A UK institution’s university-wide approach to plagiarism: First results and recommendations

Research and consultations in session 2003/2004 by a University’s Plagiarism Working Group uncovered a poor understanding of plagiarism and inconsistent handling procedures throughout its schools. In an effort to address both these issues, a strategic 2-year Action Plan was developed and rolled out beginning the following academic year in order to improve student support, staff awareness and more consistent practice overall. The plan included a pilot using the detection software service, TurnitinUK, with five of the University’s 14 schools. The pilot was only one of a series of university-wide deliberations, others included the revision and piloting of a University Plagiarism Code of Practice, implementation of school-based academic conduct officers, improved staff development opportunities and student support materials and events. One school in the University has served as a role model of good practice throughout. Noteworthy is the school’s record keeping practice since session 2001/02 of incidences of plagiarism and other academic misconduct. In the paper we present the factors such as gender, nationality and level of study that have been found linked to the incidences of plagiarism in the school. Additionally, the role plagiarism detection software plays in addressing plagiarism is explored within the collaborative and holistic approach of the Action Plan. Finally, the challenges and resistance faced by key players throughout the implementation of the first phase of the Action Plan at the University are considered and the commitment to continuous enhancement recognised.


Introduction
Plagiarism is certainly not new to higher education, and it has been widely recognised that Internet technology makes it easier to plagiarise than ever before. Academics and plagiarism experts expect the opportunities for "point-and-click" plagiarism to expand further as the Internet continues to grow (Carroll, 2005;Williams, 2002). These findings are further borne out by a 2002 survey conducted at Northumbria University in which 40% of students and 35% of the academic staff approached attribute plagiarism to the ease with which material can be copied from the Internet (Dordoy, 2002), findings consistent with a similar survey of North American students reported by McCabe (2003).
Consultation of the 'academic research services' for custom written term papers in the 1960s has evolved into diverse forms of cyber-pseudo-epigraphy facilitated by the networked research environments which include the commercial and private online availability of written academic work on any subject matter. Downloading began replacing manual transcription around the mid 1980s, and after Google's launch in 1998 the Internet became a public elixir of information of increasingly dubious origin, seemingly meeting steadily rising student demand. Plagiarism, however, is restricted to neither student incidences nor higher education. Vice Chancellors in Australia, newspaper editors in the U.S. and established authors in Germany have been caught copying written work. Increasingly plagiarism-detection service providers are in demand outside the education sector, by law firms, military agencies and information services (e.g., LexisNexis, 2005).
Australian and U.S. education institutions have been proactive in recording and addressing plagiarism for over 30 years; the first such U.S. study dates to 1940 (Drake, 1941). Paper mills are found critically discussed in U.S. literature in the early 1970s (Stavisky, 1973). Australian reports on the extent of plagiarism and reasons for cheating appear a few years later (e.g., Bushway & Nash, 1977). In 1992 The Center for Academic Integrity was founded in the U.S. in an effort to promote the values of academic integrity among students at a national level. Six years later John Barrie, a biophysics professor at Berkeley, founded Turnitin.com, the first commercially available software service developed to detect online plagiarism. In 2002, six Australian universities commissioned a comprehensive study of plagiarism using Barrie's detection software service and found that 14% of the 1925 student essays submitted were plagiarised to varying degrees (Foster, 2002). Most recently a large-scale U.S. survey was conducted with 30,000 undergraduates at 34 colleges, of which nearly 40% admitted plagiarising, up from 10% in 1999 ( McCabe, 2005).
To paraphrase Barlow (1995, p.18) U.K. Higher Education, on the other hand, seems only recently to have moved beyond the "pretending that it ain't broke means you don't have to fix it" stage. This is somewhat surprising considering the findings of the first U.K. survey of its kind in 1995 in which 50% of the surveyed undergraduate students admitted engaging in different forms of academic dishonesty (Newstead, Franklin-Stokes, & Armstead, 1996). The results hardly spearheaded institutional change, but after subsequent polls and surveys confirmed an increasingly worrying trend in UK higher education, the publicly-funded Joint Information Systems Committee Plagiarism Advisory Service (JISC PAS) was established in 2001 to help address the problem on a national level. One year earlier, Oxford Brookes University had developed a novel system of specialist officers ('Academic Conduct Officer,' ACO) to assist academic staff in dealing with students who did not comply with University academic conduct regulations. Books on plagiarism by U.K. publishers began to appear (e.g., Angelil-Carter, 2000;Carroll, 2002) and the JISC-funded, U.S. based plagiarism detection software, Turnitin ® UK, was made available at no cost to all UK higher education institutions between 2003 and 2005. Though uptake by UK institutions was not systematic, the inaugural U.K. Conference on Plagiarism sponsored by JISC PAS in 2004, gave plagiarism in higher education in the UK a further profile (JISC PAS, 2004).
One school within the authors' University has stood out internally since 2001 by recognising and recording cases of plagiarism over a period of three years to 2004/05. Its results and local educational and preventative measures enhanced the work of a University Plagiarism Working Group in the spring of 2004 in an effort to address the school's initial findings and wider concerns on an institutional-wide basis. In this paper, the school's misconduct data analysis is presented first. Measures implemented by the Working Group based on the good practice modelled by the school are then described. Finally, the results and challenges of the first set of university-wide deliberations to prevent plagiarism are explored.

