [<reflink idref="bib38" id="ref137">38</reflink>]; Pritchard & Cox, [<reflink idref="bib46" id="ref138">46</reflink>]). Our study provides a more refined and appropriate approach to measure the influence of permanent exclusion to overcome the limitations of prior work. We found that only half of the GCSE attainment gap between permanently excluded pupils and their non‐excluded peers can be attributed to permanent exclusion. This result suggests that permanent exclusion is neither the catalyst of disadvantage nor a continuation of disadvantage on the same trajectory. Rather, permanent exclusion can more accurately be understood as an accentuation of existing disadvantages whereby the existing trajectory of disadvantage becomes steeper following permanent exclusion. In other words, permanent exclusion compounds academic attainment deficiencies caused by disadvantages associated with gender, prior attainment, ethnicity, language, being looked after, poverty and special education needs.</p> <hd id="AN0186252690-25">Policy implications</hd> <p>Our study found that non‐permanently excluded pupils achieve 3.76 times more GCSE points than permanently excluded pupils, but that this difference halved when controlling for disadvantageous pupil characteristics. In other words, half of the GCSE attainment gap can be attributed to being permanently excluded and half of the attainment gap can be attributed to previous academic attainment and disadvantageous pupil characteristics. Therefore, taking the approach taken by schools in Scotland to not permanently exclude will not completely solve the GCSE attainment gap. These findings suggest that pupils in Scotland who would have been permanently excluded in England, but were not in Scotland, would still experience significantly lower GCSE attainment than their counterparts because not excluding a pupil does not remove their disadvantage (McCluskey et al., [<reflink idref="bib37" id="ref139">37</reflink>]). As such, policymakers and schools need to address the GCSE attainment gap between permanently excluded and non‐permanently excluded pupils on two fronts.</p> <p>First, in order to reduce the e gap caused by permanent exclusion, the practice should be more of a last resort. And a decision to permanently exclude should be taken in the interests of the pupil and not in the interests of the school's performance in academic league tables. Currently, England's education system is measured and funded on the basis of academic league tables, which according to Thompson et al. ([<reflink idref="bib52" id="ref140">52</reflink>]) serves as a perverse incentive when it comes to permanent exclusions that peak in the year ahead of examinations. According to McShane ([<reflink idref="bib39" id="ref141">39</reflink>]), exclusion incentivises schools to engage in the practice of off‐rolling (i.e. the removal of pupils from the school roll via unofficial means). To counter performance‐driven exclusions and off‐rolling, the results of previously permanently excluded pupils should be included in a school's data for the academic league tables.</p> <p>Second, in order to reduce the element of the gap caused by previous academic attainment and disadvantageous pupil characteristics, the chronic underfunding of the inclusion agenda should to be reversed (Done & Knowler, [<reflink idref="bib14" id="ref142">14</reflink>]). Furthermore, greater understanding of permanently excluded pupils is needed so to remove the stigma associated with being permanently excluded. Rather than blaming and punishing permanently excluded pupils (MacRae et al., [<reflink idref="bib34" id="ref143">34</reflink>]; Parsons, [<reflink idref="bib41" id="ref144">41</reflink>]), we instead argue that schools should support them in overcoming previous low academic attainment and disadvantageous characteristics. As reported by Murphy ([<reflink idref="bib40" id="ref145">40</reflink>]), permanently excluded pupils perceived that schools had misread their special