Judy Sheard, Martin Dick, Selby Markham

January 2001

This is a report on a study that aims to determine attitudes towards questionable work practices of postgraduate and graduate diploma students within the School of Computer Science and Software Engineering (CSSE). This is part of a wider project that proposes to:

• establish undergraduate and graduate students’ understanding of what constitutes cheating and plagiarism and what practices are acceptable to them in this context
• determine the extent of cheating and plagiarism amongst undergraduate and postgraduate students
• identify motivations for cheating and plagiarism, and factors that influence cheating behaviour
• suggest measures which may be taken to discourage the practice of cheating and plagiarism

The longer-term aim of the project is to assist in determining measures that can be taken to address this problem and will facilitate the development of an informed policy on student plagiarism and cheating.

Selected postgraduate and graduate diploma subjects from CSSE courses were surveyed near the end of second semester 2000. A paper questionnaire was given to students in their tutorial classes. The questionnaire contained questions to determine:

• demographic information
• students’ rating of the acceptability of various questionable work practices described in18 different scenarios
• reasons which could cause cheating
• reasons which could prevent cheating

A copy of the Questionable Work Practices survey questionnaire can be found at: ???

152 students from a total population of 384 were surveyed and 103 valid questionnaires were returned. The students surveyed were enrolled in the graduate diploma subject CSE9000, or the postgraduate subjects CSE5000, COT4310 or SFT5610, or Honours.

The following tables show the numbers of student classified according to categories used in the analysis of the results from the surveys. NR is used to indicate that no response was given to this question.

The students were classified into 2 age groups as follows:

• born in 1975 or after (younger group)
• born before 1975 (older group)

From the above data the students were further classified into low performance (fail, pass, credit) and high performance (distinction, high distinction).

The students were asked to consider 18 different scenarios, each describing a questionable work practice. For each scenario they were asked to rate how acceptable the practice was, whether they had done it, and whether they personally knew someone who had done it. For the ratings of acceptability a 5-point Likert scale was used, where 1 indicates acceptable and 5 indicates not acceptable. NR is used to indicate no response was given to the question.

For the above scenarios t-tests were used to determine any significant differences (p <= 0.05) in the means obtained for the students’ ratings of the acceptability of scenarios when classified according to course type (postgraduate or graduate diploma), study mode (fulltime or part time), gender, age group, or average course performance to date (low or high). The t-tests were not performed on the scenarios for which the results had skewed distributions, or too high or too low kurtosis i.e scenarios 7, 8, 10, 13, 14, 15 and 18

The following results were obtained:

• There were no significant differences between the means of acceptability ratings students for any scenarios based on study mode, gender, or age group.
• The postgraduate students felt that scenario 3 was more acceptable than the graduate diploma students. This scenario described the practice of showing work to a lecturer with the intention of getting the lecturer to approve answers and gain hints on an assignment.
• The low performing students felt that scenario 11 was more acceptable than the high performing students. This scenario involved the practice of copying paragraphs of text from the Internet and using it in an essay without acknowledgement.

The most acceptable scenarios were 2 and 3. These presented situations where students obtained low level assistance with assignment work. Pearson’s moment correlation co-efficient calculated for these scenarios showed a strong relationship (r = 0.57, significant at the 0.01 level).

Other acceptable scenarios (in decreasing order of acceptability) were:

• 4 (resubmitting an assignment from a previous subject for assessment in a new subject)
• 1 and 12 (getting assistance with assignment work from a friend – more assistance than described in scenarios 2 and 3)
• 9 (discovering that an assignment mark was added up incorrectly to give a higher mark, and not telling the tutor)

All of the above scenarios were significantly more acceptable than the others.

The least acceptable scenario was 10. This involved theft of an assignment from a lecturer’s pigeonhole.

Other unacceptable scenarios (in increasing order of acceptability) were:

• 14 and 15 (exam cheating)
• 13, 18 and 7 (submitting complete assignments that were someone else’s work)
• 8 and 17 (submitting assignments containing significant parts which were someone else’s work)
• 16 (using fraud to obtain an extension to an assignment)

The scenarios describing work practices which the highest numbers of students admitted to having done were:

• 3 (showing assignment work to a lecturer for guidance)
• 2 (posting to an Internet newsgroup for assistance)
• 1 (two students collaborating on an assignment meant to be completed individually)
• 12 (copying the majority of an assignment from a friend’s assignment, but doing a fair bit of work yourself)

These four scenarios were rated in the top five most acceptable scenarios. The fifth scenario which involves resubmitting an assignment from a previous subject (scenario 4), probably doesn’t rate among the most practiced because it is likely that not all students have had the opportunity to do this.

