
April 2007
Practical
Tips for Preventing Cheating on Online Exams
by David M.
Eplion, PhD, and Thomas J. Keefe, PhD
For as long as
there have been exams, there has been cheating on exams. Online
exams are no different, although they do provide some challenges
that set them apart from traditional face-to-face exams. These
include a heightened opportunity to collaborate with others, greater
possibility of using unapproved resources, and an increased
likelihood that someone other than the student is taking the test.
These problems are
not insignificant and do not have easy answers. In spite of these
very real challenges, we still wanted to offer online exams to our
students because they present a number of important benefits. These
include quicker and more accurate grading, more time to spend in
class covering important topics, and faster feedback for the
students. We decided that we needed to take steps to identify and
prevent cheating on our exams and we will share some of our
experiences in this paper.
Who we are,
what we do
We teach undergraduate business courses at a regional campus of
Indiana University. All our classes meet face-to-face for three
hours a week, but we administer our exams online outside of normal
class time. Our exams are both multiple choice and true/false in
nature, and the students take three of them throughout the semester.
On our syllabus, we
describe the nature of our online exams and also list a testing
protocol. During our first class meeting, we go over the protocol in
detail. We also include a section of the protocol on every exam. In
part, it reads, “This is not an open-book or open-notes exam. This
exam is to be taken during the allotted time period without the aid
of books, notes, or other students. You have approximately 45
seconds per question to complete this exam. This exam must be taken
online from start to finish. Do not download the test to take it or
distribute it to anyone. The statistics feature…will monitor and
report how you take this exam.”
These guidelines
ensure that our students know exactly what is expected of them. In
essence, they define for all involved what is and is not considered
cheating on our exams.
The
importance of courseware
Our first suggestion for helping would-be online exam administrators
detect cheating is to use professional courseware. We use a
proprietary system called Oncourse, but other commercially available
products such as Blackboard and WebCT are very similar.
Courseware packages
are valuable in many ways. To begin, they can be used to
password-protect the exam. This helps reduce the possibility that
unauthorized people will access the exam. Courseware also tracks the
exact time the exam was started and finished. It also can log IP
addresses that can be used to trace the location from which the exam
was completed.
Taken together,
this information can help an instructor find cases where students
may have collaborated on the exam. If two or more students take the
exam at or about the same time from computer terminals in proximity
(as indicated by the IP addresses), there is a possibility that
cheating has occurred, and we pursue these cases further.
Sometimes there is
a perfectly innocent explanation for this. Other times, it becomes
evident that cheating has occurred. However, because “cheating” is
such a charged word and can be difficult to prove, we avoid accusing
students of it. Instead, we try to document the fact that our exam
protocol has been violated.
One of our protocol
stipulations is that students not take the exam in the same room at
the same time. If we can document that this occurred, we do not
accuse the students of cheating. Instead, we simply tell them that
the protocol has been violated and that we will not accept the
results of their exam.
Courseware also
helps make sure timing protocols are not violated. We make our exams
available to students only during very narrow time windows. This
helps lessen their opportunity to recruit others to help them take
the exam. Courseware blocks access to the exam before it is
scheduled to begin and restricts access to it after the deadline.
Courseware also
keeps track of the total amount of time students spend taking the
exam. We have found that giving students an unlimited amount of time
to complete the exam creates more potential for collaborating with
others or for bringing in material that is not approved. Courseware
puts a strict timer on the exam.
If a student has
not finished the exam within the allotted time, courseware simply
submits to us the student’s progress to that point. For our exams,
we have found that allowing students 45 seconds per question gives
them enough time to think about and answer each question while not
providing so much time that they are tempted to violate our testing
protocol.
Test design
matters
While courseware is a valuable tool in the fight against cheating, a
poorly designed test can limit its effectiveness. Over the years, we
have found that there are ways to structure a test that help cut
back on cheating.
One strategy is to
scramble all the test questions for each student. Courseware also
does this for us. In addition, we are able to create a large
database of potential questions and then let Oncourse randomly
select a subset of questions for each student. It makes it very
difficult for students to collaborate when each student’s test is
markedly different from everyone else’s in the class.
We also prefer to
ask application-based questions as opposed to asking students to
just recite facts from the text. Application questions require that
students not only know the basic material, but that they also know
it well enough to apply it to practical situations. This helps
reduce the benefits of using unapproved materials on our exams.
Course
design can help as well
The final tool we use to minimize the impact of cheating is our
course design. In addition to the three exams we have students take
online, we also have them complete nearly 20 graded pretests.
Requiring that more than 20 assignments be completed online makes it
more difficult for students to recruit others to help them. It
either gets very time-consuming or very expensive for them.
We also require a
cumulative final exam that is heavily weighted and is administered
face-to-face. Because the cumulative final is worth so many points,
students must know the material if they hope to successfully
complete our course. If they have had someone else take the exams
for them over the course of the semester, they will pay for it when
taking the final.
Fortunately, we
have run statistical analyses that suggest a very strong correlation
between performance on the online exams and performance on the
face-to-face cumulative final. This suggests that we have had
success with our efforts to detect and limit cheating in our
classes. We plan to continue to employ these methods in the future.
Still, since no
strategies will totally eliminate cheating on exams, we are always
looking for new and improved ways to be even more effective. And we
welcome any suggestions or comments that the readers of this article
might have for us.
David M. Eplion
is an assistant professor of business management at Indiana
University Southeast. Contact him at deplion@ius.edu. Thomas J.
Keefe is an associate professor of business administration at
Indiana University Southeast.
|