Good practice in dealing with academic misconduct
The featured University is a modern (post-1992) university organised in four faculties with a total of 14 schools plus a Lifelong Learning (LLL) department. Some 15,000 undergraduates and postgraduates are taught by 750 members of academic staff. Nearly 30% of the students are mature-age students. Following an upsurge of interest in plagiarism in U.K. higher education in 2003, a short term Plagiarism Working Group was established by one of the University's Vice Principals. The working group was headed by one of the authors and included members from Educational Development, the secretariat, library services, quality enhancement services and the Student Association, as well as proactive members of the academic community.

The case study
As occurred at Oxford Brookes University, the appointment of an Academic Conduct Officer (ACO) 1 in the sample school in mid-session 2001/02 led to a change in approach by the school in dealing with plagiarism and associated issues. The change was prompted by the recognition that as much collusion as plagiarism was being reported and that the two were separate issues. Each was becoming a significant problem which had to be dealt with consistently.
At the end of the session, an internal report showed that there was some evidence for patterns in the occurrences, albeit based on a sample of only 25 cases. Notably: • Male students were more than twice as likely to be involved in academic misconduct as female students; • Undergraduates (UG) students were twice as likely to be involved in misconduct as taught 2 postgraduate (PG) students; • Non-native speakers of English (NNS) were 10 times more likely to be involved in misconduct than native speakers (NS).
The first two points align with the pattern of 'academic cheating' reported by Hart and Friesner, (2004). Summarising research by others, including Newstead, Franklyn and Armstead (1996), Hart and Friesner noted that cheating appeared to be associated with males rather than females and non-mature students rather than mature. Sutherland-Smith (2005) advises caution in drawing conclusions regarding NNS students, highlighting cultural differences in terms of academic conduct. Another factor is that some students seem to have an instrumental view of higher education and will do whatever it takes to get their qualification.
Three points underpinned the establishment of ACO practice in the school.
1) The school used terms in the following common ways: • Plagiarism is the unacknowledged use of the words and/or ideas of another; • Collusion occurs when two or more students submit joint work on an individual assignment with the knowledge of each.
2) There was no clear definition of major or minor misconduct in the University in the first year the ACO was appointed in the school. In the absence of other ACOs in the University and in order to establish an unequivocal definition, he proposed and implemented a very simple rule for use in the school: his applied even if the first offence was a major download in an Honours project. It worked in irst incident = minor misconduct; second and subsequent incidents = major mi T that a) both staff and students could understand the rule; and b) the school had by far the highest reporting rate of misconduct in the University in the Working Party's later survey (see below).
3) As a school dealing with a technical subject area, the students are typically not those well tudent population Details of the student cohorts were obtained from the University's Student Record System. It was soon clear that this data was unreliable for absolute analysis; for example, too many students versed in essay writing, citation and referencing. From comments made in academic miscond hearings, they may come from educational traditions in which repeating the words of a master verbatim seems to gain most marks and in which comparing two contrasting teachers' views an giving one's own opinion would not be favoured (Turner & Acker, 2002).
T was no attempt to cite a source for extracts of published work. A student who borrowed extensively from a website and listed its URL somewhere in the submission would fail be there was little of the student's own work to mark, but would not be investigated for misconduct C Table 1 shows details of th which no further action was taken, i.e. situations in which an incident was investigated but was 'not proven' to be misconduct. Also shown is the number of academic staff (from a roll of ~65) who reported misconduct. Noteworthy is the observation that it was largely the same members of academic staff who reported plagiarism each year. As the table shows, the total number of cases has fallen, reported by (approxim proportion of staff. Possible reasons for this include: • Better education of students on misconduct issues, through School talks, etc. Changes to the com S who did not matriculate or who withdrew after a short period were still included. However, ns under consideration. Two trends are clear: a) the proportion of female tudents is falling and b) the proportion of NNS students is increasing 3 . gory is the quoted that PG students were 0.6 mes as likely to be involved in academic misconduct as UG students. Another way of stating nts. All other figures in the table an be interpreted in the same way.

