Communication Cite This: J. Chem. Educ. XXXX, XXX, XXX−XXX
pubs.acs.org/jchemeduc
Crime in the Classroom: Analysis Over 26 Years David N. Harpp* Department of Chemistry, McGill University, Montreal, Quebec H3A 0B8, Canada ABSTRACT: Copying adjacent answer sheets on the oft-used multiple-choice exams can be prevented if exam material is scrambled. The latter mode has been a requirement for all exams of this type at McGill University for 26 years. Cheating by copying has been reduced to nearly zero in this time. The experiment carried out used permissive seating and a single version to see if students would still try to copy adjacent answer sheets. With over 1,400 students writing, there were no prosecutable cases detected. KEYWORDS: Ethics, Testing/Assessment
■
as a “Campus MOOC” in which students watch the online content on their own time, but report in person for two traditional multiple-choice exams and a final exam. The course has been offered in this fashion for the past three years, with enrollments of 652, 1,010, and over 1,400 this past semester (winter, 2017). In keeping with University anticheating regulations, 4 different exam versions and assigned seating were used for the exams up until this past semester’s (winter, 2017) final exam. Here, we departed from university exam procedures and used only one version of the test. The identical tests were labeled as though there were still four different versions. We also allowed students to select their own seats. Students were not informed about this shift in protocol. The goal of the experiment was to determine whether the culture (attitude) on campus that generated ∼5% exam copying that existed prior to 1991 had shifted since implementation of the anticheating regulations.2,5 The exam was carried out with 1,401 students taking a test of 82 questions. Using an excellent, statistically based analysis program8 similar to the one we originally designed,2 980,700 pairs of exams were compared for common incorrect answers. Results revealed that only two pairs of student exams emerged with any significant statistical chance of collusion. One pair had a standard deviation slightly higher than our established threshold to flag potential cheating, but we confirmed that these two students sat in different rooms, likely eliminating any chance of nonelectronic collaboration. The second pair was seated adjacent to one another, but the statistical similarity of their exams along with other criteria critical to identifying extensive copying3 fell well short of a clear-cut case of cheating. No cheating-detection program is capable of finding a small number (∼5−6) of copied answers unless the pair is essentially the only one in the class with those exact incorrect responses. Further, with nearly a million pairs for comparison, statistical anomalies are more likely encountered than with smaller sample sizes.
INTRODUCTION During multiple-choice exams, the copying of adjacent optical (bubble) answer sheets is a time-dishonored method of cheating that is as old as the exam method itself.1 In the first of our four articles on the topic of cheating on multiple-choice exams,2−5 we described our development of an effective method for detecting copying. The program we designed compares common wrong answers between every pair of students in the class; a subsequent statistical analysis highlights pairs of students likely to have shared answers. Analysis of exams both at McGill and at several other universities revealed that typically about 5% of exams showed evidence of extensive copying.2,5 Results were confirmed by statistical analysis coupled with confirmation that students whose exams were flagged for cheating were seated either next to, or behind, each other.6 Because of the program’s effectiveness at detecting copying, the Senate of McGill University adopted a series of anticheating procedures that were implemented for all multiple-choice exams beginning in 1991. The policies, which include exams with questions in scrambled ordered, “striped” seating (different courses seated in adjacent rows in exam rooms), as well as assigned student seating, have been in place since that time. In 1999, McGill’s Senate validated such data as allowable evidence in the prosecution of academic dishonesty cases. As a consequence of the careful implementation of these procedures at McGill for midterm exams and especially during final exams, copying has been reduced to nearly zero during this 26-year period. Periodic checks using the program on all pairs of students taking the exams nearly always result in no evidence of copying by any pair. Furthermore, (anecdotally) McGill students routinely convey that they do not experience their colleagues even talking about trying to cheat on exams of this type. We recently performed an experiment to determine whether this perceived shift in campus “culture” away from this form of copying has indeed taken place.
■
THE EXPERIMENT In 2014, a massive open online course (MOOC) was created for edX entitled “Food for Thought”. It was born from a similar course offered since 1981 at McGill,7 and had been offered 4 times as a public MOOC. The course was subsequently adapted © XXXX American Chemical Society and Division of Chemical Education, Inc.
