A mathematical perspective on how test-type exams are evaluated

Abstract

Typically, the test type exams with binary answers (true / false) are corrected by adding a +1 to the correct answers, a -1 to the incorrect ones and 0 to non-answered questions. This mainly obeys to a purely statistic criterion which incentives answering questions randomly when the student has the impression that she will not pass the exam with what she knows. From our point of view this method of qualification is not suitable and therefore should be amended as soon as possible. Unfortunately, it is still now massively used in many Faculties worldwide. The objective of the work is to provide an alternative method to accurately qualify such exams.