Academic integrity or academic misconduct? Conceptual difficulties in higher education and the potential contribution of student demographic factors

Academic misconduct and academic integrity are issues of importance to Higher Education Institutions (HEIs). Phraseologies and practices may con ﬂ ate unintentional mistakes with attempts to gain illegitimate advantage, with some groups potentially at higher risk. HEIs across the United Kingdom (UK) responded to a Freedom of Information Act (FOI) request, and provided data regarding their overall student demographics, the demographic details of their cases of referred and substantiated academic misconduct over a speci ﬁ c time period, as well as information about Turnitin. Available data were analysed with an estimated combined student population of 200,646 with an estimated 2718 students referred for academic misconduct. Analyses revealed some groups being referred out of proportion with their share of the student population, however, no signi ﬁ cant di ﬀ erences were observed by demographic for substantiated cases. Most of the responding HEIs indicated Turnitin was used for both identi ﬁ cation of misconduct and pedagogical purposes although some reported variations in use for pedagogy by department or module. Implications of the ﬁ ndings are considered. Limitations around available data are discussed and areas requiring more detailed consideration are identi ﬁ ed.


Introduction
A scientific-practitioner approach involves incorporating research evidence and experience within the field into one's work (Jones & Mehr, 2007).Whilst this approach may apply to many mechanisms of supporting students' development and learning in higher education (HE), this piece focuses upon academic integrity practices.We write as three academics involved in the processing of academic integrity cases at a higher education institution (HEI), and one has additional lived experience as a former student with dyslexia and background working within student services at university level.We outline preliminary evidence around protective and vulnerability factors associated with students' involvement in academic integrity processes, and consider how HEIs across the UK use Turnitin software for pedagogy and as part of their misconduct investigations.It is intended that this work will encourage holistic consideration of potential factors relevant to academic integrity practices to improve student experience of higher education and support HEIs to achieve their aims, both within the UK and internationally.
One aim of HEIs is quality assurance, within the UK it is the focus of the Quality Assurance Agency for Higher Education (QAA), which includes an Academic Integrity Charter (QAA (2020)) 1 The Charter acts to ensure that HEIs provide a certain level of standard and quality in relation to academic integrity, intending to provide a bench-line in terms of procedures providers undertake in relation to policies and practices (QAA, ND).The QAA website offers a glossary of relevant terms, defining Academic Integrity as 'Compliance with ethical and professional principles, standards, practices and consistent system of values, that serves as guidance for making decisions and taking actions in education, research and scholarship' (Tauginienė et al., 2018, pp. 7-8).A definition of Academic Misconduct is also provided: 'Any action or attempted action that undermines academic integrity and may result in an unfair academic advantage or disadvantage for any member of the academic community or wider society' (Tauginienė et al., 2018, p. 8).It appears that 'Academic Integrity' focuses on ethics and values, whereas 'Academic Misconduct' focuses on illegitimate academic advantage.'Academic Misconduct' has a negative phraseology, whilst 'Academic Integrity' potentially conflates unintentional mistakes with lack of integrity.There are concerns about the extent to which poor academic practice bleeds into academic misconduct (OIA, ND) and it is important that educators understand how one may arise from the other and how to distinguish and suggest solutions to each.One aspect is students' awareness of what constitutes good practice.
Demographics of individuals attending HEIs are changing: increased numbers of students from widening participation style initiatives and students with declared disabilities, changes in ethnic groupings (HESA, 2023a), and generally an increase of students of non-UK domicile (HESA, 2023b) have been seen in the UK, whilst, for example, Australia has recently seen large increases in international student numbers (Australian Government Department of Education, 2023).These changes in who is being taught and assessed present increased potential for unintentional mistakes to be conflated with intentional academic misconduct and lead to inequalities.To put this in context, there are other occasions when particular groups are disadvantaged disproportionately compared to others: societallevel research notes there are protective and vulnerability factors that contribute to inequalities at multiple levels (Byrne et al., 2020), for example, the Criminal Justice System (CJS).Statistics for England and Wales suggest that minority ethnic groups are often overrepresented within the CJS compared to white ethnic groups (Ministry of Justice, 2021).Institutional racism (The Institute of Race Relations, 2020a) is suggested to be a contributory factor in the overrepresentation of minority ethnic groups (Institute of Race Relations, 2020b).The CJS has certain parallels to academic integrity processes and is used here to demonstrate how broader concerns can be extended to the HE context and illustrate the need to actively identify potential risks.We do not suggest that HEIs act with intentional biases, however, students can suffer from social inequality and institutional biases if academic integrity procedures disproportionately affect certain groups.