educational needs as misbehaviour and non‐compliance. Further, the authors suggest that such behaviours were the school's way of communicating personal and social problems, which were amplified by punitive measures. We suggest that before considering permanent exclusion, mainstream schools should work with parents and other agencies to address the social and cultural factors disadvantaging pupils. However, this approach should only be taken if these factors cannot be addressed within the pupil's original mainstream school, and only if it is in the best interest of the pupil should alternative provision be sought, such that zero permanent exclusions is not the solution. This is because it does not remove disadvantages and will only encourage schools to make use of hidden forms of exclusion to achieve zero exclusions in official statistics (Power & Taylor, [<reflink idref="bib45" id="ref146">45</reflink>]). For some pupils, exclusion can serve a corrective function and other mainstream schools or alternative provision may offer a clean start with greater support for disadvantageous pupil characteristics. Keeping the subsequent GCSE performance of the pupil on the books of the originating school incentivises everyone to act in the best interests of the pupil.</p> <hd id="AN0186252690-26">Study limitations</hd> <p>Our study was only able to track 1490 pupils out of 2040 (73%) pupils who were permanently excluded from year 10 of state‐funded schools in England, such that there is a missing data issue. The permanently excluded pupils that cannot be tracked in these instances are likely to be different from those that can be tracked, as such creating a type of selection bias in these samples. In the case of our study, there were 550 permanently excluded pupils from year 10 in 2017–2018 who were not found in the National Pupil Database's Key Stage Four Data Table for 2018–2019. This means that it is unknown what happened to these pupils and their GCSE attainment.</p> <p>Some of the permanently excluded pupils that could not be tracked will have achieved zero GCSE attainment, such that the non‐adjusted difference in GCSE attainment between permanently excluded pupils and non‐permanently excluded pupils is even larger than reported. Other permanently excluded pupils may have continued their education in independent schools, personal tutoring or home schooling which the state or local authority do not fund. In which case, their data would not be included the National Pupil Database even though they may have attained GCSE points. This missing data problem could have been reduced by using GCSE results from 2019/2020 in order to capture pupils who sat their GCSEs a year later. Accordingly, missing data would have introduced bias and inconsistency in the dependent variable as GCSE marks were inflated by being based upon teacher's predicted grades in 2019/2020, rather than examination results as in 2018/2019. Whatever the extent of sampling bias in the data, it is consistent in both the unadjusted (i.e. means comparison) and adjusted (i.e. multiple linear regression) analyses being compared as both make use of the same sample.</p> <p>A further limitation of this study is condition contamination caused by off‐rolling, which is a form of unofficial exclusion using unofficial means (McShane, [<reflink idref="bib39" id="ref147">39</reflink>]). Some pupils will have been included in the data but owing to the use of unofficial means to exclude, these pupils will have been coded as non‐permanently excluded. The extent of off‐rolling is highlighted by Gill et al. ([<reflink idref="bib20" id="ref148">20</reflink>]), who contrasted 6685 officially recorded exclusions in 2016 against 48,000 pupils registered in alternative provision (AP) settings. Future research could again use data from the National Pupil Database to compare GCSE attainment of pupils in mainstream schools vs. pupils in AP, in order to capture the pupils who end up in AP through either permanent exclusion or off‐rolling. Future research should consider if these findings hold during pandemic years and post‐pandemic years to understand if the pandemic had an exacerbating influence on the GCSE attainment gap between permanently and non‐permanently excluded pupils and also if the relative influence of permanent exclusion or disadvantageous pupil characteristics altered at all.