Baseline assessment at the start of infant schooling can be like attempting to grade a beautiful sunrise on a scale of one to five and expecting that figure to convey its full glory to others. It can't, of course. Nor is there likely to be unanimity on what constitutes a grade five or a grade one and even if there was, that grade would not necessarily tell you what kind of day to expect thereafter.
Before the Government embarks on a national scheme for assessing five-year-olds, then, it needs to establish at least two things. First, that it can be done consistently and reliably; that any competent person making such an assessment would reach a similar conclusion. And second, that the information provided will serve some useful purpose. Assessing five-year-olds by their shoe size would be reliable in the sense of objective and consistent. But it is unlikely to yield much valid information about children's different achievements or propensity for learning. For any assessment to be valid it is necessary first to be sure what the purpose of it is and then that the measurement taken is relevant to that purpose.
In the case of baseline testing there are at least two purposes in various minds. One is to improve the accountability of primary schools; to provide a check on real progress by charting children's different starting points. Key stage one assessments cannot provide fair measures of pupil progress, school or local authority performance without some indicator of prior attainment. Everyone with a part to play in raising achievement needs comparable information on the value added at each stage rather than just raw results. This is recognised already by many authorities since over half have already embarked on baseline assessment. Some even plan to use it to improve targeting of extra funding.
In contrast to administrative uses for such information, however, are the desires of early years teachers for more insight into what their pupils know, are already capable of and need for the future. And since it is they who will have to cope with the extra work of baseline assessment, what they want out of it matters. A diagnostic assessment answering teachers' qualitative questions is likely to be rather different to one designed primarily to provide quantitative data. But can a single instrument be made to encompass both purposes?
The School Curriculum and Assessment Authority was asked by Gillian Shephard to look into all this. It will be consulting formally in the autumn on a national framework designed to create greater uniformity without stifling local initiative. Drawing on its proposals for the "desirable outcomes for children's learning on entry to compulsory education", SCAA has come up with three possible models (page 1).
The simplest provides 20 short statements relating to the child's reading, writing and mathematical development, which the teacher is asked to tick "yes" or "no". This has the potential to provide a clear and consistent numerical score for each child, while making few demands on the teacher. On the other hand, it offers little insight into the child's needs or wider attainment. Its narrow curricular focus ignores speech let alone social skills and, given the tendency of assessment to influence the curriculum, its backwash effects could prove counterproductive. It is as though when asked to rate that glorious sunrise you simply judged it by whether it gave enough light to read a newspaper.
The model at the other extreme provides six headings based on SCAA's "desirable outcomes": personal and social development; language and literacy; mathematics; knowledge and understanding of the world; physical development; and creative development. These areas of learning are intended to focus teachers' comments on the child's achievements over the first half term. This provides no numerical scores of any kind. It systematises observations to some extent but hardly informs them and it depends upon the skill and insight of the teacher whether it provides any helpful information .
The rapprochement suggested between these two extremes is to have both in the national framework. This would simply mean two instruments rather than one, with the numerical tail wagging the descriptive dog. Teachers would rightly feel the formative part was there as a palliative, like teacher assessments with national curriculum tests.
A third and more promising approach relies upon matching children to detailed grade descriptions devised around the "desirable outcomes" and the early key stage one levels. These helpfully prompt observations and detailed qualitative judgments by the teacher about each child which may well help to ensure every child's education gets off to the best possible start. These gradings can also be ascribed numerical values. But SCAA has still to show these to be better indicators of future attainment than shoe sizes.