Researchers (e.g., Denisova-Schmidt, 2016) have considered how cultural expectations of HE may result in differing attitudes towards academic integrity.Research has found that non-US students self-reported higher rates of academic misconduct, however, this resulted from US students having a more lenient attitude to what constitutes academic misconduct (Chapman & Lupton, 2004;Lupton et al., 2000).Meanwhile, Dawson and Sutherland-Smith (2018) identify that an honour system within US HE may deter academic misconduct, though this does not extend to UK HE due to differing perspectives on moral elements of academic integrity, illustrating how differences in academic culture can have far-ranging consequences even within two closely comparable cultures such as the UK and the USA.Robey et al. (2022), examining a southwestern US HEI, observed that whilst white and citizen students were more likely to encounter academic misconduct processes, underrepresented racial groups and non-citizens received higher sanctions for the same levels of misconduct.
There is similar potential for disadvantage by virtue of disability.Ableism, or focus on 'able-bodiedness' as the ideal state, has been considered in the academic context, operating in a manner similar to institutional racism (Dolmage, 2017).To return to the CJS comparison, people with a broad range of disabilities (including mental ill-health and neurodiversity differences) are over-represented, with adjustments not always identified or met in a system not well designed to meet their needs (Criminal Justice Joint Inspection, 2021; EHRC, 2020).In the context of academic integrity, disability has received little research attention, and where consideration has taken place, need for a social justice approach is highlighted (Pagaling et al., 2022).
It has been acknowledged that EDI requires more focus within HE (Advance, 2020), with the importance of acknowledging intersectionality also identified (Nichols & Stahl, 2019).Cho et al. (2013) explain that intersectionality involves the complex interplay of inequality and disadvantage, where different qualities or aspects of an individual's experience can expose them to different levels of (dis)advantage.Researchers have suggested that differences in ethnicity, culture and nationality can be subsumed into a wider category of 'race' (Mwangi, 2014), that ethnicity and nationality can often intersect (Joseph, 2006) and that intersecting traits can have a multiplicative rather than additive impact (Anderson & Collins, 1995).Student ethnicity and nationality (Poyrazli et al., 2010;Wilton & Constantine, 2003), second language status (Khawaja et al., 2017) and experiencing disability (Holloway, 2001) have been linked to increased stress.Stress may lead to increased acceptance and undertaking of academic misconduct (Ip et al., 2016;Tindall et al., 2021;Tindall & Curtis, 2019) and an association with decreased self-control (Tindall et al., 2021) which may link to positive attitudes towards plagiarism (Tindall & Curtis, 2019).Stress negatively impacts decision-making and risk-assessment, and can lead to habitual behaviours (Porcelli & Delgado, 2017), therefore such students may unwittingly fall into habits of academic misconduct.
Intersectionality warns us that whilst the exact pattern of (dis)advantage may be complex, factors such as ethnicity, cultural background and disability are often key.HEIs make laudable efforts to be responsive to these factors and make provision for specialist support for those with declared or anticipated need.One common practice includes an introduction to academic integrity processes at the induction level, which may include an introduction to the similarity index software Turnitin.Used in over 15,000 institutions across 140 countries (Turnitin, 2021), Turnitin may be used as a teaching tool to good effect, especially amongst those reporting poor academic writing skills or studying in a second language environment (Orlando et al., 2018), although it is not evident if this is a primary use within institutions.This software is not without its critics, however -Weber-Wulff (2019) highlights how use is often around detection of potential misconduct and an associated fixation with 'similarity scores'.Other practices to orient students into academic writing conventions include embedded support within the institution such as student services and academic skills, whilst there is often tailored support for international students.For students with needs associated with disability there is, within the UK, support incorporated within the Disabled Students' Allowance (DSA) system and reasonable adjustments (RA) available under the Equality Act (2010).It is clear that there is much commitment and provision involved, however, it is difficult to know how successful these efforts are in the context of academic integrity practices, with this kind of evaluation being an important aspect of transparency within the Academic Integrity charter and its worldwide equivalents.
In response to this, the current study examined the numbers of academic misconduct referrals and 2 number of substantiated cases amongst UK HEIs during the final semester of the academic year 2020/2021, and the demographic breakdown of their overall student population and that of their misconduct cases: both students referred and for students who had an academic misconduct case sustained.The broad aims were to provide a preliminary examination of the potential contribution of anticipated protective and vulnerability related factors to academic misconduct practices in HEIs, and to provide a preliminary understanding of the ways that HEIs use Turnitin for pedagogy and for identifying cases of academic misconduct.Whilst the research focus is on the UK, it is hoped that the findings will be of use to educators within the international HE community.

