Throughout the past two decades assessment has operated on two fronts. First has been the continuing interest in large-scale, standardised testing, which affords governments and countries data for accountability and reporting purposes. Second has been the increasing interest in assessment within a learning culture (Shepard, 2000). Broadly speaking, this has concentrated on formative assessment for improving learning and has generated a proliferation of phrases seeking to highlight vital connections between assessment and learning (for example, 'assessment for/as learning'). Each of these fronts can be understood as giving priority to particular assessment activities and contexts. In the case of standardised testing, usually undertaken to generate data for systems' purposes, the context is necessarily controlled, with variables such as time and place fixed and regulated.