“…However, the need to ask questions that can be marked by computer leads to reduction in validity, as this tends to focus questions on procedural aspects (Broughton, Hernandez-Martinez and Robinson, 2017). The final answer input into a computer is not necessarily enough to establish partial credit (Rønning, 2017), a normal part of assessment in mathematics (Genemo, Miah & McAndrew, 2016), leading to a solution with a minor error potentially being marked as completely wrong (Greenhow, 2015). Some systems attempt to address this by breaking questions into parts, steps or sub-questions, prioritising procedural aspects and so reducing validity (Quinney, 2010).…”