Materials and methods
Participants 166 HEIs were approached via FOI request using the www.whatdotheyknow.comwebsite, which enables sending requests for information held by public authorities in the UK, in accordance with the Freedom of Information Act (2020).Public authorities will have a designated FOI officer who will receive and respond to such requests, with requests and responses publicly visible on the website.
The Universities UK members list was used to identify HEIs to contact.Due to clerical error, approximately 4 institutions were not contacted.A cross-sectional, between participants quasi-experimental design was used to investigate differences in academic misconduct referral and substantiation by demographic group.

Materials
Institutions were asked questions relating to the demographics of their entire student body, the number of students referred for academic misconduct during the final semester of the academic year 2020/2021 and a breakdown of this number by demographics.They were asked the number of student cases of academic misconduct found proven during this time, broken down by demographic group.Institutions were asked if they used Turnitin software to identify cases of academic misconduct, and if their institution gave students access to Turnitin to understand academic misconduct prior to submission of work.

Procedure
Ethical approval was gained from the authors' faculty ethics committee.Institutions were asked questions via FOI request between 25.07.2022 and 22.08.22.Automatic reminders were sent at regular intervals in the event of non-response.Data collection closed 5.10.2022.All institutions received at least 20 working days to respond.

Initial stages
Following the sending of requests several institutions were removed at various stages, visualised in Figure 1.Early responses indicated that some colleges within institutions were not always directly involved in assessment, and so 30 institutions were removed from consideration.Of the remaining 133 institutions, we received responses from 103.Twenty-two institutions cited copyright clauses or gave us reason to believe we could not use their response datatheir information is not discussed hereafter.Of the remaining institutions, 16 provided data that was usable to assess impact of demographics on cases of referral for academic misconduct and impact of demographics on cases of academic misconduct being substantiated (our statistical analyses).Not all institutions provided information for all analyses, therefore any estimate of the total student population and the proportions engaged in the academic integrity process would be misleading, as the number of contributing sources varies in places.We simultaneously analysed responses from the remaining institutions' answers to our questions about Turnitin use for identification, (Question 1) and use for pedagogy (Question 2).There was a wide variation in depth and detail of responses, therefore, to provide a more cohesive picture, responses were categorised into response options for each Turnitin-related question.The number of institutions that provided data we could quantify in this way is displayed below.

Statistical analyses
Chi-square goodness-of-fit tests were used to determine if referral and sustained rates for Academic Misconduct differed by demographics from those of the overall student population or for sustained cases, if the sustained demographics differed from that of the referred population.To meet the assumptions for chi-square, categories with an expected count of less than 5 were removed from analysis.For all chi-squares, Bonferroni correction was applied to allow for multiple testing, with the threshold p value of .006(.05/8) being used.Where a chi-square goodness-of-fit test produced a significant effect, standardised residuals were calculated for each cell, in keeping with guidelines for this type of work (cf Agresti, 2007;Sharpe, 2015) and enabling identification of categories making the greatest contribution to the significant result.The standardised residuals which were greater than 3 were identified as making a meaningful individual contribution (Sharpe, 2015).

Referrals by demographic
Students assessed for English language competence before commencing study were referred significantly more than would be expected from their share of the student population, whilst those who were not assessed were referred less than might be expected (see Table 1).Some nationalities were referred more frequently than would be expected, whereas others were referred less; this was also found amongst different ethnicities.Notably, students of Pakistani ethnicity were referred disproportionately more frequently, unlike students of Pakistani nationality.
Finally, students with no declared disability (Category A) were referred more than would be expected, whilst some other disability groups were referred less than would be expected (Disability categories B, E, F, G and I covering Social/communication impairment; Long standing illness or health condition; Mental health condition; Specific learning difficulty and finally Disability, impairment or medical condition not listed elsewhere, respectively).
These findings can be conceptualised as demographics associated with risk and protective factors at the referral level, see Table 1.It should be noted that groups with 5 or fewer members were omitted.