</p> <p>A final study limitation is that the <emph>R</emph><sups>2</sups> for the model produced to test the second hypothesis is 0.312, such that 68.8% of variance is explained by variables not in this model. In other words, there is omitted variable bias in predicting capped GCSE points and factors beyond permanent exclusion and disadvantageous pupil characteristics play an important role in predicting capped GCSE points. Although this study focused on disadvantageous pupil characteristics, future research could consider other groups of variables, such as pupil behaviour, including absences from school. However, caution should be taken to ensure control variables are distinct from the independent variable of permanent exclusion. For example, absences in the year of a permanent exclusion may be a symptom, rather than a predictor, of exclusion. Furthermore, it should be considered that the data collected for the National Pupil Database is evolving and a limitation of this study is that scale data for a composite prior academic attainment measure has only been available since 2020–2021 and may have been more effective in modelling capped GCSE points than the less granular data used in this study. Future research should make use of this continuous data.</p> <hd id="AN0186252690-27">FUNDING INFORMATION</hd> <p>This research received no specific grant from any funding agency in the public, commercial or not‐for‐profit sectors.</p> <hd id="AN0186252690-28">CONFLICT OF INTEREST STATEMENT</hd> <p>The authors declare that there are no conflicts of interest regarding the publication of this paper.</p> <hd id="AN0186252690-29">DATA AVAILABILITY STATEMENT</hd> <p>Research data are not shared.</p> <hd id="AN0186252690-30">ETHICS STATEMENT</hd> <p>This study was conducted in accordance with the BERA's Ethical Guidelines for Educational Research and the ethical standards of the London Metropolitan University. Ethics approval was obtained from the London Metropolitan University Ethics Committee prior to the commencement of the research.</p> <ref id="AN0186252690-31"> <title> REFERENCES </title> <blist> <bibl id="bib1" idref="ref20" type="bt">1</bibl> <bibtext> Achilles, G. M., McLaughlin, M. J., & Croninger, R. G. (2007). Sociocultural correlates of disciplinary exclusion among students with emotional, behavioral, and learning disabilities in the SEELS national dataset. Journal of Emotional and Behavioral Disorders, 15 (1), 33 – 45.</bibtext> </blist> <blist> <bibl id="bib2" idref="ref6" type="bt">2</bibl> <bibtext> Berridge, D., Brodie, I., Pitts, J., Porteous, D., & Tarling, R. (2001). The independent effects of permanent exclusion from school on the offending careers of young people. Great Britain, Home Office, Research, Development and Statistics Directorate.</bibtext> </blist> <blist> <bibl id="bib3" idref="ref3" type="bt">3</bibl> <bibtext> Blyth, E., & Milner, J. (1993). Exclusion from school: A first step in exclusion from society? Children & Society, 7 (3), 255 – 268.</bibtext> </blist> <blist> <bibl id="bib4" idref="ref16" type="bt">4</bibl> <bibtext> Bowman‐Perrott, L., Benz, M. R., Hsu, H. Y., Kwok, O. M., Eisterhold, L. A., & Zhang, D. (2013). Patterns and predictors of disciplinary exclusion over time: An analysis of the SEELS national data set. Journal of Emotional and Behavioral Disorders, 21 (2), 83 – 96.</bibtext> </blist> <blist> <bibl id="bib5" idref="ref82" type="bt">5</bibl> <bibtext> Burtless, G. (1995). The case for randomized field trials in economic and policy research. Journal of Economic Perspectives, 9 (2), 63 – 84.