This paper presents findings from the Australian Research Council funded ‘Disciplinarity, Knowledge and Schooling’ project (DISKS) which investigates knowledge-building practices in Australian secondary schools and gave rise to the ground-breaking notions of ‘semantic waves’ (Maton, 2013) and ‘power pedagogy’ (Martin, 2013). In this paper, we investigate student writing in senior secondary school Ancient History. We focus on how students use evidence in their responses to different types of exam questions. Our research question focuses on the extent to which key features of responses to short answer questions appear in extended responses and vice versa. This focus arose through findings that teachers in our study tended to view short answer questions as a ‘mini’ version of extended responses and prepared students accordingly. The similarities and differences are important to identify as extended responses make a significant contribution to the overall exam grade.

To better understand the use of evidence in responses to different types of exam questions, the study draws on the dimension of Semantics in Legitimation Code Theory (Maton, 2013). We use the newly developed wording and clausing tools (Doran & Maton, 2018, forthcoming) to analyse the relative strength of context dependence in responses to Year 12 exam questions. Context dependence is particularly relevant to how students use evidence, as it involves relating the concrete particulars of specific historical artefacts, events, and the behaviours of historical figures to more abstract concepts in the discipline of history that are not bound to one historical setting. Our analysis tracks relative shifts in context dependence in student texts to generate semantic profiles of their exam responses.

Findings show that although teachers may use the writing of short answer questions as preparation towards the high-stakes extended writing tasks, short answer responses are not ‘minature’ versions of extended responses. We argue that the differences are teachable and propose the use of model texts to make these features visible to students. Beyond the timeframe of secondary school education, learning to use evidence, particularly for the development of arguments, may provide a robust foundation for tertiary level writing tasks where students need to control degrees of context dependence.