Cases substantiated
Chi-square tests were used to examine if there were significantly disproportionate ratios of cases sustained or dismissed relating to English language competence, disability, ethnicity or nationality amongst those referred for academic misconduct.Again, categories with an expected count of less than 5 were removed from analysis.
There were no significant differences between students with English language competence assessed and those not assessed in proportion of academic misconduct cases sustained (X 2 (1, N = 250) = 0.34, n.s.), or between expected and observed sustained cases in our other demographic factors of nationality X 2 (23, N = 1600) = 39.36,p n.s.; ethnicity X 2 (10, N = 1329) = 14.39,p n.s.; or disability X 2 (3, N = 2028) = 1.05, p n.s., suggesting that cases are sustained in proportion of share of referred population regardless of these demographic factors.

Use of Turnitin
Responses to Question 1 (use for identifying possible academic misconduct) and Question 2 (use for pedagogy) were coded into three categories, 'Yes', 'No' and 'Turnitin was used inconsistently' (across department, school, module or assessment).Responses not fitting into categories were excluded from this analysis.The majority (sixty-six; 95.65%) used Turnitin to help staff identify possible academic misconduct, three (4.35%) reported that use of Turnitin varied across department and assessment and none of the institutions reported not using Turnitin in this fashion at all.For Question 2 the majority (forty-five; 73.77%) used Turnitin to help students understand and identify possible academic misconduct prior to submission, eight (13.11%) reported that use of Turnitin varied across department and module, whilst eight reported not using Turnitin in this manner.

