Professor David Jesson of York university issued the warning after conducting an analysis of information and communication technology results which, he said, suggested that pupils had been marked too severely.
Last week, The TES revealed that all 8,000 pupils taking AQA's applied GCSE in ICT had their results upgraded after only 0.5 per cent of candidates got C or better on one paper.
But Professor Jesson said that even after these changes had been made results were still very low.
Only 33 per cent of more than 40,000 pupils who took applied ICT achieved a C or better, compared with a national average for traditional GCSEs of 59 per cent.
Any attempt to suggest that this was because the cohort who took the exam were of relatively low ability was contradicted when the figures were analysed in detail, said Professor Jesson.
He analysed the key stage 2 results of pupils who had taken the ICT exam and found that they were slightly higher than average.
He then divided the pupils into five groups, according to their performance in the KS2 tests in 1999.
Some 95 per cent of those in the highest-performing group at KS2 gained a C or better in English GCSE. For maths GCSE, the figure was the same; for science it was 93 per cent.
In applied ICT, it was only 60 per cent. The other seven applied GCSEs last year - offered in subjects including business, science and manufacturing - all had lower proportions of pupils gaining C or better than the average for academic GCSEs.
Professor Jesson said that the ICT figures contradicted certain ministerial promises last year that no pupils would be disadvantaged by taking the new courses, which were examined for the first time last year.
"There is clearly a very significant case to answer and until it is, schools are likely to abandon these courses as offering little of value to their pupils or to themselves," he said.
A Qualifications and Curriculum Authority spokesman said: "As with all new qualifications, QCA is keeping every aspect of the applied ICT GCSE under review."
The QCA would be carrying out its own analysis of the results, similar to that conducted by Professor Jesson.