The hardest decisions teachers have to make during their assessment of children at Key Stages 1 and 2 concern those who fall at or near the borderline between levels. The well-illustrated Exemplifications of Standards in the three core subjects produced by the School Curriculum and Assessment Authority are not necessarily any help in this grey area.

This is partly because they have to illustrate clearly what does constitute work at a given level. But also because, as they emphasise, teacher assessment is not expected to be based on written work alone. The final judgment should be made on the basis of how a child performs across a range of contexts, taking strengths and weaknesses into account and using knowledge of a child’s whole performance in class, not just completed pieces of written work.

Advisers and head teachers are now building up some good practice in this area which they are willing to share. Dorothy Kavanagh, this year’s president of the Association of Assessment Inspectors and Advisers, says teachers need to become more confident of their judgment.

“The evidence for a given level does not have to come from what is written down. It can be in the teacher’s own head, a question of professional observation in the classroom and assessment,” she says.

She suggests that schools should keep a portfolio of children’s work from across the whole school so that different classes can be compared. Making time at a staff meeting to discuss a child’s work in a given subject is a useful exercise, she thinks. As little as 15 minutes might be enough to allow a teacher to explain his or her judgment and justify to colleagues why a certain level has been decided on.

This is particularly useful for children on the borderline where the final decision may depend as much on a teacher’s knowledge of a child in the classroom as on written work. “Written work may give an impression of something higher or lower, so it is good practice for the class teacher to justify a different decision to colleagues. It encourages teachers to think around a judgment of what a child is doing. This is especially important in the early years when so much is dependent on process and so little on written work. ”

The majority of children fall fairly easily into levels, Ms Kavanagh says, so it is particularly important to discuss the borderline cases. Making these decisions is not a science, she says. It is all about professional judgment, but teachers must be able to justify their decisions to others.

Stephen Munby of the quality development service in Oldham makes a similar point. “We emphasise ‘typical’ performance and best fit with the level descriptions.” But it is important for schools and departments to standardise their judgments through the use of portfolios of work, he says.

Accurate judgments, Mr Munby thinks, come from teachers developing a thorough knowledge of the meanings of the various levels and of the progress of their individual pupils. The first can be gained by a combination of: u reference to the levels when planning work,u designing an assessment scheme to accompany work units,u occasional reference to levels when judging achievement in a project or unit of work, u discussing and moderating pupils’ work with colleagues, and u examining and discussing the SCAA exemplification materials.

Knowledge of a pupil’s individual progress can be built up through marking work, observing pupils, asking questions to test knowledge and understanding, through formal tests and through record-keeping.

Sue Matthew, head of an Oxford first school, also emphasises co-operation between teachers, especially over the most difficult, borderline decisions. She and her staff have found it helpful to divide children within each level into “middle, top and bottom” groups. This is also useful when discussing children’s progress with parents.

The school sets aside part of in-service training time to look at examples of work and come to a consensus as a staff about where the borderline lies. “It is hard work,” Mrs Matthew says. “It is time-consuming. Each subject is different and each child is different.”

She and her colleagues have found it useful to look at a range of work in a subject and a particular attainment target and also to look at the range of work produced by an individual child, particularly if it seems to stretch across the boundary between levels.

“This approach has made the borderline decisions easier. We work as a staff and we now find that we usually agree.”