</bibtext> </blist> <blist> <bibl id="bib6" idref="ref63" type="bt">6</bibl> <bibtext> Carlile, A. (2012). An ethnography of permanent exclusion from school: Revealing and untangling the threads of institutionalised racism. Race Ethnicity and Education, 15 (2), 175 – 194.</bibtext> </blist> <blist> <bibl id="bib7" idref="ref55" type="bt">7</bibl> <bibtext> Clarke, E. (2024). Voices from the edge: Girls' experiences of being at risk of permanent exclusion. British Educational Research Journal, 50 (2), 855 – 875.</bibtext> </blist> <blist> <bibl id="bib8" idref="ref67" type="bt">8</bibl> <bibtext> Clegg, J., Stackhouse, J., Finch, K., Murphy, C., & Nicholls, S. (2009). Language abilities of secondary age pupils at risk of school exclusion: A preliminary report. Child Language Teaching and Therapy, 25 (1), 123 – 139.</bibtext> </blist> <blist> <bibl id="bib9" idref="ref7" type="bt">9</bibl> <bibtext> Daniels, H., & Cole, T. (2010). Exclusion from school: Short‐term setback or a long term of difficulties? European Journal of Special Needs Education, 25 (2), 115 – 130.</bibtext> </blist> <blist> <bibtext> Demie, F. (2021). The experience of black Caribbean pupils in school exclusion in England. Educational Review, 73 (1), 55 – 70.</bibtext> </blist> <blist> <bibtext> Demie, F., & Strand, S. (2006). English language acquisition and educational attainment at the end of secondary school. Educational Studies, 32 (2), 215 – 231.</bibtext> </blist> <blist> <bibtext> Department for Education. (2019). Permanent and fixed period exclusions in England 2017 to 2018. Gov.UK. https://<ulink href="http://www.gov.uk/government/statistics/permanent‐and‐fixed‐period‐exclusions‐in‐england‐2017‐to‐2018">www.gov.uk/government/statistics/permanent‐and‐fixed‐period‐exclusions‐in‐england‐2017‐to‐2018</ulink></bibtext> </blist> <blist> <bibtext> Department for Education. (2023). Suspensions and permanent exclusions in England. Gov.UK. https://explore‐education‐statistics.service.gov.uk/find‐statistics/suspensions‐and‐permanent‐exclusions‐in‐england</bibtext> </blist> <blist> <bibtext> Done, E. J., & Knowler, H. (2020). Painful invisibilities: Roll management or 'off‐rolling' and professional identity. British Educational Research Journal, 46 (3), 516 – 531.</bibtext> </blist> <blist> <bibtext> Duffy, G., Robinson, G., Gallagher, T., & Templeton, M. (2021). School exclusion disparities in the UK: A view from Northern Ireland. Emotional and Behavioural Difficulties, 26 (1), 3 – 18.</bibtext> </blist> <blist> <bibtext> Early, E., Miller, S., Dunne, L., Thurston, A., & Filiz, M. (2020). The influence of socio‐economic background and gender on school attainment in the United Kingdom: A systematic review. Review of Education, 8 (1), 120 – 152.</bibtext> </blist> <blist> <bibtext> Farquharson, C., McNally, S., & Tahir, I. (2024). Education inequalities. Oxford open Economics, 3 (1), i760 – i820.</bibtext> </blist> <blist> <bibtext> FFT Education Datalab. (2024). GCSE results 2024: The main trends in grades and entries. Exams and Assessment. https://ffteducationdatalab.org.uk/2024/08/gcse‐results‐2024‐the‐main‐trends‐in‐grades‐and‐entries/</bibtext> </blist> <blist> <bibtext> Fletcher, J., Strand, S., & Thomas, S. (2015). The educational progress of looked after children in England: technical report 1: secondary school progress and attainment. Nuffield Foundation. https://research‐information.bris.ac.uk/ws/portalfiles/portal/56549074/EducationalProgressLookedAfterChildrenTechnical_Report_1_Nov2015.pdf</bibtext> </blist> <blist> <bibtext> Gill, K., Quilter‐Pinner, H., & Swift, D. (2017). Making the difference: breaking the link between school exclusion and social exclusion. Institute for Public Policy Research. https://ippr‐org.files.svdcdn.com/production/Downloads/making‐the‐difference‐summary‐october‐2017.pdf</bibtext> </blist> <blist> <bibtext> Gorard, S. (2012). Who is eligible for free school meals? Characterising free school meals as a measure of disadvantage in England. British Educational Research Journal, 38 (6), 1003 – 1017.