Remarking of Examination Answer Scripts – Finding a Standard for Quality Assurance

The value of remarking of examination answer scripts is often limited to knowing how many changed symbols and ad hoc identification of the most glaring errors made by the initial markers. In any marking action performed by people there will always be the occasion where the marker has to make a judgment call on whether or not to award a mark for a specific answer. To be fair towards the candidates the markers often use a system of marking per question to minimize the effect of this subjectivity and attain better overall consistency. It is a generally accepted educational principle that there may be small variations in marking by different markers, even if they use the same memorandum. However, the question still remains of when is the difference acceptable and when not. With this paper the authors examine the changes in marks during remarking of a large number of subjects in a large-scale examination such as the Senior Certificate Examination in Gauteng over a number of years. The analysis is done in terms of changes in two-percent intervals on raw marks to determine a standard against which individual subjects can be measured. Raw marks are specifically used as to eliminate the effects of standardization. The analysis is spread over a number of years to determine patterns and eliminate once-off individual problems. The effect of deliberate interventions introduced by the province to improve the quality of marking is shown in terms of changes in remarking patterns. The remarking of the subject Physical Science of one cycle is analyzed in detail to show the nature and frequency of common marking errors made during initial marking, with reference to the relationship between the nature of the question (multiple choice, calculations, etc.) and the type of error.