Effects of Apophenia on Multiple-Choice Exam Performance

There is a broad literature on the various issues related to effective exam construction applicable to both on-ground and online course delivery. These guidelines tend to support rather close contact between the instructor and the exam. However, to remain competitive, both textbook and course management providers have developed technologies to automate many aspects of exam construction. As test construction becomes automated, the possibility of inadvertently deviating from demonstrated or intuitive guidelines increases. Two experiments were conducted to examine the degree to which apophenia (perceiving patterns in random data) might negatively influence multiple-choice exam performance among college students. Experiment 1 indirectly demonstrated the extent to which certain answer patterns seemed to be tolerated among students (maximum of three repeated answers) in comparison with what might be expected from randomly generated exams from Blackboard. Experiment 2 directly examined the effects of answer patterns on exam performance. Participants' performance declined as the underlying answer patterns became more obvious, and this effect appeared to be particularly strong for the upper level psychology students. The importance and implications of these findings with regard to automated test construction were discussed, and a recommendation is provided.