Teacher bias in assessing five-year-olds could lead to flaws in the data used to prove how much progress their pupils make, a leading assessment expert says .

From 2002, children could be assessed at the end of reception year, instead of at the beginning, using one national baseline assessment.

Reception class teachers will assess what children have learned during the foundation stage and whether they have achieved early learning goals, such as counting to 10 and writing their name.

The data would show how successful early-years education has been and the progress made during key stage 1.

Birmingham education authority and the National Foundation for Educational Research have been contracted by the Qualifications and Curriculum Authority to devise the national scheme. Testing of the scheme will begin in November.

But Dr Peter Tymms, from Durham University, one of the bidders for the QCA contract, claimed the scheme being developed could depend too heavily on teacher assessment rather than testing. He said: “The scheme tends towards teacher observations, rather than objective assessment. It would involve teachers, who know pupils well, filling in rating sheets.

“But unless the scheme involves some kind of objective assessment, in the long run, we will not get good information.”

He said schools in affluent areas may rate their children at a lower level because practically all the children in the class are higher achievers. On the other hand, a school which has lower-ability children could rate some pupils highly because they achieve more than their classmates.

“Reception-class teachers want to be seen to have achieved early learning goals but Year 1 teachers want to show value added during KS 1 and therefore would benefit from lower scores as their baseline.”

He said where assessment was linked to a teacher’s own success there was a risk of flawed data.

Dr Tymms devised Performance Indicators for Primary Schools - which 4,000 schools use. It involves teachers and pupils going through pen-and-paper tests in reading, writing and use of number.

Teacher assessment is seen by some as a more acceptable way of gauging the progress of very young pupils.

Birmingham developed the Signposts assessment which is based on cumulative teacher observations and is now used by about 12 LEAs. Head of assessment at the council, David Bartlett, said Signposts had a very high level of consistency. Data from over nearly 10 years showed a very good correlation between assessment and results at key stage 1.

A spokeswoman for the QCA said that no decision had been made on what the scheme would consist of.