Student Performance Assessment

'Authentic assessment' is a term used largely in the USA where the intention is to design assessment which moves away from the standardized, multiple-choice type test towards approaches where the assessment task closely matches the desired performance and takes place in an authentic, or classroom, context. Student performance-based assessment, more commonly called student performance assessment, aims to model the real learning activities that we wish pupils to engage with, for example, written communication skills and problem-solving activities, so that assessment does not distort instruction.

The Standard Student Assessment Tasks outlined in the blueprint for the National Curriculum assessment programme in England and Wales (DES, 1988) are good examples of student performance assessment. Performance assessments demand that the assessment tasks themselves are real examples of the skill or learning goals, rather than proxies. They support good teaching by not requiring teachers to move away from concepts, higher order skills, in depth projects etc to prepare for the tests. The focus is more likely to be on thinking to produce an answer than on eliminating wrong answers as in multiple choice tests... insights about how to develop and evaluate such tasks come not from the psychometric literature... but from research on learning in subject matter fields' (Shepard, 1991). However, when such tasks are required to support psychometric principles such as reliability and standardization, in order to be used in accountability settings, they fall short since that is not the purpose for which they have been designed.

The issue for student performance assessment, as some see it, is how can tasks developed from, for example, diagnostic interviews be adapted for large scale administration and offer some level of confidence in comparability of results (which is necessary for accountability purposes). An alternative view is that we cannot force student performance assessment into a psychometric model and that what we need is a range of approaches: more formal testing on a psychometric model for monitoring and accountability purposes and teacher-based approaches on an educational assessment model for assessment to support learning. This still leaves us with the question of whether assessment for certification and selection purposes can be more broadly conceived (as for example, the GCSE) to offer both beneficial impact on teaching and sufficient reliability for public credibility.

Trickle down testing is characterized by standardization first and foremost and may be paper and pencil or student performance assessment; a good test is one that has high reliability, validity and efficiency and whose assessor remains a neutral observer; the results are largely used for accountability purposes (and in the UK we would add certification); the need for efficient scoring means that the 'fidelity' of results may be sacrificed; testing occurs at most once a year; the content represents a shallow sample from a broad domain; tests are timed; results are reported summatively, often with norm-referencing and involve considerable delay.

Although American writers refer to the need to change the culture of teachers if we are to move them away from a reliance on norms, and to change their belief that formal exams and tests are necessary in order to make students work hard, the situation in the UK is different. We have not had the same reliance on standardized tests as in the USA: our public exams sit firmly within the student performance assessment model while authentic assessment in the guise of RoA and pupil portfolios have been widely accepted as good assessment techniques.