RESEARCH REPOSITORY
This is the author’s final version of the work, as accepted for publication
following peer review but without the publisher’s layout or pagination.
The definitive version is available at
:
http://dx.doi.org/10.1007/s10805-017-9278-x
Curtis, G.J. and Clare, J. (2017) How Prevalent is Contract
Cheating and to What Extent are Students Repeat Offenders?
Journal of Academic Ethics, 15 (2). pp. 115-124.
http://researchrepository.murdoch.edu.au/id/eprint/36543/
Copyright: © 2017, Springer Science+Business Media Dordrecht
It is posted here for your personal use. No further distribution is permitted.
1
RUNNING HEAD: PREVALENCE OF CONTRACT CHEATING
How prevalent is contract cheating and to what extent are students repeat offenders?
Guy J. Curtis (corresponding author)
Murdoch University, School of Psychology and Exercise Science, Murdoch University, 90
South St., Murdoch, Western Australia, 6150, Australia. g.curtis@murdoch.edu.au, Ph: +61 8
9360 2268, fax: +61 8 9360 2253
Joseph Clare
Murdoch University, School of Law
THIS PAPER MAY BE CITED AS FOLLOWS:
Curtis, G. J., & Clare, J. (2017). How prevalent is contract cheating and to what extent are
students repeat offenders? Journal of Academic Ethics, 15, 115-124.
doi:10.1007/s10805-017-9278-x
2
Abstract
Contract cheating, or plagiarism via paid ghostwriting, is a significant academic
ethical issue, especially as reliable methods for its prevention and detection in students’
assignments remain elusive. Contract cheating in academic assessment has been the subject
of much recent debate and concern. Although some scandals have attracted substantial media
attention, little is known about the likely prevalence of contract cheating by students for their
university assignments. Although rates of contract cheating tend to be low, criminological
theories suggest that people who employ ghostwriters for their assignments are likely to re-
offend, and little is known about re-offence rates in this form of academic misconduct. We
combined previously-collected datasets (N = 1378) and conducted additional, and previously-
unreported, analyses on self-report measures of contract cheating prevalence. We found that
few students (3.5%), on aggregate, ever engaged in contract cheating but this varied
substantially among samples (from 0.3% to 7.9%). Of those who ever engaged in contract
cheating, 62.5% did so more than once. The data also suggested that engagement in contract
cheating is influenced by opportunity. These figures may help policy makers, and researchers
who are creating contract cheating detection methods, to estimate base rates of contract
cheating and the likelihood of re-offence.
Keywords: contract cheating, ghostwriting, prevalence, plagiarism, academic integrity
3
Contract cheating is a breach of academic ethics in which students enter into an
agreement to pay another person to complete assessments on their behalf (Clarke and
Lancaster 2006; Walker and Townley 2012). Typically, this involves paying the other person
to write an unsupervised assessment such as an essay, report, or computer code (Clarke and
Lancaster 2006; Walker and Townley 2012). Contract cheating appears to be a subset of the
form of plagiarism defined by Walker (1998) as “ghostwriting”. Walker defined plagiarism
via ghostwriting as occurring when students submit work written by another person as if it
was their own, but this definition is silent on the issue of payment, which is an inherent
feature of contract cheating. Contract cheating has been a particularly hot topic of late, both
in the academy and in the media.
A recent headline-grabbing scandal in Australia was the discovery of a custom-
writing website called MyMaster (Visentin 2015). The numbers of students involved
appeared to be high, with reports of nearly 1000 students using the site (McNeilage and
Visentin 2014). However, such reports of the prevalence of contract cheating may be
misleading when framed in terms of the number of students doing it, because these figures do
not take into account the size of the group from which the students are drawn. The university
identified as having the most potential student users of the MyMaster contract cheating
website had 128 requests to the site for assignments (McNeilage and Visentin 2014).
Although the raw number of 128 requests appears to be shocking, this would represent only
0.3% of students at that university if each request was from a unique student. Unfortunately,
the prevalence of contact cheating can appear to be overblown with the media’s focus on raw
numbers; one thousand students cheating makes a better headline than 3 in every 1000
students cheating. And, such media reports may cause the kind of panic over contract
cheating that academic authors on the topic suggest is unwarranted (Walker and Townley
2012).
4
Walker and Townley have stated that “the prevalence of contract cheating is
unknown” (2012: 27). This statement seems to be too definitive. From our assessment of the
literature, we would, instead, contend that little is known about the prevalence of contract
cheating. The aim of our report is simple; we wish to contribute some new information to the
literature on the prevalence of contract cheating. This aim is consistent with the recent plea of
Wallace and Newton, who state, in relation to contract cheating: “[T]he single greatest need
is for more high-profile research in this area, to educate educators about the existence and
detail of the problem” (2014: 236).
There are some important reasons why estimating the prevalence of contract cheating
among university students may be worthwhile. First, it may be helpful to policy makers to
have some guidance as to likely the extent of the problem. Second, some potential avenues
for automated detection of contract cheating may rely on estimates of its prevalence.
Busy academics in massified higher education systems have limited time to
investigate suspected academic misconduct, and often rely on text-matching software to assist
in the detection of plagiarism (Walker and Townley 2012). However, automated methods for
detecting contract cheating are not well-formed or in widespread use, partly because these
require, but lack, data on the prevalence of contract cheating. For example, textual analysis
methods can estimate the probably that two documents (e.g., an exam where the student’s
identity is verified and an unsupervised essay) were penned by the same author (Afroz et al.
2014; Koppel and Winter 2014; Stamatatos et al. 2015). But, should an academic trust a
computer-generated estimate that 75% of his or her students have engaged in contract
cheating? Similarly, persistent discrepancies in marks between unsupervised and supervised
assessments may act as red flags for potential contract cheating (Clare 2016). But, should an
academic trust a computer-generated finding that students who have never before been
accused of using ghostwriters have done so? Such methods use probabilistic decision rules
