Only 44 per cent achieved a level 5 or above in trials of the key stage 3 ICT tests in May. Last year, 80 per cent did so in English, 80 per cent in maths and 77 per cent in science.
Pupils take the tests on-screen, which are then marked by computer. This year 172,225 pupils, from 1,762 schools, took the test intended to become statutory in 2008. Results improved on last year, when only 36 per cent achieved a level 5 or above.
Teachers on the TES on-line staffroom have commented on the relatively poor results, and the fact that, for some the scores were well below their assessment of the pupils.
One said: "Our results...bore no relation to student ability whatsoever. We had 51 per cent level 5 plus, but TAs [teacher assessments] were 94 per cent and our GCSE results are regularly...85 per cent plus A* to C. So there's definitely something wrong with them!"
Another said: "[Pupils'] results came back with 60 per cent level five or better (compared to TA of 90-plus per cent). We get 86 per cent A* to C at GCSE, so the test is hardly reflective of that."
Another teacher reported "great" results in ICT, but warned that a feeling of achievement had soured after the deputy head had compared them with the "glowing" maths and science figures.
In a letter to schools, Sue Walton, of the National Assessment Agency, admitted that the 2006 ICT results were worse than teacher assessment levels. Further study was needed to find out why.
She said: "A key factor on pupil performance appears to be whether or not pupils have had adequate practice. Pupils who are well-prepared generally perform better than those who are not."
Some schools have reported serious teething problems with the technology.
Ms Walton's letter said that many schools had been "slow" to install software, and that nearly one in three secondaries had left practice until just before their pupils took the test.
National key stage 2 and 3 results are expected by September, with the exception of key stage 3 English, which will be released in the end of the autumn.