In Idea 10 of Graham Gibbs’ blog, he discusses the idea that “Students’ marks are often determined as much by the way assessment is configured as by how much students have learnt“. In exam paper setting season, it is good practice to review the performance of last year’s students on the previous exam. If students did more poorly than expected, we often think about how hard the students worked or even the attendance at lectures. If students did better than expected, we may look at the interventions or changes that were made to the teaching. However, Gibbs argues that it is the assessment regime itself that may be the most important factor in determining how the class as a whole performs. When considering the design of this year’s exam, it is worth considering how it is made up as well as its content.
For an exam, students have to study all topics in some depth unless they are willing to take risks on likely topics. Their study time is likely to be shared across all topics, so that each topic will get less time than it would for a piece of coursework. Exam marks reflect this lower amount of time spent on each topic. To reduce any bias in the difficulty, it is often a good idea to set exams without too wide a choice of questions. An exam which randomly samples the entire curriculum and is purposefully designed to assess core and advanced understanding will generally give a reasonable indication of what students have learnt across the whole curriculum. However, exam marks are usually lower than marks for coursework for precisely this reason.
Graham Gibbs is one of the UK’s most well known advocates of improving university learning and teaching. With his colleagues (Habeshaw & Habeshaw), he was responsible for the very popular ’53 Interesting Ideas’ books series that have supported countless teachers (new and experienced) navigate their way through tricky classroom and curriculum dilemmas. Gibbs’ new blog 53 Powerful Ideas all Teachers Should Know About offers a research summary of key issues.