2002/
assuming that such errant data was spread over all types of students, the data could be used for relative analysis. Table 2 shows details of the numbers of students taught on campus within the school in the UK in the three sessio s Analy Table 3 shows the principal results of the analysis of Tabl The table entries mean that the ratio of the first category to the second cate figure. For example, in 2002/03, the PG/UG figure is 0.6. This means ti this is that the ratio of PG to UG student involvement is 0.6:1.
In 2003/04 and 2004/05, the same relative rate had risen to 1.2, i.e. that PG students were slightly more likely to be involved in misconduct than UG stude c The interesting points to note, bearing in mind the problems with the population data, are: • Female students had an anomalous rate of misconduct in 2003/04; otherwise the initial finding (that males are more than twice as likely as females to be involved in misconduct) • demic misconduct to slightly more likely.
igh. A separate analysis of A furth n the overall figure by level of study. Table 4 shows the results. The figures should be interpreted in the same way as elative rates of misconduct by level of study and first language, 2002-2005 onduct by taught PG and UG students. Thus r PG students in 2002/03, NNS students were 2.8 times as likely to be involved in academic t, whilst misconduct awareness education has the potential to have an pact on the behaviour of undergraduates (students are in the school typically for two to three pments for session 05/06 he 2005/2006 academic year has seen a number of developments evolve in the featured school ersity of staff attitudes to plagiarism and promote the s school of the Turnitin®UK detection service will be increased significantly. A few members of staff are piloting its use for text-based will • ine submissions will be compared to those already • lish ies which are used to investigate individual cases. is confirmed. PG students have moved from being significantly less likely than UG students to be involved in aca • NNS students are more than four times as likely to be involved in misconduct as NS students. This is lower than the 2001/02 estimate but still h UG NNS in programmes delivered overseas confirms this result. er analysis of the NNS/NS figures was carried out by breaking dow for Table 3. It could be inferred tha im years), this appears not to be true for taught postgraduates (typically with the school for one year).

Develo
T in order to further inform the Univ educational dimensions of Turnitin®UK's use.
• During the current session, the use by thi coursework. Students are allowed to see their originality reports and to resubmit if problems are found. It is intended that all student dissertations, at both Honours level and Masters level, will be submitted to Turnitin®UK by the students at draft stage. This allow students to see where problems exist and to correct them before their work becomes a formal submission. The work will be submitted again to Turnitin®UK at hand-in stage as a final check.
One member of staff has developed software to analyse student submissions to an onl assessment system. All student submitted, in the current or previous sessions. A staff perception survey is being conducted in the school to try to further estab attitudes to academic misconduct and the polic

