Problematising high-stakes assessment in statistics.

Abstract

Statistics emerged as a discipline to address pressing practical problems. In the UK, this has not been reflected in school statistics curricula, where students often work with small-scale invented data to develop mastery of statistical technique. Recent curriculum reforms set out to improve this situation; students are expected to work in class with a large authentic data set, and to demonstrate appropriate skills on high-stakes assessment. Here, we analyse all the first set of examination papers containing statistics for the new GCE qualification, and also questions using statistical graphs from the GCSE qualifications in summer 2017. We show that there is very little emphasis on statistical skills such as interpreting data and drawing conclusions, and a great deal of emphasis on technical skills. Contexts are (for the most part) banal. Several questions ask students to use inappropriate procedures. We believe systemic flaws have resulted in assessment which is not fit for purpose. We call for curriculum reform, and offer examples of how things might be done better both in curriculum and in assessment.