</bibtext> </blist> <blist> <bibtext> Gross, J., & McChrystal, M. A. (2001). The protection of a statement? Permanent exclusions and the SEN code of practice. Educational Psychology in Practice, 17 (4), 347 – 359.</bibtext> </blist> <blist> <bibtext> Harland, L. (2014). Educational attainment of children and young people in the looked‐after care system. Community Practitioner, 87 (11), 25 – 27.</bibtext> </blist> <blist> <bibtext> Hodgson, P., & Webb, D. (2005). Young people, crime and school exclusion: A case of some surprises. The Howard Journal of Criminal Justice, 44 (1), 12 – 28.</bibtext> </blist> <blist> <bibtext> Humphrey, N., Wigelsworth, M., Barlow, A., & Squires, G. (2013). The role of school and individual differences in the academic attainment of learners with special educational needs and disabilities: A multi‐level analysis. International Journal of Inclusive Education, 17 (9), 909 – 931.</bibtext> </blist> <blist> <bibtext> Ilie, S., Sutherland, A., & Vignoles, A. (2017). Revisiting free school meal eligibility as a proxy for pupil socio‐economic deprivation. British Educational Research Journal, 43 (2), 253 – 274.</bibtext> </blist> <blist> <bibtext> Jackson, M. (2012). Bold choices: How ethnic inequalities in educational attainment are suppressed. Oxford Review of Education, 38 (2), 189 – 208.</bibtext> </blist> <blist> <bibtext> Krezmien, M. P., Leone, P. E., & Achilles, G. M. (2006). Suspension, race, and disability: Analysis of statewide practices and reporting. Journal of Emotional and Behavioral Disorders, 14 (4), 217 – 226.</bibtext> </blist> <blist> <bibtext> Kulz, C. (2019). Mapping folk devils old and new through permanent exclusion from London schools. Race Ethnicity and Education, 22 (1), 93 – 109.</bibtext> </blist> <blist> <bibtext> Lessof, C., Ross, A., Brind, R., Harding, C., Bell, E., & Kyriakopoulos, G. (2018). Understanding KS4 attainment and progress: Evidence from LSYPE2. Department for Education. https://assets.publishing.service.gov.uk/media/5bc4b616e5274a361d74a6da/Understanding_KS4_LSYPE2_research‐report.pdf</bibtext> </blist> <blist> <bibtext> Li, Y. (2021). Entrenched inequalities? Class, gender and ethnic differences in educational and occupational attainment in England. Frontiers in Sociology, 5, 601035.</bibtext> </blist> <blist> <bibtext> Luke, N., Sinclair, I., & O'Higgins, A. (2015). The educational progress of looked after children in England: technical report 2: relating care to educational attainment and progress. Nuffield Foundation. https://<ulink href="http://www.education.ox.ac.uk/wp‐content/uploads/2019/05/Technical‐Report‐2.pdf">www.education.ox.ac.uk/wp‐content/uploads/2019/05/Technical‐Report‐2.pdf</ulink></bibtext> </blist> <blist> <bibtext> Lupton, R., Thomson, S., Velthuis, S., & Unwin, L. (2021). Moving on from initial GCSE 'failure': Post‐16 transitions for 'lower attainers' and why the English education system must do better. Nuffield Foundation. https://<ulink href="http://www.nuffieldfoundation.org/wp‐content/uploads/2021/02/Post16‐transitions‐for‐lower‐attainers‐Final‐report.pdf">www.nuffieldfoundation.org/wp‐content/uploads/2021/02/Post16‐transitions‐for‐lower‐attainers‐Final‐report.pdf</ulink></bibtext> </blist> <blist> <bibtext> MacRae, S., Maguire, M. E. G., & Milbourne, L. (2003). Social exclusion: Exclusion from school. International Journal of Inclusive Education, 7 (2), 89 – 101.</bibtext> </blist> <blist> <bibtext> Mainwaring, D., & Hallam, S. (2010). 'Possible selves' of young people in a mainstream secondary school and a pupil referral unit: A comparison. Emotional and Behavioural Difficulties, 15 (2), 153 – 169.</bibtext> </blist> <blist> <bibtext> McCluskey, G., Cole, T., Daniels, H., Thompson, I., & Tawell, A. (2019). Exclusion from school in Scotland and across the UK: Contrasts and questions. British Educational Research Journal, 45 (6), 1140 – 1159.