Language as a confounding variable in the diagnosis of misconceptions

The study reported in this paper examines the possibility that language con fusion can be mistaken for misconception. A written test, consisting of 20 multiple choice items that had been used by other authors to diagnose misco nceptions in physics was administered to a sample of matric level, English first language, South African students. A sub-sample of these subjects was then interviewed to explore the reasoning behind their choices in the writt en lest. An analysis of the interviews showed that in several cases, distra cters were chosen by subjects who did not hold the target misconceptions th e items were intended to diagnose, and that language usage, for example mis interpretation of the question text, was frequently the reason behind the c hoices. The findings indicate that caution is needed in interpreting studen ts' responses to multiple choice questions.