Consider Jack. He's nearly five and he can't hold a pencil properly, but you should see him operate his PlayStation 2 controller. He refuses to blend his letter sounds, but astonishes you by recognising the word saxophone. (His dad has one.) He plays nicely with Rebecca for half-an-hour one day, but refuses to look at her the next. He's contrary, surprising, and constantly changing - just as he should be at four years old.
So how should he be assessed? If Jack were Australian, Danish, or Finnish, he wouldn't be under the school microsope. Even at the age of six, his achievements would be celebrated rather than measured. Photos of him busy in the sandpit might adorn the noticeboard. Or he might go home with a glorious annotated portfolio of his work. In high-achieving Finland, his progress would simply be tracked through frequent chats with his parents - and himself. But in England reception-class children have to be formally assessed, just as they are in liberal Wales, where baseline assessment is established for children entering reception.
For Jack, that entails the controversial foundation-stage profile. His teacher selects one of nine statements, little pen portraits of stages in six areas of learning, to describe the boy's progress towards the early-learning goals. Since 2004, the scores of Jack and his classmates have been published as national statistics. This infuriated members of the National Union of Teachers, who warned that the profile was a "key stage 0"
in disguise and would be used in value-added league tables.
The unions were not the only ones unhappy. In 2004 Ofsted described the profile as time-consuming and bureaucratic. What's more, it failed to tell teachers anything useful. Schools were assessing children twice to find out what they needed to know (see box, left).
"Assessment takes time and money so you need to be sure you have a good process that produces good information," says Peter Tymms who runs Pips, an assessment scheme based at Durham university which is used in 4,000 UK schools. The Government profile does not achieve this, he says. It lacks reliability. Two people doing it would not necessarily get the same answers. It does not continue past the foundation stage so cannot monitor Jack's long-term progress or, indeed, its own validity. If it suggested that Jack had a special need, there is no mechanism to check two or three years later to see if it was right.
Another flaw, says Professor Tymms, is that the under-fives vary tremendously, and the profile does not do justice to high-achievers.
Even more crucially, he asks, what is the profile for? Is it to help Jack's teachers get to know him, to plan his education, or to pinpoint a special need? If so, it must tell them more about the boy than they already know from teaching him. It won't, says Professor Tymms, because, unlike Pips, it is based purely on observation. "No one wants tests or exams for young children," he says, "but you do have to pose tasks, problems and questions in order to probe a child's deeper understanding.
"There are people who reject that out of hand. They are overlaying their own angst from their O and A-level days on young children. It is a mantra, a dividing line, that is holding us back in terms of giving good professional information to people working with young children."
Yet what Professor Tymms sees as a weakness, others view as a virtue. "The profile was welcomed because it wasn't just a test, a snapshot of a child's abilities," says Anne Nelson, director of Early Education, a leading voluntary organisation for practitioners. "Its greatest strength is that it is based on the long-term observation of children by people who know them.
After all, a robot could give a child a test."
Just because Jack might answer a question one day doesn't mean he will be able to do so the next, so, Anne Nelson argues, "teachers' observations over time and in different contexts are most useful. Some children feel more freedom out of doors to show what they can do and reveal abilities there that they might not do indoors on a similar activity."
She does have some worries. Staff have to guard against limiting children's progress by teaching to the early-learning goals, an easy temptation in these target-driven days. And she is anxious about the national statistics.
"Once I studied four children who each achieved the same overall score, but of course their profile was entirely different. You have to take great care how you use the statistics as the profile is subtle - it gives you a whole picture."
"The whole picture" is what Jan Comrie, a Cardiff headteacher, says she gets from the effective early learning programme (Eel), now employed in more than 70 local education authorities. It is based on observation, but its focus is rather different to that of the profile, although it did have an influence on the profile's development.
Created by Professor Christine Pascal of Worcester university, Eel is a self-evaluation tool for nurseries and pre-schools. "It's very structured,"
says Mrs Comrie. "You start by evaluating everything - your school documentation, your relationships with parents, your staff qualifications.
It's a huge undertaking.
"Then you start observing the children. First of all you 'track' them, seeing how often they visit the areas you've set up. Maybe no one is using the shaving foam on the activity table - so you ask yourself why."
At the heart of the programme is the "child involvement scale". This enables teachers to judge, not if Jack can count three building blocks, but how deeply engaged he is in playing with them. His facial expressions, posture, concentration, creativity and persistence are observed - and scored.
This score is not used to assess him, says Mrs Comrie, head of Grangemouth nursery school and an Eel trainer. Rather, all the results are analysed by the school so that it can see what it is doing well and not so well.
"Eel looks at attitudes," says Mrs Comrie. "It promotes looking at a child holistically and checking that what you are providing suits their individual needs and interests. It's about building from where the children are at, rather than a ticklist of whether they can recognise yellow, red and green."
The emphasis is on cognitive, personal and social development, she says, as is the case in Denmark and Finland. One Danish setting, visited by Ofsted in 2003, recorded whether six-year-olds were happy to socialise, their attitudes to learning, their behaviour, confidence and sense of security.
"Only once a child showed interest were notes made on their achievements in reading, writing and maths."
Very different from England, where the profile will note Jack's inability to use phonics to read simple words.
At least he isn't alone. Late last year newspapers reported that roughly two-thirds of reception pupils had not achieved the early learning goals in writing and reading. The Department for Education and Skills insists this isn't a problem since the goals are not targets, but aspirations. Still, perhaps Jack is lucky that he can't read the headlines yet.