Until this year, teachers were able to select test materials from various catalogues, the suggestion being that their natural inclination was to choose tests which they knew would suit their classes.

However, in preparation for the move to a national survey of assessment - a switch away from the current system of school-by-school collated results - over the last year teachers have had to use randomly selected tests from a national electronic bank.

The latest annual survey of 5-14 attainment - the last to be published before its planned replacement in 2005 by the Scottish Survey of Achievement - shows that overall, attainment levels have increased slightly in reading, writing and mathematics, although there is a dip in the percentage of pupils attaining Level A in reading at the P2-P3 stage.

At P2, attainment dropped from 51.9 per cent (2002-03) to 45.9 per cent (2003-04), and at P3 from 87.5 per cent to 85.5 per cent.

A spokeswoman for the Scottish Executive said: “The reason for the dip in P2 and P3 in reading is the switch over to new electronic tests. The feedback from schools and local authorities is that the new Level A tests differ from the old ones.”

Judith Gillespie, development manager of the Scottish Parent Teacher Council, who participated in the assessment action group which worked on changes to the way that attainment is assessed, said: “Because there is a new system of providing the materials, you would expect some kind of adjustment in achievement levels. In fact, you would have expected a bigger dip in every area.

“The fact that there is only a dip in Level A is really quite remarkable.

It probably suggests that in the past, teachers at this stage were selecting things that were not challenging enough to evaluate where the kids had got to.

“There used to be a disparity between AAP (Assessment of Achievement Programme) and 5-14. It showed that schools were not doing as well under the AAP as under the 5-14 results.

“Now they are trying to bring things into line with one another by using the AAP questions in a kind of random generated way to produce tests for 5-14,” she added.

Peter Peacock, the Education Minister, said: “We don’t expect our youngest pupils to reach the first assessment level - level A - until primary 3. But we are still seeing significant numbers of pupils reaching that level a year earlier. This is despite the introduction of a tougher test at that level.

“The steady progress of results across the board in the later years, particularly S2 maths, is also an encouraging step forward as we strive to equip our children to reach even higher in future.”

The new electronic tests were introduced over 2003-04 in preparation for the new Scottish Survey of Achievement which will in turn build on the AAP survey, and will from May 2005 test pupils in P3, P5, P7 and S2 in English language, mathematics, science and social subjects.

TRENDS BEHIND THOSE FIGURES

The Scottish 5-14 results up to June 2004 using the new tests showed:

* attainment in English reading at Level A has decreased at P2 from 51.9 per cent (2002-03) to 45.9 per cent (2003-04); and at P3 from 87.5 per cent to 85.5 per cent over the same period:

* by S2, the percentage of pupils attaining Level E in reading had risen from 60.7 per cent in 2002-03 to 64.5 per cent;

* in English writing there was a slight dip at P2 in Level A - 43 per cent (2002-03) to 42.1 per cent (2003-04), and at P3 - 85.6 per cent to 85.4 per cent; by S2, the percentage attaining Level E had risen from 51.4 per cent to 52.1 per cent;

* levels of attainment in mathematics have continued to increase at all stages, most notably at S2, where the percentage attaining Level E in 2002-03 was 54.6 per cent and in 2003-04 was 59.5 per cent;

* at all stages, fewer pupils attained target levels in English writing than in mathematics and reading. For stages P2 to P4, attainment in mathematics is highest and from P6 to S2 reading has the highest attainment;

* the percentage of pupils attaining target levels was higher for girls than boys for all subjects and stages.