</bibtext> </blist> <blist> <bibtext> McCluskey, G., Riddell, S., Weedon, E., & Fordyce, M. (2016). Exclusion from school and recognition of difference. Discourse: Studies in the Cultural Politics of Education, 37 (4), 529 – 539.</bibtext> </blist> <blist> <bibtext> McCrystal, P., Percy, A., & Higgins, K. (2007). Exclusion and marginalisation in adolescence: The experience of school exclusion on drug use and antisocial behaviour. Journal of Youth Studies, 10 (1), 35 – 54.</bibtext> </blist> <blist> <bibtext> McShane, J. (2020). We know off‐rolling happens. Why are we still doing nothing? Support for Learning, 35 (3), 259 – 275.</bibtext> </blist> <blist> <bibtext> Murphy, R. (2022). How children make sense of their permanent exclusion: A thematic analysis from semi‐structured interviews. Emotional and Behavioural Difficulties, 27 (1), 43 – 57.</bibtext> </blist> <blist> <bibtext> Parsons, C. (2005). School exclusion: The will to punish. British Journal of Educational Studies, 53 (2), 187 – 211.</bibtext> </blist> <blist> <bibtext> Parsons, C., & Castle, F. (1998). The cost of school exclusion in England. International Journal of Inclusive Education, 2 (4), 277 – 294.</bibtext> </blist> <blist> <bibtext> Parsons, S., & Elliot Major, L. (2024). Not just grades: The far‐reaching consequences of failing to gain a grade 4‐9 in English and Maths GCSEs. UCL Centre for Longitudinal Studies. https://discovery.ucl.ac.uk/id/eprint/10194892/1/CLS‐Working‐Paper‐2024‐3‐Not‐just‐grades_consequences‐of‐failing‐to‐gain‐a‐grade‐4‐9‐in‐English‐and‐Maths‐GCSEs.pdf</bibtext> </blist> <blist> <bibtext> Pirrie, A., Macleod, G., Cullen, M. A., & McCluskey, G. (2011). What happens to pupils permanently excluded from special schools and pupil referral units in England? British Educational Research Journal, 37 (3), 519 – 538.</bibtext> </blist> <blist> <bibtext> Power, S., & Taylor, C. (2020). Not in the classroom, but still on the register: Hidden forms of school exclusion. International Journal of Inclusive Education, 24 (8), 867 – 881.</bibtext> </blist> <blist> <bibtext> Pritchard, C., & Cox, M. (1998). The criminality of former "special educational provision" permanently "excluded from school" adolescents as young adults (16–23): Costs and practical implications. Journal of Adolescence, 21 (5), 609 – 620.</bibtext> </blist> <blist> <bibtext> Sabates, R. (2008). Educational attainment and juvenile crime: Area‐level evidence using three cohorts of young people. The British Journal of Criminology, 48 (3), 395 – 409.</bibtext> </blist> <blist> <bibtext> Shuttleworth, I. (1995). The relationship between social deprivation, as measured by individual free school meal eligibility, and educational attainment at GCSE in Northern Ireland: A preliminary investigation. British Educational Research Journal, 21 (4), 487 – 504.</bibtext> </blist> <blist> <bibtext> Strand, S. (2013). The limits of social class in explaining ethnic gaps in educational attainment. British Educational Research Journal, 37 (2), 197 – 229.</bibtext> </blist> <blist> <bibtext> Strand, S., & Fletcher, J. (2014). A quantitative longitudinal analysis of exclusions from English secondary schools. University of Oxford.</bibtext> </blist> <blist> <bibtext> Strand, S., Malmberg, L., & Hall, J. (2015). English as an additional language and educational achievement in England: An analysis of the National Pupil Database. Educational Endowment Fund.</bibtext> </blist> <blist> <bibtext> Thompson, I., Tawell, A., & Daniels, H. (2021). Conflicts in professional concern and the exclusion of pupils with SEMH in England. Emotional and Behavioural Difficulties, 26 (1), 31 – 45.</bibtext> </blist> <blist> <bibtext> Velthuis, S., Lupton, R., Thomson, S., & Unwin, L. (2018). The characteristics and post‐16 transitions of GCSE 'lower attainers'. Nuffield Foundation. https://hummedia.manchester.ac.uk/institutes/mui/igau/lower_attainers_working_paper_oct2018.pdf</bibtext> </blist> <blist> <bibtext> Wong, Z. (2023). The alarming epidemic of school dropouts: Causes and consequences. Journal of Educational Sciences Research, 13 (3), 1.