Is cheating by answer copying habitual or opportunistic?
Written by: Dennis Maynes, Chief Scientist, Caveon Test Security
For some time, I have wondered whether cheating behavior is opportunistic or adopted. Do students cheat because they have acquired the behavior? Or, do they cheat because they were able to exploit an opportunity? In this short essay, I have investigated this question empirically. I shouldn’t rely upon folklore and common misperceptions to answer this question.
Experience has taught me students who fail and then retest have higher rates of non-independent test taking[1] than other students. When they fail again and again, they will have pressure to adopt cheating as a way to pass the test. This is easily explained because the stakes and pressure are high. The clock is ticking. Students must pass end-of-course tests before graduating. What about those students who feel no such urgency? How many will adopt cheating as a behavior?
At Caveon, we have begun research in this area. In this short essay, rates of independently taken tests for seventh graders who became eighth graders are analyzed. This group was chosen because they presumably do not experience significant pressure to pass the test. The overall rates of independently and non-independently taken tests are shown in Table 1.
Table 1: Independent test taking rates for 7^{th} graders who became 8^{th} graders
Column 1: The student took both the math and reading tests independently in 8^{th} grade | Column 2: The student did not take both the math or reading tests independently in 8^{th} grade | Percent of 7^{th} graders | |
Row 1: The student took both the math and reading tests independently in 7^{th} grade |
96.35% |
1.57% |
97.92% |
Row 2: The student did not take both the math and reading tests independently in 7^{th} grade |
1.92% |
0.16% |
2.08% |
Percent of 8^{th} graders |
98.27% |
1.73% |
100.00% |
In Table 1, about 2% of the tests were not taken independently each year. If there were no students who repeated the offense of non-independent test taking, we would expect the percentage in Row 2 and Column 2 to have been equal to 0.036%. Instead, the observed value was four and one-half times greater than expected[2]. Thus, it appears that a small percentage of these 7^{th} grade students who became 8^{th} graders (about one eighth of one percent) have adopted a pattern of non-independent test taking. It is important to observe that (1) about half of the students detected by the statistic were probably innocent, and (2) only about one percent of students at these grade levels appear to be cheating.
In answer to my question, it appears that opportunistic answer-copying or panic cheating was about seven times more likely than potential cheating by repeat offenders for this group of 7^{th} graders. Thus, while some students may have adopted a pattern of cheating, it appears that most answer-copying is opportunistic. Even so, I am delighted when repeat offenders decide to mend their ways. Last spring one student tweeted, “I used to be a heavy cheater, til i got my FCAT revoked.” I was gratified to read this because one of the purposes and uses of data forensics is to encourage students to take tests ethically, honorably, and with integrity.