GCSEs play a crucial role in secondary education throughout England and Wales, and the process of marking them, which entails extensive human judgement, is a key determinant in the futures of many sixteen-year-olds. The aims of our study were to investigate the cognitive strategies used when marking GCSEs and to interpret them within the context of psychological theories of human judgement.

Two GCSE examinations were considered: an intermediate tier Mathematics paper, which used a 'points-based' marking scheme, and a foundation tier Business Studies paper, which used a 'levels-based' scheme. For each subject, a group of six experienced examiners marked four identical script samples each. The first three of these samples were marked silently. Whilst marking the fourth sample, the examiners were asked to 'think aloud' concurrently. Using a semi-structured interview schedule, the examiners were later questioned about their marking experiences retrospectively.

A qualitative analysis of the verbal protocol data enabled us to propose a tentative model of marking, which includes five distinct cognitive marking strategies: matching, scanning, evaluating, scrutinising, and no response. These strategies were broadly validated not only in the retrospective interviews with the participating examiners, but also by other senior mathematics and business studies examiners.