</bibtext> </blist> </ref> <aug> <p>By Stephen Hills; Matthew Walker; James Guinn and Aubrey Kent</p> <p>Reported by Author; Author; Author; Author</p> </aug> <nolink nlid="nl1" bibid="bib13" firstref="ref1"></nolink> <nolink nlid="nl2" bibid="bib42" firstref="ref2"></nolink> <nolink nlid="nl3" bibid="bib29" firstref="ref4"></nolink> <nolink nlid="nl4" bibid="bib44" firstref="ref5"></nolink> <nolink nlid="nl5" bibid="bib24" firstref="ref8"></nolink> <nolink nlid="nl6" bibid="bib38" firstref="ref9"></nolink> <nolink nlid="nl7" bibid="bib46" firstref="ref10"></nolink> <nolink nlid="nl8" bibid="bib35" firstref="ref13"></nolink> <nolink nlid="nl9" bibid="bib15" firstref="ref14"></nolink> <nolink nlid="nl10" bibid="bib36" firstref="ref15"></nolink> <nolink nlid="nl11" bibid="bib28" firstref="ref17"></nolink> <nolink nlid="nl12" bibid="bib50" firstref="ref18"></nolink> <nolink nlid="nl13" bibid="bib10" firstref="ref21"></nolink> <nolink nlid="nl14" bibid="bib34" firstref="ref24"></nolink> <nolink nlid="nl15" bibid="bib41" firstref="ref25"></nolink> <nolink nlid="nl16" bibid="bib25" firstref="ref37"></nolink> <nolink nlid="nl17" bibid="bib53" firstref="ref38"></nolink> <nolink nlid="nl18" bibid="bib22" firstref="ref39"></nolink> <nolink nlid="nl19" bibid="bib48" firstref="ref43"></nolink> <nolink nlid="nl20" bibid="bib21" firstref="ref44"></nolink> <nolink nlid="nl21" bibid="bib26" firstref="ref45"></nolink> <nolink nlid="nl22" bibid="bib17" firstref="ref46"></nolink> <nolink nlid="nl23" bibid="bib23" firstref="ref48"></nolink> <nolink nlid="nl24" bibid="bib32" firstref="ref50"></nolink> <nolink nlid="nl25" bibid="bib19" firstref="ref51"></nolink> <nolink nlid="nl26" bibid="bib18" firstref="ref56"></nolink> <nolink nlid="nl27" bibid="bib16" firstref="ref57"></nolink> <nolink nlid="nl28" bibid="bib49" firstref="ref64"></nolink> <nolink nlid="nl29" bibid="bib31" firstref="ref65"></nolink> <nolink nlid="nl30" bibid="bib27" firstref="ref66"></nolink> <nolink nlid="nl31" bibid="bib51" firstref="ref68"></nolink> <nolink nlid="nl32" bibid="bib11" firstref="ref69"></nolink> <nolink nlid="nl33" bibid="bib30" firstref="ref71"></nolink> <nolink nlid="nl34" bibid="bib54" firstref="ref72"></nolink> <nolink nlid="nl35" bibid="bib33" firstref="ref77"></nolink> <nolink nlid="nl36" bibid="bib43" firstref="ref79"></nolink> <nolink nlid="nl37" bibid="bib47" firstref="ref80"></nolink> <nolink nlid="nl38" bibid="bib12" firstref="ref97"></nolink> <nolink nlid="nl39" bibid="bib37" firstref="ref139"></nolink> <nolink nlid="nl40" bibid="bib52" firstref="ref140"></nolink> <nolink nlid="nl41" bibid="bib39" firstref="ref141"></nolink> <nolink nlid="nl42" bibid="bib14" firstref="ref142"></nolink> <nolink nlid="nl43" bibid="bib40" firstref="ref145"></nolink> <nolink nlid="nl44" bibid="bib45" firstref="ref146"></nolink> <nolink nlid="nl45" bibid="bib20" firstref="ref148"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ1475322
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: The GCSE Attainment Gap: Assessing the Influence of Permanent School Exclusion
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Stephen+Hills%22">Stephen Hills</searchLink> (ORCID <externalLink term="https://orcid.org/0000-0001-5894-1418">0000-0001-5894-1418</externalLink>)<br /><searchLink fieldCode="AR" term="%22Matthew+Walker%22">Matthew Walker</searchLink> (ORCID <externalLink term="https://orcid.org/0000-0002-3080-0530">0000-0002-3080-0530</externalLink>)<br /><searchLink fieldCode="AR" term="%22James+Guinn%22">James Guinn</searchLink> (ORCID <externalLink term="https://orcid.org/0009-0005-5451-1644">0009-0005-5451-1644</externalLink>)<br /><searchLink fieldCode="AR" term="%22Aubrey+Kent%22">Aubrey Kent</searchLink> (ORCID <externalLink term="https://orcid.org/0009-0002-8537-4409">0009-0002-8537-4409</externalLink>)
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22British+Educational+Research+Journal%22"><i>British Educational Research Journal</i></searchLink>. 2025 51(3):1481-1497.
