Effective discrimination in mark schemes
Effective assessment is often measured in terms of mark re-mark reliability: the extent to which the same mark would be awarded to a candidate by two different examiners or, alternatively, by the same examiner on two different occasions. Reliability, however, can come at the expense of validity: the extent to which an assessment tests the skills it purports to test.
The validity of an assessment can be compromised by any element of the specification, the examination paper or the mark scheme. With the use of a hypothetical illustration and a case study, this paper explores the effect that a mark scheme can have on the validity of an assessment; focussing particularly on a mark’s worth across the mark range for an item.
It proposes a number of techniques which might be used to identify areas of weakness in a mark scheme and areas where the mark scheme does not provide for effective discrimination between candidates.