During my first year as an assistant professor, while I was teaching an introductory child development class, two students requested to take a make-up exam. They both presented written documentation for their absence, following the instructor’s policies stated in my syllabus. I arranged to have the teaching assistant (TA) proctor their make-up exam. The TA met both students, put them in separate but adjacent rooms, asked that all personal belongings be left outside the testing room, and instructed the students to submit the completed exam to a receptionist when they retrieved their belongings. The TA planned to pick up the completed exams from the receptionist and to score the multiple-choice section before returning both exams to me. The exam consisted of multiple-choice and short-answer questions, plus two essays. Each student wrote a predetermined “codename” on the exam pages so that the exams could be graded blindly, and then exam grades were recorded on a master sheet that linked the codename to the student ID number. After each exam, I would complete an item-analysis of the multiple-choice items to determine which items might be bad (e.g., poorly written, confusing), too easy, or even too difficult (e.g., less than 10% correct response) in order to maintain or discard them from the test bank I used for subsequent exams. The exams were graded by section, not by exam, to ensure consistency in applying the rubric for nonobjective items (e.g., short-answer and essays).

As I was recording the point totals by section for the two make-up exams I noticed that each student received the same total score – 77.5 points out of 100 – although scores for each section varied slightly (e.g., 40 and 41 points out of 49 on the multiple-choice section, respectively, for each student). Then I noticed that of the multiple-choice items that were marked incorrect for each student, 10 items for one student and 11 for the other, they both incorrectly answered the same six items, and each selected the same incorrect response option. I was stunned by this pattern and was suspicious that the two students collaborated, especially since these test scores were between one-half and one full grade higher than their previous exam score, but I wanted to be sure. As a new assistant professor, I had not previously had to deal with student cheating.