Analysis of 600 international studies suggests that standards rise if assessment is used to identify children's learning needs. Maureen O'Connor reports.
Few education reforms or initiatives come with promises of quantifiable gains in achievement. But Professor Paul Black and Dr Dylan Wiliam of King's College, London, are quite specific about the improvement possible through the use of "formative" assessment or classroom evaluation in British schools.
Their conclusions are based on a year-long survey of 600 research studies across the world, involving 10,000 learners from nursery school to undergraduate levels. They conclude that teacher assessment which diagnoses pupils' difficulties and provides constructive feedback leads to significant learning gains.
The approach helps low attainers even more significantly than the rest. The effect is to reduce the spread of attainment in a group while at the same time raising performance overall, so tackling Britain's apparently intractable problem of a "tail" of underachievement and consequent alienation, truancy and crime.
Researchers estimate learning gains by comparing experimental groups with classes taught and tested in the usual way. The experiments surveyed by the King's College team have produced consistent and substantial learning gains - from the United States to Saudi Arabia, and in subjects from maths to PE.
A Portuguese study of 246 children with 25 teachers which introduced pupils to learning objectives and assessment criteria, then asked them to assess their learning on a daily basis, produced progress twice as rapid as that in a control group doing the same work.
In America more than 800 kindergarten children had their learning needs diagnosed regularly by teachers over eight weeks. Activities and tasks were then matched to the individual child. The experimental group achieved significantly higher scores in final tests of reading, mathematics and science. In addition, fewer children in the experimental group were diagnosed as having special educational needs and only one child in 71, as opposed to one in five in the "normal" class, was referred for special education.
The quality of the feedback offered to pupils appears to be crucial if formative assessment is to produce the maximum gains. But there seem to be no curriculum limits to improvement. A US study of 18 PE teaching projects involving reinforcement, feedback and corrective help produced very significant learning gains.
The most modest improvement achieved by the 600 research projects would raise the performance of the average pupil to the level of the top 35 per cent of students. In Britain, this level of improvement would push GCSE performance up by between one and two grades per subject.
The most substantial effect of formative assessment found in the survey, if replicated in the mathematics tests used for the Third International Maths and Science Study, would take England from the middle of the 41-country league table to the top five.
So exactly what are teachers doing in classrooms where formative assessment has such a remarkable effect? The survey indicates five factors seemingly crucial for success and a further five that hinder learning (see box).
In the UK, the researchers argue, the role of formative assessment by teachers has been officially recognised but almost all the resources and attention have been devoted to the end-of-key-stage tests. Little or no attention has been paid to teacher input into these final grades, and at GCSE the importance of teacher assessment has been deliberately reduced. This has reduced teachers' interest in their own assessment function.
"With hindsight, it can be seen that the failure to perceive the need for substantial support for formative assessment was a serious error," say Professor Black and Dr Wiliam.
They propose a four-point scheme for the development of formative assessment in British schools. Teachers, they say, will not take up new ideas - even those based on extensive research evidence - if they are presented simply as general principles.
The first step should therefore be to establish a small number of groups of schools in different environments to develop methods of formative assessment on a collaborative basis. The aim would be to see what works best in different contexts.
Schools should be given extra support for this development work and evaluators should be available to collect evidence of effectiveness.
The next stage should be one of gradual dissemination of good practice. Progress will only come as teachers find their own ways of incorporating formative assessment ideas into their own classrooms. This will take time.
Meanwhile, the damage done by short external tests - their tendency to dominate teaching and to encourage drilling to produce answers to out-of-context questions - and the interaction between formative and external testing should be studied. A reduction of the content of the curriculum when it is reviewed in 2000 might also be helpful, the researchers say.
Finally, they think that although further research is needed to clarify some aspects of formative assessment, this should not be used as an excuse to delay beginning it. The learning gains on offer are too substantial for that.
How BEST to raise STANDARDS
What works . . .
* regular classroom testing and the use of results to adjust teaching and learning rather than for competitive grading; * enhanced feedback between the teacher and the taught which may be oral or in the form of written comments on work; * the active involvement of all pupils; * careful attention to the motivation and self-esteem of pupils, encouraging them to believe that they can learn what is being taught; * time allowed for self-assessment by pupils, discussion in groups and dialogue between teacher and pupils.
... and what doesn't * tests which encourage rote and superficial learning, even when teachers claim they wish to develop understanding; * failure by teachers to discuss and review testing methods between themselves; * over-emphasis on the giving of marks and grades at the expense of useful advice to learners; * approaches which compare pupils in a way which persuades them that the purpose is competition rather than personal improvement, and which demotivate some pupils; * feedback, testing and record-keeping which serves a managerial function rather than a learning one.