Plagiar
The initial findings by the featured school clearly indicated that a central role should be played y educative measures to prevent and minimise plagiarism. In light of these results, the 2003 ing a 10-question survey to measure and evaluate the hool) cases (noteworthy here is that the highest rate of incidence outside the featured school was only 16 cases); • The s awa n was agreed among Working Group members that the polarisation of the student experience ce ce across all schools, a two year Action lan was drawn up which included a total of 10 measures to be implemented in a staged fashion 2004/05 and 2005/06 within an education, prevention, detection and • Academic Conduct Officers (ACOs) appointed in all 14 schools staff, and educative materials written for students All u ard. The two pilots, running for the duration of the t session, followed by a universityide implementation. The Code of Practice was drafted by a small team drawn from wed and to the 13 volunteers from 5 schools, expanded to 18 to include ism working group survey b Working Group responded by develop educative-preventive-detection and disciplinary procedures in place across the University. Response rate by the 14 schools and Lifelong Learning was 100%. The results exposed three main areas of concern: • detection described as ad hoc by 93% of the respondents ranging from 0 to a maximum of 91 (featured sc • deterrence proactively practiced in only two out of 14 schools and LLL (13%); and an overall unsatisfactory perception by staff of other staff as role models of good practise.
re ults reflected themes from a Student Association report in which the inconsistency of re ess, school guidelines and disciplinary procedures were raised with the Working Group. It across the schools, and particularly for distance learners and the overseas students in terms of plagiarism, contributed to reducing the quality of student experience and lowering academic morale. Purposeful, strategic action at all levels of the institution, benchmarked by good practi in the most proactive schools, was felt to be essential.

The two year action plan
In order to promote a threshold standard of good practi P between the academic years discipline continuum. An extant (but little known) institutional-wide definition of plagiarism was adopted from the Student Disciplinary Regulations (SDRs): Plagiarism: 'unacknowledged incorporation in a student's work either in an examination or assessment of material derived from the work (published or unpublished) of another'.
Primarily, the four strategic actions taken in the first year included: • Educative and awareness raising seminars and workshops for • a systematic Turnitin ® UK service pilot • a Plagiarism Code of Conduct pilot and consultation fo r actions were approved by Academic Bo Ac ion Plan, began with self-selected schools for the first w Educational Development, two academic faculties, and Secretariat and Management services. It was circulated in three schools at the beginning of Semester 1 of academic year 2004/05. Together with the Student Association, the original authors and pilot participants revie revised the Code in Semester 2 for renewed implementation in all schools in the following academic year. The revised Code was again approved by Academic Board. Modifications SDRs were made and approved also.
The detection software Turnitin ® UK was piloted from within Educational Development in two phases, initially on a small scale with interested ACOs. Recognising that software could only be one tool of many in the effort to deter • Evaluate the software as a staff support tool ays in reducing plagiarism The a ision to subscribe (or not) to the service afte t online in WebCT self-study training developed to facilitate and manage the pilot and support its d g th t participants completed a 20 question survey after the first four months of peration. The response rate was 78% (n=14). The results are summarised in the order of the t: he ed less than 3 hours of preparation before assignment submission. The high quality of information returned in the 178 Originality Reports was

•
Only two active users deployed the service with their students. These gave students opportunities to practice referencing by submitting • e arism was not such a problem that it warranted the use of such a service. Active users, however, felt at • databases students had access to. One respondent criticised the need to use the service alongside other detection methods such as search engines. academic misconduct and to support academic staff, the pilot's main objectives were to: • Assess the reliability and usability of the detection software • Evaluate the educational potential of the service • Address staff perception of the role that detection software pl st ff experience would play a key role in the dec r i became subscription-based in August 2005. Academics were inducted to the service participants. Here all software documentation and additional online resources which addresse web literacy, showcased student tutorials or linked to essay banks and ghost writing services were made available. Correspondence between the administrator of the training module and pilot participants occurred via email, online discussion board or phone. Staff were encouraged to use the software as an educational tool with their students, not solely as a detection tool durin periods of assessment. Staff were given the opportunity in an online environment to share wi one another their experiences and advice. A comprehensive survey at the end of Semester 1 marked the end of Phase I of the small scale pilot followed by a revised and slightly expanded pilot in Phase II.