The scenarios describing work practices which the highest numbers of students admitted to having known someone personally who had done them were:

• 3 (showing assignment work to a lecturer for guidance)
• 5 (submitting a friend’s assignment from a past running of the subject)
• 1 (two students collaborating on an assignment meant to be completed individually)
• 12 (copying the majority of an assignment from a friend’s assignment, but doing a fair bit of work yourself)
• 4 (resubmitting an assignment from a previous subject in a new subject)

The scenarios describing work practices which the least numbers of students admitted to having done were:

• 14 (hiring someone to sit an exam for you)
• 7 (hiring a person to write your assignment for you)
• 10 (taking a student’s assignment from a lecturer’s pigeonhole and copying it)
• 18 (swapping assignments with a friend, so that each does one assignment, instead of doing both)
• 13 (copying all of an assignment given to you by a friend)
• 16 (obtaining a medical certificate from a doctor to get an extension, when you are not sick)
• 6 (being given the answer to a tutorial exercise worth 5% by a class mate if the computer you used has problems)
• 8 (copying another student’s assignment from their computer without their knowledge and submitting it)
• 15 (using a hidden sheet of paper with important facts during an exam)

These scenarios were also all rated as highly unacceptable.

The scenarios describing work practices which the least numbers of students admitted to having known someone personally who had done it were:

• 10 (taking a student’s assignment from a lecturer’s pigeonhole and copying it)
• 14 (hiring someone to sit an exam for you)
• 7 (hiring a person to write your assignment for you)
• 8 (copying another student’s assignment from their computer without their knowledge and submitting it)
• 18 (swapping assignments with a friend, so that each does one assignment, instead of doing both)

For this question students were asked to indicate the likelihood that each reason would cause them to cheat. A 5-point Likert scale was used, where 1 indicates not at all and 5 indicates highly likely. NR is used to indicate no response was given.

t-tests were used to determine any significant differences (p <= 0.05) in the means obtained for the students’ ratings of likelihood of each reason causing cheating when classified according course type (postgraduate or graduate diploma), study mode (fulltime or part time), gender, age group, or average course performance to date (low or high). The t-tests were not performed on the scenarios for which the results had skewed distributions, or too high or too low kurtosis.

The only significant differences were between full and part time students. The full time students indicated that exams that were too hard and feeling that they couldn’t afford to fail would be more likely to cause them to cheat than the part time group.

For this question students were asked to indicate the likelihood that each reason would prevent them from cheating. A 5-point Likert scale was used, where 1 indicates not at all and 5 indicates highly likely. NR is used to indicate no response was given.

t-tests were used to determine any significant differences (p <= 0.05) in the means obtained for the students’ ratings of likelihood of each reason causing cheating when classified according to course type (postgraduate or graduate diploma), study mode (fulltime or part time), gender, age group, or average course performance to date (low or high). The t-tests were not performed on the scenarios for which the results had skewed distributions, or too high or too low kurtosis.

The only significant differences were between the two age groups. The older student group indicated that not knowing how to cheat and cheating being against their moral values would be more likely to prevent them from cheating than the younger student group.

No further analysis was done as these results were positively skewed and had high kurtosis.

Cross tabulations were performed to determine any significant differences (chi-squared <= 0.05) in the responses classified according to course type (postgraduate or graduate diploma), study mode (fulltime or part time), gender, age group, or average course performance to date (low or high).

These showed that the male students would be more inclined to write an assignment for a fee than the female students, and the fulltime students would be more inclined to write an assignment for a fee than the part time students.

There were no differences in responses based on course type, age group or average course performance.

Cross tabulations performed showed that there were no significant differences in the responses classified according to course type (postgraduate or graduate diploma), study mode (fulltime or part time), gender, age group, or average course performance to date (low or high).

No further analysis was done as these results showed an extreme modal value.

No further analysis was done as these results showed an extreme modal value.

For the above two questions, t-tests showed that there were no significant differences (p <= 0.05) in the means obtained for the students’ ratings of how strongly their teachers or the University feel about cheating when classified according course type (postgraduate or graduate diploma), study mode (fulltime or part time), gender, age group, or average course performance to date (low or high).

A Pearson’s correlation between these two questions showed a strong relationship between the responses (r = 0.69, significant at the 0.01 level)

Cross tabulations were performed to determine any significant differences (chi-squared <= 0.05) in the responses classified according to course type (postgraduate or graduate diploma), study mode (fulltime or part time), gender, age group, or average course performance to date (low or high).

These showed that the younger age-group students are more aware of the university regulations than the older age group.

There were no differences in responses based on course type, study mode, gender or average course performance.