THE MOST COMMON PROBLEMS IN ASSESSMENT PLANNING

Abstract

Learning about resolving common problems means that modes of assessment become less deterministic – real world problems rarely pop up in a form that can be readily solved by an equation or two. This often means the use of projects, portfolios and dissertations. Even simple mathematical problems can give rise to differences of opinion in marking but in the case of the less structured assessments – where answers may be ‘better’ or ‘worse’ rather than ‘right’ or ‘wrong’ – the assessment issues are much starker. When classes are large and the number of assessors equally substantial there are issues of reliability. This paper looks at some of the issues around the reliability of assessment in such circumstances and uses a large sample, drawn from the marking of Masters dissertations in a related area, to examine issues of double marking. This study suggests a wide disparity between individual markers and that the practice of using a third marker, when the disparity in an individual case falls outside a given range, does not necessarily improve reliability.