The Turnitin ® UK pilot survey results
Turnitin®UK pilo o above stated pilot objectives for the pilo • The detection software service: All active service users (7/14 respondents) agreed that t software was easy to use and requir identified, but long return times (>24 hrs) during peak submission periods were criticised. The service was unavailable twice during Phase I, which proved frustrating. The small breadth of the database was criticised.
The educational potential of the service: All but one respondent agreed that plagiarism was a serious problem in their school.
drafts to the service, recognising the potential for improving the quality of their work. One respondent recommended student use be discouraged to prevent misuse.
Staff support: Respondents who did not use the service argued either that it was too tim consuming to learn, using conventional search engines was better or that plagi well-assisted, remarking that using the service saved them time and could act as a deterrent. Checking for collusion was found to be especially helpful. All active users anticipated using the service again and would recommend its use to colleagues. ACOs using the service commented that the additional cases of detected student plagiarism the school since piloting the service proved extremely stressful and hard to manage.
Staff expectations of service: In the comments section of the survey two respondents reported disappointment with the service for failing to check for matched text from

Additio
The Tu within mments by 30% of the spondents ranged from "Anything that may cause students to fail seems deeply unpopular" to y the inaction to even gross misconduct." And cynically, "What's the sk." about sity is in wider stitutional debate about academic integrity and plagiarism. The holistic, transparent and crossapproach taken by the Working Group on all fronts has proven successful in a en ot clearly nd of the first year of the Action Plan less than half of designated ACOs had regularly attended staff development seminars. Only two had regularly made use of wances ; • number as been increased during 2005/2006, now including sessions addressing problematic trends flagged by the featured school data, such as the apparent increases in ism ent ation nal survey results rnitin®UK survey exposed a recurring theme, namely the lack of extensive support the schools themselves for efforts to address plagiarism. Co re "Peers have felt let down b use? The penalties for plagiarism appear so slight that some students perceive it worth the ri Finally, one participant remarked "There are senior lecturers who don't seem to either care plagiarism or want to take the time to deal with it." In personal talks with affected pilot participants one of the authors confirmed that pressure to ignore plagiarism in order to submit grades in a timely fashion or improve retention rates was not uncommon.

