Low-stakes assessment is supposed to improve educational practice by providing feedback to different actors in educational systems. However, the process of assessment from design to the point of a final impact on student learning outcomes is complex and diverse. It is hard to identify reasons for substandard achievement on assessments, let alone the means that should be undertaken to improve the educational setting. Furthermore, it is difficult to show a causal link between educational reforms and change in test achievement over time. This paper examines the potential impact that low-stakes testing initiatives have in engineering higher education on educational improvement by examining two case studies. It discusses how the design of constructs and particular assessment conditions may foster or limit educational improvement and outlines conditions to ensure that low-stakes assessment designs achieve educational improvement.