Discussion
Overall, referrals suggest a pattern of certain groups being referred more or less than others proportionate to their share of the student population.Meanwhile at the substantiation stage, no groups are disproportionately represented relative to their population at referral stage.The results around referral suggested that students who were assessed for English language competence were referred more than would be expected from their share of the student population and those who were not assessed were referred less than expected.Whilst meeting a certain level on a test can show a simplistic understanding of skill, it does not reveal the cognitive load involved in undertaking a task.Nevertheless, the University and Colleges Admissions Service (UCAS) states that applicants who do not hold English as a first language must take a test to establish a sufficient standard of English to study at HE level in the UK, the most common being the International English Language Testing System (IELTS) (Hu & Trenkic, 2021).Concerns about this test include lack of validity due to its revision to be used as both a test of academic ability and a useable part of the immigration system (Read, 2022); the ability for students to pass via coaching and retesting resulting in non-generalisable results (Hu & Trenkic, 2021); and that HEIs tend to accept students with lower scores than the IELTS recommends (Feast, 2002;IELTS, 2023).However, institutions may consider contextual information, therefore our findings provide a preliminary indication of areas to consider further, for example, the contribution of English language ability on different types of academic misconduct.Bretag et al. (2019) observed that contract cheating occurred more frequently amongst students studying in an additional language, whilst Maxwell et al. (2006) found that in cases of plagiarism there was no difference between students who studied in their first language and students studying in an additional language.
The ability to read and select literature, create notes and plans and then paraphrase and synthesise ideas in a structured narrative is a multilevel task requiring considerable skill and practice to develop into a confident and proficient authorial identity (Elander et al., 2010).Therefore, students studying in an additional language may disproportionately struggle with aspects of academic integrity compared to those students who study in their first language, cultural differences in education systems may compound this (Clark & Yu, 2021) as well as the increased stress from studying in another language (Khawaja et al., 2017).Whilst this was anticipated as a risk factor, nationality was also anticipated to have a role.
Students with Chinese, Greek, Indian, Kuwaiti, Nigerian or Uzbekistani nationality were all referred more than would be expected, whilst those of Canadian, Chilean, French, German, Italian, Polish, Singaporean, British or American nationality were referred less, suggesting nationality may be a risk factor to the academic experience.This is supportive of Lupton et al. (2000) and Chapman and Lupton (2004), especially the replication of the lower rate of academic misconduct amongst Polish students (Lupton et al., 2000).UK nationals are likely to have a lifetime of experience within the UK education system and be familiar with UK standards, expectations and elements of the hidden curricula.Students who are not UK nationals may experience extra challenges of being away from their support systems, and although this type of transitional challenge is the case for many students to a degree (Dangoisse et al., 2020), international students may be unfamiliar with UK expectations.Additionally, different nations value different academic skills (e.g., copying verbatim as a sign of respect, Bista, 2011) and so nations may vary in attitudes, beliefs and practices regarding education and assessment, which may offer an explanation (Clark & Yu, 2021;Hayes & Introna, 2005;Hendy et al., 2021).
Meanwhile, some groups of ethnicities (Indian, Pakistani) were referred more frequently than would be expected, whilst others were referred less (Bangladeshi, Asian Other, Mixed Asian, White and Other), and some ethnicities were referred as would be expected for their share of the student population (Arab, Chinese, Black African, Black Caribbean, Mixed Caribbean, Other Mixed and Prefer not to say).The impact of culture rather than nationality supports the findings of Kasler et al. (2021) who reported differences in misconduct levels reflecting culture in students who shared Israeli nationality, a contrast to the findings of Robey et al. (2022), where referrals were more frequent amongst home nationalities and white ethnicities compared to underrepresented groups and international students, yet in line with Beasley (2016).The apparent contradiction with Robey et al. (2022) may be methodological: nationality and ethnicity were collapsed into dichotomous categories, whereas we attempted a more granular approach.
This approach enabled the suggestion of a complex interplay between nationality, ethnicity and culture: students of Pakistani nationality were not referred more than expected, whilst students of Pakistani ethnicity were referred more, suggesting a possible disconnect between educational culture in Pakistan and that of members of the Pakistani diaspora.It may be that students from Pakistan studying in the UK have a different mosaic of protective and vulnerability factors to UK students of Pakistani ethnicity, consistent with suggestion membership of multiple minority groups can have a multiplicative negative or protective effect (Parra & Hastings, 2018).Previous studies have examined the impact of nationality (e.g., Clark & Yu, 2021;Hayes & Introna, 2005;Hendy et al., 2021) or ethnicity (Kuntz & Butler, 2014;Martin, 2012) on awareness of plagiarism; this is the first study to our knowledge to consider both.Students who study abroad may not be characteristic of the overall student culture in their home nation (Hayes & Introna, 2005) and cultural differences in perceptions of Academic Ability may exist between students and HEIs (Kuntz & Butler, 2014).Therefore, it is not sufficient to consider nationality or ethnicity alone, but together, in combination with other relevant factors.
We anticipated declared disability to present a potential relevant risk factor, and our findings suggested that some specific disabled groups (Disability categories B, E, F, G, which incorporates Autism Spectrum Disorder, mental health difficulties, longstanding illness and Specific Learning Difficulties) were referred less frequently than would be expected for their share of the student population, whilst students with no declared disability were referred more than would be expected.These findings highlight that from an intersectional perspective the absence or presence of disability alone should not be the only indictor of consideration, that a student's current emotional state is also important.
It is also possible that for students who do not have a disability, progression through education may have been achievable without need for excesses of effort or additional skill training, and they therefore may not have engaged with academic authorship skills training, as they did not anticipate requiring it (Chen & Van Ullen, 2011).Similarly, Dangoisse et al. (2020) suggested that students without a disability can sometimes be overconfident in their abilities in the early stages of their academic careers and initially focus more energy into the social side of university life.Overconfidence in academic abilities may also extend to those around library research skills (including referencing and citing) (Angell, 2015) and this has been linked to poorer ability to identify and avoid unintentional plagiarism (Chen & Van Ullen, 2011), although confidence in abilities and library research skills have not always found to be strongly linked (Gustavson & Nall, 2011).However, this increased confidence and focus on social life may lead to an unwillingness to engage in what are viewed as unnecessary distractions, such as attending library courses on academic skills and how to avoid misconduct.
