Fairness in using negative marking for assessing true/false questions.

Abstract

Multiple choice questions are a popular method of testing as they are easy to mark. However, they also lend themselves to guessing. True/False questions are particularly prone to guessing. This can be alleviated by deduction of marks for incorrect answers. The University of KwaZulu-Natal currently uses true/false questions in its assessments, which are answered by means of pencil-and-paper sheets that are optically scanned, marks being calculated according to a standard formula. A trial of a proposed new computer program revealed mark discrepancies compared to the current scheme. This study evaluates the program’s marking scheme ‘internal negative marking’ – whereby an overall negative mark for an item consisting of a stem and several true/false questions is ‘rounded up’ to zero. We enumerate the cause of the discrepancies, demonstrate that the latter scheme diminishes the penalty for guessing and may encourage strategic students to leave out parts of the curriculum when studying.