– Name: Avail
  Label: Availability
  Group: Avail
  Data: Wiley. Available from: John Wiley & Sons, Inc. 111 River Street, Hoboken, NJ 07030. Tel: 800-835-6770; e-mail: cs-journals@wiley.com; Web site: https://www.wiley.com/en-us
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 17
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2025
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Reports - Research
– Name: Audience
  Label: Education Level
  Group: Audnce
  Data: <searchLink fieldCode="EL" term="%22Secondary+Education%22">Secondary Education</searchLink>
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Foreign+Countries%22">Foreign Countries</searchLink><br /><searchLink fieldCode="DE" term="%22Educational+Attainment%22">Educational Attainment</searchLink><br /><searchLink fieldCode="DE" term="%22Exit+Examinations%22">Exit Examinations</searchLink><br /><searchLink fieldCode="DE" term="%22Achievement+Gap%22">Achievement Gap</searchLink><br /><searchLink fieldCode="DE" term="%22At+Risk+Students%22">At Risk Students</searchLink><br /><searchLink fieldCode="DE" term="%22Disadvantaged%22">Disadvantaged</searchLink><br /><searchLink fieldCode="DE" term="%22Scores%22">Scores</searchLink><br /><searchLink fieldCode="DE" term="%22Secondary+School+Students%22">Secondary School Students</searchLink><br /><searchLink fieldCode="DE" term="%22Inclusion%22">Inclusion</searchLink><br /><searchLink fieldCode="DE" term="%22Access+to+Education%22">Access to Education</searchLink><br /><searchLink fieldCode="DE" term="%22Barriers%22">Barriers</searchLink><br /><searchLink fieldCode="DE" term="%22Expulsion%22">Expulsion</searchLink><br /><searchLink fieldCode="DE" term="%22Equal+Opportunities+%28Jobs%29%22">Equal Opportunities (Jobs)</searchLink>
– Name: Subject
  Label: Geographic Terms
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22United+Kingdom%22">United Kingdom</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1002/berj.4133
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 0141-1926<br />1469-3518
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Permanent school exclusions continue to be a topic of keen interest to UK schools and policymakers. The debate over the practice has recently intensified owing to the perceived negative outcomes directly resulting from the exclusion event. Research has indeed shown that pupils who have been permanently excluded are at a greater risk for a variety of negative life outcomes when compared with their non-excluded peers. However, that disadvantaged groups are disproportionately represented among those excluded has not been accounted for in empirical testing. Accordingly, previous measures of the influence of permanent exclusion may have over-estimated its negative consequences because they have not controlled for disadvantageous pupil characteristics. This is a critical limitation of the research owing to the influence of confounding variables and sample selection bias. Using the National Pupil Database and a full cohort of UK pupils (N = 590,092), our analysis tracked a sample of 1490 pupils permanently excluded in year 11 of the English education system in 2018/2019. Using capped GCSE points as the academic attainment variable, we find that permanently excluded pupil scores were nearly 25 points lower than their non-permanently excluded peers. However, when controlling for disadvantageous pupil characteristics, this difference was cut roughly in half. As such, we conclude that permanent exclusion is neither the catalyst of disadvantage nor a continuation of disadvantage on the same trajectory, but rather an accentuation of existing disadvantage. In other words, the existing trajectory of disadvantage gets steeper following the permanent exclusion event. Therefore, considering that the GCSE attainment gap found is equally attributable to both permanent exclusion and disadvantageous pupil characteristics, policymakers should both limit permanent exclusion to being a last resort and provide additional support for pupils at risk of being permanently excluded. Including a permanently excluded pupil's GCSE attainment in their former school's academic league table data incentivises schools to act in the best interests of these highly disadvantaged and vulnerable pupils.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2025
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1475322
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1475322
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1002/berj.4133
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 17
        StartPage: 1481
    Subjects:
      – SubjectFull: Foreign Countries
        Type: general
      – SubjectFull: Educational Attainment
        Type: general
      – SubjectFull: Exit Examinations
        Type: general
      – SubjectFull: Achievement Gap
        Type: general
      – SubjectFull: At Risk Students
        Type: general
      – SubjectFull: Disadvantaged
        Type: general
      – SubjectFull: Scores
        Type: general
      – SubjectFull: Secondary School Students
        Type: general
      – SubjectFull: Inclusion
        Type: general
      – SubjectFull: Access to Education
        Type: general
      – SubjectFull: Barriers
        Type: general
      – SubjectFull: Expulsion
        Type: general
      – SubjectFull: Equal Opportunities (Jobs)
        Type: general
      – SubjectFull: United Kingdom
        Type: general
    Titles:
      – TitleFull: The GCSE Attainment Gap: Assessing the Influence of Permanent School Exclusion
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Stephen Hills
      – PersonEntity:
          Name:
            NameFull: Matthew Walker
      – PersonEntity:
          Name:
            NameFull: James Guinn
      – PersonEntity:
          Name:
            NameFull: Aubrey Kent
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 06
              Type: published
              Y: 2025
          Identifiers:
            – Type: issn-print
              Value: 0141-1926
            – Type: issn-electronic
              Value: 1469-3518
          Numbering:
            – Type: volume
              Value: 51
            – Type: issue
              Value: 3
          Titles:
            – TitleFull: British Educational Research Journal
              Type: main
ResultId 1