Responding to our article on the unreliability of multiple-choice tests (Teaching, THES, February 4), Gareth Holsgrove (Letters, THES, Feb 11) dismisses attempts to use penalty scoring (or negative marking, the deduction of whole or fractional marks for wrong answers) as means of minimising guessing: "The only thing they have in common is that none works."
That is far from true, but there can be problems. Students may score low if over-reluctant to take risks - notably when questions test more than one concept at a time, with knowledge sometimes "partial" (an intrinsic fault of many MCQ tests).
Our counter-argument is that if students do not trust their knowledge enough to answer, it is hard to assert that they functionally possess it.
Students may also be indignant at losing marks (but not at gaining them unearned) if the system and its alternatives are not fully explained. It helps to point out that, without penalty scoring, they are in effect penalised if they "underscore" by chance while their test-mates "overscore".
We wonder if the antipathy to penalty scoring shown by some examiners stems from a gut feeling that deducting marks is unfair. In its favour, penalty scoring (for examinees, examiner and teachers), with guessing thereby minimised, provides more accurate statistics on the individual and general level of understanding of each test item (that is whether answers are right, wrong or omitted). Finally, it avoids the intellectually dishonest obligation to guess.
Richard F. Burton David J. Miller Institute of biomedical and life sciences, Glasgow University.