Received: July 7, 2017 Revised: October 23, 2017
A
DOI: 10.1021/acs.jchemed.7b00426 J. Chem. Educ. XXXX, XXX, XXX−XXX
Journal of Chemical Education
■
Communication
(4) Harpp, D. N.; Hogan, J. J. Crime in the Classroom. Part III. The Case of the Ultimate Identical Twin. J. Chem. Educ. 1998, 75, 482− 483. (5) Harpp, D. N. Crime in the Classroom Part IV. Conclusions. J. Chem. Educ. 2008, 85, 805−806. (6) Wikitionary entry for “the proof of the pudding is in the eating”. https://en.wiktionary.org/wiki/the_proof_of_the_pudding_is_in_ the_eating (accessed Nov 2017). (7) Fenster, A. E.; Harpp, D. N.; Schwarcz, J. A. Chemistry for the Public. Part III. “The World of Chemistry”. A Course for All. J. Chem. Educ. 1993, 70, 819−821. (8) Wesolowsky, G. O. Detecting Excessive Similarity in Answers on Multiple Choice Exams. Journal of Applied Statistics 2000, 27, 909− 921. (9) Stoltzfus, K. To Stop Exam Cheats, Economists Say, Try Assigning Seats. Chron. High. Educ. 2015, 62 (8); http://www. chronicle.com/article/To-Stop-Exam-Cheats/233741 (accessed Nov 2017).
CONCLUSIONS The idea was to try to discover if the students’ peers were still motivated to take advantage of the exam system. In the two gymnasia where this exam took place, it would not take extraordinary measures for students to switch versions when the seating in such a mass setting was taking place. Further, a cursory examination of adjacent exams could have shown identical patterns of paragraph size and shape, which might trigger copying of the adjacent answer sheet. The lack of any evidence of cheating on the exam (far below the average 5% rate detected prior to 19912,5) suggests that consistently applied, clearly communicated institutionalized exam regulations are effective to not only prevent copying, but also result in a measurable shift in campus away from this method of academic dishonesty. The students receive a clear message that McGill University cares about academic honesty and is willing to make the effort to protect the credits students earn from this type of examination. We suggest that it is the institution-wide adoption of anticheating exam protocols9 that has resulted in the positive change in student behavior. Two referees appropriately noted that the lack of cheating could be attributed to two different phenomena. Either students cannot easily get away with the “crime” owing to the anticheating protocols, or the protocols (having been in place for many years for all exams at McGill) have removed the temptation to even try to cheat. I believe the latter is the more likely conclusion, because all McGill multiple-choice exams have been scrambled for more than a quarter century. However, the experiment that both referees suggested should be tried. The exam should have permissive seating as well as identical exams either with the same version number or no designation at all. What is also of interest is that with essentially zero reports of cheating over this long time period there are very rarely campus or other legal issues to adjudicate. Such cases are not only complicated to prosecute, but painful to experience when cheating is demonstrated.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. ORCID
David N. Harpp: 0000-0001-8228-6038 Notes
The author declares no competing financial interest.
■
ACKNOWLEDGMENTS The author thanks McGill University for continued vigilance and support of academic integrity and Professor K. S. Harpp, Colgate University, for helpful discussions.
■
REFERENCES
(1) Trewin, S. A.; Fuchs, A. H. History of Psychology: Robert Yerkes’ Multiple-Choice Apparatus, 1913−1939. Am. J. Psychol. 2007, 120, 645−660. (2) Harpp, D. N.; Hogan, J. J. Crime in the Classroom. Detection and Prevention of Cheating on Multiple-Choice Exams. J. Chem. Educ. 1993, 70, 306−311. (3) Harpp, D. N.; Hogan, J. J.; Jennings, J. S. Crime in the Classroom. Part II. An Update. J. Chem. Educ. 1996, 73, 349−351. B
DOI: 10.1021/acs.jchemed.7b00426 J. Chem. Educ. XXXX, XXX, XXX−XXX