In the steelworks of Sheffield, 40 years ago, there were two ways of measuring the temperature of a hot metal ingot. One was to apply a pyrometer to it and read the scale. The other was to call up the foreman, who would spit on the ingot, listen to the sizzle and then announce the temperature to the nearest 10 degrees Celsius. You could use the second method, of course, only if you had total confidence in the professionalism and experience of the spitter.
The same question - to what extent assessment should be balanced between measurement and judgment - has hung over the development of the national curriculum from the start. And we are not simply talking about the distinction between school-based assessments and tests set from outside.
Statutory teacher assessment can be done in two distinct ways. You either dismantle the curriculum into lots of little parcels of knowledge (for which read "statements of attainment") and then tick them off as the child masters them, or you look through the child's work and your own records, and then say "I've taught Jane for two years and in my judgment her work fits the description for level 3."
The distinction between these two models of teacher assessment is a key difference between the existing and revised versions of the national curriculum. The difficulty is that many teachers are going to take a lot of convincing. Peter Lacey, a professional officer for mathematics at the School Curriculum and Assessment Authority, for example, believes that "moving to the notion that the end-of-key-stage assessment is a rounded judgment will be difficult for some".
Talking to teachers supports this. Over the past few years many of them have felt pushed into mechanistic teacher assessment systems that needed lots of evenings and weekends spent on paperwork. To tell them now that they need only spit on the red-hot ingot is to invite doubt and uncertainty, especially among teachers who have come into the profession since 1988. As one Warwickshire head put it: "Assessment co-ordinators have hammered and hammered for some years now on the need for evidence of level in every statement of attainment all the way through the key stage."
Angela Walsh, another professional officer for maths at SCAA, accepts that a culture shock is on its way, and is at pains to give reassurance, especially to schools which are working with assessment models they like and trust. "The thing is to move slowly away from whatever systems have been set up - evolution, not revolution."
One central piece of advice, she believes, is that teachers are no longer going to be looking at levels for individual pieces of work.
"The big message is getting away from using levels within the key stage. It's not now a matter of monitoring levels, but of monitoring progress."
The new model in action in a school, therefore, might go something like this: The teacher guides and monitors the child's progress through the key stage, referring to the programmes of study and the school's planned curriculum with its assessment opportunities.
Clearly, for her own professional purposes, she has to assess the pupils as she goes. Thus she may keep records and possibly a small collection of work in order to have evidence of standards. She may even take a school (or departmental) portfolio to other schools to make comparisons.
There is, though, no statutory requirement to do any of these things. (Still to come is joint SCAAOFSTED guidance to help schools monitor the reliability of their statutory teacher assessments.) Most importantly, though, there will definitely be no need to tick off statements of attainment, because they no longer exist. SCAA officers hope to avoid an initial tendency by some - perhaps in the belief that they are being helpful to teachers - to take the level descriptions to pieces, and set them out as assessable fragments, thus effectively reinventing statements of attainment.
Hence SCAA's heavy and constant emphasis on the holistic nature of the level descriptions. As SCAA spokesperson Madeleine Moore explains it: "With level descriptions you are looking for a rounded judgment of the pupil's performance. We're really very anxious to get people away from tick boxes."
Statutory teacher assessment with the use of levels comes only at the end of the key stage, when the teacher makes a judgment about which level description best fits each child's performance. (This concept of "best fit" figures strongly in SCAA's explanations of level descriptions.) To help in this, she will undoubtedly look back at whatever records and evidence she or the child has kept, and she may also refresh her understanding of "levelness" by looking in a school or departmental portfolio containing selected children's work. She may also reach on the shelf for the collection of examples of pupils' work - "exemplification material" - which SCAA will by then have produced. In the end, though, she will be expected to use her professional judgment in a way that teachers have not so far been free to do. Thus, having judged that a child's work is at, say, level 5, she will look at the descriptions for levels 4, 5 and 6, and confirm her judgment on the basis of "best fit".
SCAA does not expect teacher assessment to be a bureaucratic or stressful business, and the implication is that teachers will develop what Peter Lacey called "an internalised fix on levelness" which will make the job much less time-consuming than this description implies.
Exemplification material - samples of assessed pupils' work at the first three key stages, in the three core subjects, with supportive text explaining how levels were arrived at - will be a key reference point, especially during the first years of the use of level descriptions. Very importantly, there will be a section - covering levels 1, 2 and 3 at key stage 3 - aimed specifically at supporting teachers of special needs pupils.
There is reassurance for teachers that when the material appears in June 1995 it will have been produced over a nine-month period by teams of experienced classroom teachers, working in a wide range of schools to the post-Dearing national curriculum. Clearly, they are gathering valuable experience in the revised approach to teacher assessment.
These teachers are generally approving of the opportunities offered by level descriptions. Jerry Swain, head of English at Prince Henry's School, Otley, made a point which I heard from other secondary heads of department.
"Assessment by level description is more in tune with the kind of judgment we have become really skilled at in running 16-plus and GCSE. Under the old Orders, teaching was becoming assessment-driven because teachers felt they had to exemplify every single aspect of the statement of attainment, and it became a really artificial approach to English teaching."
Much the same feeling is evident at the other end of the age range, at Christchurch Infants - a multi-ethnic school in Ilford - where teacher Gaynor Sparrow is helping to produce examples of work in key stage 1 English.
Her message to classroom colleagues is generally a positive one, and she sees improved logic in the revised teacher assessment requirements, particularly if, as is the practice at Christchurch, work is adequately planned with assessment in mind.
"The programmes of study, if they're used efficiently, will produce a range of work so that you can use the 'best fit' approach. It's much more fair. It creates a more rounded picture of the child's performance - you can say that a child's work, perhaps, demonstrates this, has elements of that, but not so much of something else. Therefore she's not really at level X but she's fully functional at level Y. We've done it, and it works!" At Christchurch, for example, the work in the portfolio represents a range, chosen from a "control group" of six pupils within each class. The choice of pupils for this group, explained Gaynor Sparrow, is carefully done so that their work will illuminate the borderlines between levels. "It's no good picking somebody who is in the middle of a level. You need to pick children who hover between one level and another."
The portfolio is then a tool used to help the moderation of classwork both by teachers and support staff, first within the year group and then across the school - "So that teachers not currently working in Year 2 are trained in the process."
"Moderation," explained Gaynor Sparrow, "is not a one-off process. Instead of spending two hours looking at umpteen pieces of work we spend half-an-hour at a time on a limited number." Also involved in producing exemplification work is Chiquita Henson, deputy head of Deer Park School, Cirencester, Gloucestershire. She, and the other teachers involved in the work "have made every effort to produce a realistic and useful document. We've been careful to select work that shows coverage, and we have not engineered special situations for that to take place."
It was left to Deer Park's head, David Crossley, who is deeply committed to an open approach to teacher assessment, with full parent and pupil involvement, to throw in a note of caution. "Statements of attainment were very specific, and they were responsible for shifting teachers away from making vague judgments like 'good effort'. The level descriptions are not so specific, and the danger is that we won't talk about the the detail of attainment as much as we once did."
Some other teachers have their own reservations. More than one infant head, for example, was clearly doubtful about the prospect of giving up a detailed teacher assessment structure built up and refined over the years.
The message to SCAA is that they will really have to help teachers to relearn how to fly. As Jerry Swain of Prince Henry's put it, "Teachers need training and reassurance in the exercise of their own judgments."
The exemplification materials will presumably provide some of this reassurance - especially if, as the participating teachers have been requesting, the whole package is sent to all schools, so that teachers can see progression from key stage 1 to the end of key stage 3.