Some groups appeared better able to avoid misconduct.Where some specific disability groupings are referred less than their proportion, a potential explanation may be that when a marker sees a RA notification on a script that they apply extra leniency and do not refer, emphasising the importance of careful monitoring by institutions to enable an equal playing field.Alternatively, there is some evidence that students with a disability arrive with an expectation of the need for hard work and a positive view of HEIs as supportive environments (Dangoisse et al., 2020) which may lead to increased engagement with and internalisation of the need to develop strong academic skills.Looking at the disability groupings specifically, these groupings are likely to be eligible for disability-related support such as DSA and RA which can include mentors, tutors, dyslexia tutors and assistive technology.As such, our findings may indicate that tailored and well delivered support is working well in supporting these groupings in their development of academic skills, a notable and positive finding.
Further, whilst we have identified apparent risk and protective factors for referrals it is important to note that in all of the areas we investigated, the number of cases substantiated were in similar proportions to cases referred for each group.This may suggest there is no additional positive or negative bias at the stage of substantiation (beyond that of the initial referral) but requires further study.Additionally, the majority of responding institutions also reported using Turnitin for pedagogical purposes.The Higher Education Academy (HEA) (2014, now Advance HE) has recommended the use of formative learning about academic integrity using similarity matching software.A small number of institutions reported Turnitin use for pedagogical purposes varied across departments and modules, meanwhile the majority of institutions used Turnitin to help staff identify possible academic misconduct and none of the institutions reported not using Turnitin in this fashion at all.This is consistent with the wider literature suggestion that many institutions use this type of software as a means of detecting plagiarism and academic misconduct rather than as a pedagogical aid (Mphahlele & McKenna, 2019).Institutions may misunderstand the purpose of the similarity index of Turnitin software (Foltýnek et al., 2020), however, the responses obtained do not permit detailed insight into these perceptions.Similarly, detail was not given around the use for student learning and development, and there was variation around the use for pedagogical purposes.This suggests such use is not standardised and therefore potentially not fair and balanced which could further exacerbate the risks to some groups of encountering the academic integrity system.To mitigate for this, academic integrity systems, processes and cases should be transparent and well audited.
We anticipated a FOI request would be a good means of carrying out this type of investigation: researchers have identified the potential challenges and advantages of using FOI for research (Savage & Hyde, 2014) with a suggestion it could be used more widely (Clifton-Sprigg et al., 2020;Savage & Hyde, 2014).We found that not all institutions responded to our requests or provided usable data.. Whilst we acknowledge the wording of our questions may have impacted responses, some responses did not feel to be following in the spirit of transparency: stating for example, that data remained under copyright and could not be used without permission.In our requests we used the term 'semester', partially to reduce the amount of time involved administratively for HEIs, by considering only part of the academic year.Some HEIs did not have available data for a particular semester, only having data for the entirety of the academic year.This suggests that ability to identify factors such as problems resulting from assessments in particular time periods may not be readily identifiable in audit and review.Some HEIs reported that the amount of time required to respond to the request would exceed that of FOI request requirements.Institutions should be willing and able to send clear and transparent records of these simple materials easily if they are carrying out careful monitoring and reviewing but it is not clear that they are.This has EDI risks and implications for the Academic Integrity Charter.
The Academic Integrity Charter however isn't without its difficulties: it notes that academic misconduct puts a student's academic and career future in jeopardy (QAA, 2020), a narrow view of impact on students.Reviewing of the QAA website and, more broadly, the academic integrity literature presents a narrative focused largely on students who are aiming to gain illegitimate advantage (Perry, 2010).Indeed, a recent QAA document focuses largely on 'cheating' (QAA, 2022).The term 'pedagogy' is not noted nor student 'journey'.Instead, there is a focus on wrongdoing.If students are truly intending to gain an illegitimate advantage and are permitted to gain from such, this is a problem and is unfair for all involved.However, from a psychological stance, actions and what causes them are complex and multiple explanations can explain an action.Students are learning, they may be confused, make a mistake, or lack skill and experience in their academic authorship skills.The lack of focus within this debate on students as learners developing skills in academic conventions is disappointing.
This paper begins to highlight specific elements of challenge in delivering quality pedagogical experiences for a diverse student body.It discusses a specific area of conundrum and potential controversy, where delivery of a quality education needs to be fit for purpose within the global HE marketplace.The shift in those who need us to teach and assess them has increased the complexity of accessibility and delivery of quality.Faced with this complexity comes the difficulty of understanding the pedagogical challenges, needs and goals of students and ensuring our systems and processes are fair and equitable.This study holds implications for research, education, EDI initiatives and importantly institutions' approach to students and to learning, both in the UK and beyond.To best support both students in their academic endeavours and institutions in maintaining their academic standards we argue a broad consideration of academic integrity and academic misconduct as conflicting processes is required; to achieve this it is important to change the narrative around academic misconduct and academic integrity.

Notes
1.Although the QAA has recently ceased to become the designated quality body in England, a brief internet search of publicly available information spanning the previous 12 months (at time of writing) suggests that this Charter is still informing practice in at least some HEIs and likely also was at the time of data collection.2. In this paper we use the terms 'academic authorship' and 'academic quality' as well as 'academic misconduct'.However, the questions asked of institutions used the term 'Academic Misconduct' as it is a widely used term in the literature and in practice.

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

Table 1 .
Protective and risk factors associated with referral.
HIGHER EDUCATION RESEARCH & DEVELOPMENT