Discussion
The second year of the Action Plan sees the University recognising its need to continue to plan strategically and enhance its practice in this area. Additionally, the Univer in institutional number of ways. Academic and support staff, Academic Board, management and student communities of practice have had the opportunity to contribute to the enhancement and development of the project. Their feedback has ensured cyclical consideration of measures and improvements to the Action Plan already within the first year. More improvements have be made for the second year based on formal and informal observations and discussions, summarised as follows.
• Academic Conduct Officers (ACO): Following the recommendations made by the featured school, ACOs were appointed. Their responsibilities, however, were n defined. By the e the detection software despite repeated efforts to include all ACOs in the pilot. The increased workload of the most proactive members of the ACO team has proven controversial. Half of the ACOs were changed for 2005/06, and guidelines drafted for all. All ACOs currently receive personal support from a member of Educational Development, and sessions are being held regularly to facilitate understanding of procedures, exchange good practice and to encourage networking. Workload allo are being discussed. Overall attendance at the staff development sessions has improved networking is increasingly vibrant. Four out of 15 ACOs now use Turnitin®UK regularly.
The five staff development events in 2004/2005 that addressed plagiarism were reasonably well attended but largely by the same members of academic staff: The of events h postgraduate offences and amongst international students. Though a new plagiar information website for staff and students was originally planned to go live during semester 1 of 2005/06, it has been postponed due to resourcing problems. New stud material has included informative handouts at matriculation, entries in programme handbooks and alerting screensavers at key times of the year, all additional to student diary, freshers' guide inserts and the posters previously available. Plagiarism inform and useful links are included in the institution's electronic library resources and in a newly developed, pilot online induction site for distance students (Napier, 2005). The Plagiarism Code of Practice is in operation university-wide in a revised format: Academic Board has recognised that revisions are an iterative process. More will Disciplinary Regulations of the University. In common with many institutions, such regulations are developed over a number of years and need a review of underlying philosophy rather than changes of detail. The Plagiarism Working Group, reconstituted as the Academic Conduct Working Group, is currently revising the SDRs, using the experiences of Carroll and Appleton (2005) as a basis. The major changes under consideration are: o a move from minor and major misconduct (which many seek to categorise numeric way o revising the penalties associated with breaches of academic regulations to properly reflect the modular structure of programmes; dealing with all breaches through a common struct la ned to present these changes to the Academic Board in t fo session 2006/07.
Turnitin ® UK: Despite the disappointing rate of actual software uptake by pilot participants (7 out of 14 sur face inductions to the service. This has been made compulsory for potential use order to promote good practice and dispel common misconceptions of the service which are a) Turnitin ® UK is a magic bullet; and b) Student use encourages misuse. The breadth of the database is steadily growing with currently 10% of academic staff and 7% of students enrolled as users. Eleven of the fifteen schools have been represented at the workshops. One school has imposed mandatory training for all its academic staff. At least one other is negotiating compulsory submission of all student coursework to Turnitin ® UK for the next academic year. The educative dimension of Turnitin ® UK is being recognised more widely, encouraged in part by a Scottish Funding Council (SFC project which includes the exploration of Turnitin ® UK's assessment features as a m of improving assessment practice in UK HE (SFC, 2005 to current news events, research findings and web resources are circulated to keep the engaged but busy academic abreast of relevant national developments regarding academic integrity, assessment and higher education practice. Because availability of the text matching system has been recognised as having significant implications for its use in teaching, learning, and assessment, workshops at many more levels are being sought. Turnitin®UK demons at LTA level have been carried out in selected schools and changes to management and operational procedures are being considered there. More sessions are planned, including student inductions to the service in the current semester and events as part of the general academic staff induction pro the next academic year.
Although the l effectively on an institutional basis, another school and Educational Development have chosen to llocate a member of staff research money for the area. Additionally, a second member of staff strator d by the 2004 survey carried out by FreshMinds and JISC PAS which reported a mere 3% detection rate of plagiarism (FreshMinds, ty nd in ink l values to tween the high expectations that HE has of the written work of its students nd the low quality academics may feel pressured to accept, one academic bemoaning, "We t base nd data set for wider University development and for University Action Plans. The first e of the Action Plan proved successful in raising awareness of the issues that r than rking Online Learning at apier University, Edinburgh. Originally Christina trained as a chemist, researching in spectroscopy at e adsorbate-metal interface. Teaching science and mathematics in traditional and online mode followed f e has taught many aspects of information systems and e-commerce, concentrating on the essons suggested by the featured school data have not yet been addressed a from Educational Development has been charged with assisting the Turnitin®UK admini to manage the institutional-wide implementation of the system, which also plans forward for a plugin to WebCT Vista by the next academic year. The number of academics who have not engaged with issues of educational integrity is a matter of concern. This is not a local problem, as evidence 2004). If continued, such a figure will thwart the impact of even the most exemplary of measures to address the issue. It would seem to be in the primary interest of all Universi educators to engage in efforts to reduce plagiarism, but reported school data and the Turnitin ® UK pilot survey results suggest otherwise. Why? Part of the answer can be fou the paper, "Why professors ignore cheating". There, Keith-Spiegel and Tabachnick (1998) l factors such as emotional stress, lack of time and institutional disinterest in traditiona academics' apathy.
A recent study carried out at the University College of Northampton, UK sheds some light on the growing disparity be a expect so much and accept so little" (Pickard, 2006, p. 217). The authors of this paper advise that the University featured here has chosen to engage strategically from evidence-based and quality enhancement perspectives rather than to ignore the issues. The authors further sugges that UK universities should be serious about promoting academic integrity, and that in order to do this they will need to recognise and respond to the requirements of students and of staff engaged in measures to prevent and confront plagiarism and other academic misconduct.

Conclusion
First lessons, innovation and record keeping within one school have provided an evidence a 2004/05 phas surround plagiarism and in the use of a consistent strategic framework to enhance support for both staff and students. The initial uptake of that support by academic staff has been slowe anticipated, the reasons for which are being explored currently in a separate study. The Wo Group and pilot studies, however, have already paved a way to increase education about and reduce the incidence of plagiarism. Central record systems are being updated in sensitive ways.
Regulations are being refined iteratively and, simultaneously, avenues of communication with the student body are being developed. Small research monies have been identified in order to continue to enhance collegial understanding of plagiarism as a phenomenon.