Alex Scharaschkin

I am Director of Research and Compliance at the education charity and exam board AQA. I run AQA’s Centre for Education Research and Practice (CERP), and am also Executive Secretary of the Association for Educational Assessment, Europe. At AQA my responsibilities include AQA’s programme of research and its dissemination; maintenance and management of awarding standards; technical regulatory advice; and oversight of regulatory compliance activity.

My research interests include the philosophy of measurement in education and the social sciences, the nature of validity arguments, and the application of quantitative and interdisciplinary approaches in assessment. I started my D.Phil part-time in October 2015 and my supervisor is Professor Jo-Anne Baird.

The working title of my D.Phil is ‘Constructs, valuations and counterfactuals: a mathematical theory of assessment’. This work draws on mathematics (category theory–the modern mathematical theory of structure), cognitive psychology (prototype theory of concepts), and philosophy (logics and semantics of counterfactual propositions). It aims to use these to ground a fairly general theory of assessment, going beyond more traditional approaches such as those based on representational measurement theory. It starts by examining outcome spaces for assessments, and the possible meaning of structures on outcome spaces that discriminate between, and ascribe value to, performances. It examines whether, in the same way that assessment procedures can be conceived of as generating outcome spaces, they also generate ‘statement spaces’ about the candidates being assessed, and the relationships between structures on outcome spaces and on statement spaces. Implications for conceptions of reliability, validity and comparability will be considered, as well as for aggregation of assessments of constructs from parts (such as items) to wholes (such as qualifications). The project will also explore how statistical approaches such as CTT (classical test theory) and IRT (item response theory) fit within such a framework. It will aim to gather empirical information on how assessors learn how to recognise good/bad examples of a construct, and examine whether such valuations can be modelled based on attributes of performances.

College: St. Anne’s