Using multiple test trials was something I had never considered until found myself in a newly assigned course with an old syllabus. The previous course, which consisted of 310 total points, included 140 (45 percent) testing-based points. In addition to a 100-point final exam, there were four 10-point quizzes. I was intrigued by the quiz design format that allowed students to take the quiz up to three times over the course of a week, with the average score added to the grade book.
I wondered whether allowing multiple test attempts was an educational folly. But for 13 percent of the course points, I decided it was worth investigating patterns of student usage and performance before dismissing a new idea.
Of the 20 students enrolled in the course, multiple trials (at least two) were utilized by 13 (65 percent) in the first quiz, 12 (60 percent) in the second quiz, eight (40 percent) in the third quiz, and seven (35 percent) in the fourth quiz. When comparing the average first attempt to the average multiple trial score, I found variation in student performance. Students improved their scores by an average of .71 points (close to a full letter grade) with multiple attempts on the first quiz versus a negative result (-.03) on the third exam, and two smaller gains for quizzes 2 (+.23) and 3 (+.24).
The syllabus also gave students an option of stopping if they liked their scores on the first attempt. Using a 10 or 9 (the lowest A performance for the course) as a cutoff score, I found that nine (45 percent) students stopped with a “liked” score in the fourth quiz compared with three (15 percent) students in the first quiz. In the first quiz I glimpsed a gamesmanship strategy I found fascinating: a student who earned 8 points in the first attempt and scored the maximum in the second attempt. I thought the student would have “held” with an average score of 9. But the student used the third trial to score 10 points again, raising the average to 9.33. In the second and fourth quizzes, I saw a magnification of this effect: multiple 9/10/10 patterns. Rather than holding with a 9 “A-hand,” a student decided to draw two other quiz cards to move to a 9.67, a stronger A.
The final exam had been designed along a more traditional format: a onetime shot during a time limit of 90 minutes within an 11-hour window. After several polite student requests for an extended time because of multiple exams, I decided to use the same multiple-attempt format, allow an additional 30 minutes to take the exam, and lengthen the time window to 25 hours. I was pleased to find that students who requested additional time available took the exam within the extended testing window. As for multiple final exam attempts, only three (15 percent) students took advantage of the opportunity; they earned an average improvement of 6 points (half a letter grade).
I ultimately decided to allow multiple final exam trials when my inquiry toward the end of the course showed a spread in student performance. After the final, student performance variation continued. Five (25 percent) students who tested well (scored high and/or took advantage of multiple attempts) and submitted strong individual projects (55 percent of the grade) earned top grades. Of the other 15 students, I estimate 13 (87 percent) could have earned at least a half grade higher by taking advantage of multiple test trials. I was amazed at the many missed opportunities.
An interesting benefit came after the semester ended when a “grovel” email arrived from a student who requested rounding up her percentage to the next grade. In showing I was unable to replicate her higher percentage, I pointed out that of five testing situations in which she had multiple attempts available, she had taken advantage only two times. As I looked more closely at the student’s online pattern, I found an irregularity: the student had an open final exam first attempt with scores for second and third trials. I felt comfortable indicating that the student could have earned a better grade, but there were many lost opportunities. The biggest benefit of multiple testing attempts came when the course evaluation report showed high student ratings. The goodwill generated by the security of having additional testing opportunities available was viewed positively by students.
Despite my initial skepticism, my one-course investigation of student usage and performance provides compelling evidence for applying multiple attempts to some testing situations in the future. A final exam might not be the best option in some courses, but using multiple-trial testing worked well as an instructor-pacing method. Even if students decide not to take tests more than once, multiple test trials provide a learning opportunity.
Aimee J. Luebben, EdD, OTR, FAOTA is a professor of occupational therapy at the University of Southern Indiana.
Excerpted from Offering Multiple Test Trials: Educational Folly or Learning Opportunity? November 2008, Online Classroom.