Beruflich Dokumente
Kultur Dokumente
I. INTRODUCTION
VER one in ten students surveyed have admitted
to copying programs in courses with computer
assignments [1]. In those technology-oriented courses, where
much of the work is in digital format, copying coursework
has become virtually effortless. It is not surprising then at
some universities disproportionately large shares of violations to academic integrity have been in technology-oriented
disciplines [2]. The issue is further complicated in large classrooms where the increased number of students makes detecting
plagiarized work even more difficult due to time constraints [3].
So how can cheating be effectively reduced without either radically changing course designs, or requiring draconian measures
which interfere with the learning process?
This study examined a large technology-oriented course,
which used a technological tool to detect plagiarism in direct
response to blatant plagiarism occurring in course projects
using Microsoft Access. The change in students behavior,
in terms of the number of students detected who plagiarized
projects, was analyzed at points before and after the implementation of the tool. This paper illustrates how technology can
help alleviate some difficulties in discovering plagiarism and
enforcing policies against plagiarism.
This paper describes the course, the problem, the tool, and
the process for using the tool. An analysis of seven semesters
MCCART AND JARMAN: TECHNOLOGICAL TOOL TO DETECT PLAGIARIZED PROJECTS IN MICROSOFT ACCESS
167
168
TABLE I
PROPERTIES
MCCART AND JARMAN: TECHNOLOGICAL TOOL TO DETECT PLAGIARIZED PROJECTS IN MICROSOFT ACCESS
169
TABLE II
EXAMPLE 1
TABLE III
EXAMPLE 2
If the database titles were identical and were not the default
title (db1), those projects were marked as duplicates. Table II
shows an example of three projects meeting these criteria.
Three students each submitted their individual projects and
each project has a unique filename. However, the projects
A. Example 1
170
TABLE IV
EXAMPLE 3
TABLE V
EXAMPLE 4
B. Example 2
If students submitted individual projects with identical database creation dates and titles and the titles were the default title
of db1, those projects would be marked as potential duplicates
(Fig. 2, step 3b). Table III shows this example. Student Foo and
student Bar potentially copied projects because the database creation dates and titles are identical for both projects. However,
since the titles are the default db1, it can not be determined
with certainty if the two projects are duplicates. The likelihood
that two projects would have identical creation dates and neither
student changed the database title from the default is remote but
it is possible. Therefore, the percentage of likeness and other details need to be examined to determine if projects meeting these
criteria were copied. These details will be discussed later in the
paper.
C. Example 3
Suppose two projects were submitted that had identical database creation dates but the database titles were different. The
possibility still exists that these two projects were copies. Further details had to be examined to determine if this were the case
(Fig. 2, step 4). The database title can be changed; however, the
need to do this is not intuitively obvious. Most students incorrectly assumed that changing the actual filename of the database
was enough to change the title. On rare occasions, students did
change the title after copying the database from another student.
Table IV shows an example of this.
If the database creation date and title were the only properties examined, one would assume that these students just happened to create their projects at exactly the same time. This
situation is unlikely, but possible. CCPE examined one other
property when such coincidences were discovered to determine
whether projects were duplicates. As discussed previously, the
database has a last updated properties date that identifies when
the properties of the database rather than the actual data have
been updated.
In Table IV, the database creation date is identical for the
two projects, yet the titles are different. CCPE then looks at
the last updated properties date. Student Browns last updated
properties date has been changed from its initial value, which
would have been the same as the database creation date. This
difference in dates shows that the title was probably changed and
MCCART AND JARMAN: TECHNOLOGICAL TOOL TO DETECT PLAGIARIZED PROJECTS IN MICROSOFT ACCESS
171
VIII. ANALYSIS
Projects from 19 assignments over seven semesters were collected and processed with the CCPE tool. Averaged over 10
runs, the typical time to process an assignment consisting of 627
projects took less than 19 seconds, resulting in a substantial decrease over the time required for the manual method.
Projects were categorized based on when the student took
the course (i.e., fall and spring semesters versus summer
semesters). Since students from other universities often take
the course during the summer, the categories were created for
more consistent comparisons between semesters. Students that
attended this university during the regular school year had ties
to each other, which provided them more opportunities to cheat
than students from other universities had.
Table VI and Fig. 5 present the results of the fall and spring
semesters, while Table VII and Fig. 6 present the results of the
summer semesters. Each table lists the number of projects submitted and how many were detected as duplicates using CCPE,
while the figures show the change in percentage between
semesters. The group size indicates the number of projects
that were copied from one project, including the original. For
instance, a group size of five means there was one original
project and four other students copied that original project.
In fall 2004, 6.7% of all projects turned in were direct duplicates of another project. Fall 2004 was also the only semester
with groups of four and five students copying the exact same
project. As CCPE was not yet being used, fall 2004 is considered the baseline for fall and spring semesters without any technological mechanisms to detect plagiarized projects.
CCPE was first introduced and used in spring of 2005. Each
regular school year semester was compared to the baseline
semester using a two-proportion z-test to determine if there
was a significant difference in the proportion of students
plagiarizing work between semesters [13]. Additionally, the
average percentage of duplicate projects was compared to the
baseline semester. Semesters using CCPE had a significantly
compared to
lower percentage of copied projects at
the baseline semester (see Table VIII).
172
TABLE VI
FALL AND SPRING SEMESTERS
TABLE VII
SUMMER SEMESTERS
TABLE VIII
FALL AND SPRING SEMESTER COMPARISONS
TABLE IX
SUMMER SEMESTER COMPARISONS
In summer 2004, CCPE was also not yet being used. Therefore, summer 2004 is considered the baseline for summer
semesters without any mechanisms to detect duplicate projects.
Like the regular school year, summer semesters using CCPE
MCCART AND JARMAN: TECHNOLOGICAL TOOL TO DETECT PLAGIARIZED PROJECTS IN MICROSOFT ACCESS
173
James A. McCart received the B.S. degree in information systems from Purdue
University, West Lafayette, IN, and the M.S. degree in management information
systems from the University of South Florida (USF), Tampa, in 2002 and 2006,
respectively.
He is working towards the Ph.D. degree in management information systems at USF. His research interests include software testing, security, and
development.
Jay Jarman received the B.S. degree in computer science from East Tennessee
State University, Johnson City, and the M.S. degree in management information
systems from the University of South Florida (USF), Tampa, in 1996 and 2006,
respectively.
He is working towards the Ph.D. degree in management information systems
at USF. He served eight years in the U.S. Air Force in the information management field. He has 20 years experience in developing and managing information
systems as well as teaching and developing information technology courses for
software companies.