Interpretation: a choice of two models

Fischer Family Trust model
This uses value-added data to generate a rounded picture of children's and schools' performances. Working with LEAs, it takes data from that collected nationally on pupil performance at the different key stages, Sats, GCSEs and A-level. In addition to this it also gathers information on geographical location, gender, schools' ethnic profile and postcode.

It then produces two sets of analyses. One compares the progress of a group of children with the same cohorts nationally. A school can compare its performance with the national picture but can also view where it rates next to schools in similar socio-economic backgrounds.

"It's about trying to provide a balance between comparing schools but also trying not to lose sight of the fact that we want to aim for all schools making progress, irrespective of background," says Mike Treadaway, leader of the trust's research. "You need to have both in there to know the differences."

The other analysis sets estimates for what a group of children might achieve in three or four years' time, based on their progress thus far and on what others have achieved previously. They are informed expectations not targets, Treadaway is keen to stress. "We don't think they should be a target because there will be some children who buck the trend. Rather, this approach tries to encourage positive thinking about a child's potential."

The trust's data goes out electronically to the LEAs with which the trust works, and some are already starting to feed elements of it into their own information management systems. It's not a straightforward process, however, and compatibility with Fischer and other models is a key issue for companies.

New DfES/Ofsted data model
One of the challenges of comparing schools lies in the different approaches to data that they and LEAs use. All national data sets are built up from the same data produced by exam boards: it's the way they are analysed that varies. The way data is mapped can tell different stories, which can make comparing like with like a tricky proposition.

Part of the difficulty has been that the DfES and Ofsted have never used the same data, though that's about to change. They have agreed a joint data set which both Ofsted inspectors and Sips will use. Called RAISEOnline, it will replace Pat (Pupil Achievement Tracker) and Panda (Performance and Assessment Reports) from 2006.

The new unified system will provide schools with a tool to review performance data in greater depth and compare it against national data. It will also include the existing target-setting and question-level analysis elements of Pat.

A DfES spokesman says: "The department is committed to developing a more sophisticated value-added methodology for all departmental purposes. Ofsted will use the same model in its Pandas. An initial model was piloted last year and we are developing the model further."