Estimating sub-domain scores within national assessments. A case study using the Key Stage 2 statutory mathematics test in England

Each year, hundreds of thousands of 10–11-year-olds in England sit the Statutory Assessment Tests (SATs). Alongside overall mathematics scores, schools receive results for eight mathematics sub-domains intended to highlight specific areas of strength and weakness. However, little empirical evidence exists regarding the reliability or practical value of these sub-domain scores. This paper investigates this issue using data from England’s Key Stage 2 mathematics assessment. Drawing on descriptive analyses, exploratory factor analysis and multidimensional item response theory, we demonstrate that the school-level sub-domain scores derived from the test lack the reliability, utility and precision necessary for meaningful diagnostic use. We argue that the current practice of reporting Key Stage 2 mathematics sub-domain scores to schools should either be discontinued or supported by a redesigned assessment that can validly fulfil this diagnostic purpose.