This paper reports a 3-year (1999-2001) Australian study of teacher judgement of student writing. It analyses teachers' talk to discover how they arrive at such judgements. It focuses on the processes teachers use as they read and appraise student writing, as distinct from judgements recorded as numerical or letter grades. It identifies and discusses a set of data-based indexes the teachers rely on to constitute their judgement. In so doing, the 'global' standard-setting of external assessment (judging the quality of student work against stated standards), and the 'local' of teacher judgement (based on the richness of what teachers bring to the task) are reconsidered. This study notes how teacher judgement of student coursework may be intertwined with and shaped both by officially authorized curriculum materials, syllabus documents, and assessment practices, and by other essentially private, local ways of knowing.