The Nature and Value of Formative...
DRAFT �� King���s College London 1 2/6/04 The Nature and Value of Formative Assessment for Learning Paul Black with the King���s College London Assessment for Learning Group (Christine Harrison, Clare Lee1, Bethan Marshall, Dylan Wiliam2) Abstract The this paper has two foci. The first is to present an account of how we developed formative assessment practices with a group of 36 teachers. This is then complemented by a reflection on the productive and positive experience of these teachers, in the light of learning principles, of changes in the roles of teachers and pupils in the task of learning, and of effects on the self-esteem and motivation of pupils. Attention then shifts to the second focus, which is on the ways in which these teachers struggled with the interface between formative assessment and summative testing. The conclusion is that the potential of enhanced classroom assessment to raise standards may never be fully realised unless the regimes of assessment for the purposes of accountability and certification of pupils are reformed. The starting point In 1998 we published a review summarising the results from over 250 articles by researchers from several countries. Our main purpose was to survey the evidence about the effects of improving formative assessment on pupils��� performance (Black & Wiliam, 1998a). This review showed that there was a strong body of evidence to support a claim that formative assessment practices can raise standards. At the same time, the published evidence also showed that such practices were only weakly developed in most classrooms. Three main problems were identified. The first was that the assessment methods that teachers use are not effective in promoting good learning.The second was that marking and grading practices tend to emphasise competition rather than personal improvement. The third problem was that assessment feedback often has a negative impact, particularly on pupils with low attainments who are led to believe that they lack ���ability��� and are not able to learn. However, whilst these conclusions were a spur to action, they could not provide recipes for improvement. The reported surveys and experiments lacked the detail that would enable teachers to implement the practices in classrooms. It seemed to us then that teachers would need a variety of living examples of implementation, by teachers with whom they could identify and from whom they could both derive conviction and confidence that they could do better. 1 Now assessment adviser with the Warwickshire LEA 2 Now with the Educational Testing Service, Princeton, NJ, USA
DRAFT �� King���s College London 2 2/6/04 To publicise the results we also published a 20-page booklet Inside the Black Box: Raising standards through classroom assessment. (Black & Wiliam, 1998b). Over 40000 copies have been sold and the work is widely quoted. Thus it was clear that teachers were interested in what we had to say. So we decided to start on the exploratory development work that was called for. To do this, we needed to collaborate with a group of teachers willing to take on the risks and extra work involved, and to secure support from their schools and their Local Education (district) Authorities (known in the UK as LEAs). The funding for the project was provided through the generosity of the Nuffield Foundation. We were fortunate to find, in the Medway and Oxfordshire LEAs, advisory staff who understood the issues and who were willing to work with us. Each authority selected three secondary schools, spanning a range of catchment backgrounds they included one boys��� and one girls��� school, the other four being mixed. Each school selected two science and two mathematics teachers. We discussed the plans with the head of each school, and then called the first meeting of the 24 teachers in January 1999 . In the subsequent second year of the project we added 12 teachers of English, two from each of same schools. The project became known as the King���s -Medway -Oxfordshire Formative Assessment Project (KMOFAP). The findings from this work have been published both as a short booklet for teachers (Black et al. 2002) and as a book for teachers (Black et al. 2003) The ways in which the partners involved worked together has been written up elsewhere (Black & Wiliam 2003). For the present purpose, it is the outcomes that are important. The findings presented here are based on the observations and records of visits to classrooms by the King���s team, records of meetings of the whole group of teachers, interviews with and writing by the teachers themselves, and a few discussions with pupil groups. Throughout the development of the project, members of the King���s team have responded to numerous invitations to talk to other groups of teachers and advisers: over three years they have made over 100 such contributions. These have ranged across all subjects, and across both primary and secondary phases. In addition, there has been sustained work with some primary schools. All of this makes us confident that our general findings will be of value to all, although some important details may differ between different age groups and different subjects. Some of the results of our work have been incorporated into the Key Stage 3 initiative and into the primary strategy, whilst researchers and teachers in the project have been appointed consultants to national programmes in Scotland and in Jersey to develop assessment for learning. The concept In our short publication for teachers (Black et al. 2002) we included the following extended definition of our idea of formative assessment : Assessment for learning is any assessment for which the first priority in its design and practice is to serve the purpose of promoting pupils��� learning. It thus differs from assessment designed primarily to serve the purposes of accountability, or of ranking, or of certifying competence.
DRAFT �� King���s College London 3 2/6/04 An assessment activity can help learning if it provides information to be used as feedback, by teachers, and by their pupils in assessing themselves and each other, to modify the teaching and learning activities in which they are engaged. Such assessment becomes ���formative assessment��� when the evidence is actually used to adapt the teaching work to meet learning needs. The detail and precision here have been found essential, because many teachers and researchers seem to have misunderstood the term. One example of misunderstanding is that it is about any assessment conducted by teachers, and in particular that giving a test every week and telling the pupils the marks constitutes formative assessment. It does not. Unless some learning action follows from the outcomes, such practice is merely frequent summative assessment. Another example is the belief that it includes portfolio assessment where that is developed with the aim of replacing or supplementing the information which externally imposed tests are designed to provide. Again, there is no formative assessment in such practice except insofar as there is active feedback to change and improve pupils��� work as the portfolios are built up. In general, any test or assessment at the end of a piece of learning is too late for formative purposes, precisely because it is at the end, so there is no opportunity to use its results for feedback to improve performance of the pupils involved. The learning gains Although our review of the international research literature showed that enhanced formative assessment produces gains in pupil achievement, we were clear that it was important to have some indication of the kinds of gains that could be achieved in normal classrooms, and over an extended period of time. The criteria for ���normal��� were that the content of the teachers��� lessons would be that of their normal plans, and that the only measures of pupils��� performance would be those derived from their normal testing procedures. Since each teacher in the project was free to decide the class with which they would work on these ideas, we discussed with each teacher individually how to set up a ���mini-experiment��� for each teacher using data available within the school. Each teacher decided what was to be the ���output��� measure for their class, using, as appropriate, data from national examinations and tests and/or results of the school���s own tests. For each project class, the teacher identified a control class. In some cases this was a parallel class taught by the same teacher in previous years (and in one case in the same year). In other cases, we used a parallel class taught by a different teacher and, failing that, a non-parallel class taught by the same or different teacher. Where the project and the control classes were not strictly parallel, we controlled for possible differences in ability by the use of ���input��� measures, such as scores on a standard Cognitive Abilities Test, or school test scores from the previous year. The data covered changes over one whole school year ( for a full report see Wiliam et al. 2004). To enable us to aggregate the results across the teachers, we used the standardised effect size. For the nineteen teachers on whom we had reliable data, the average effect size was around 0.3. Such improvement, produced across a school, would raise a school in the lower quartile of the UK national performance tables to well above average. It was clear, therefore, that, far from having to choose between teaching well and getting good test results, teachers can actually improve their pupils��� results by working with the ideas that developed in our