Reading recovery, a core part of the Government's multi-million-pound literacy programme, is being dropped in Queensland, Australia, because of fears that the gains children make do not last.
Britain has already given pound;5 million towards a three-year pilot of Every Child a Reader, a scheme that uses Reading Recovery with other less intensive programmes.
But from September, the Government has earmarked pound;144m for reading plus a maths programme, Every Child Counts, to help 30,000 children a year by 2011.
But in the Australian state, funding is being gradually withdrawn because officials there say improvements made by pupils were temporary. The $A20m (pound;9m) allocated for the programme will be spent on teacher training instead.
Literacy expert Kevin Wheldall, director of the Macquarie University Special Education Centre, said: "I am truly amazed that the UK Government is investing further in Reading Recovery when others elsewhere are bailing out.
"The logic of employing Reading Recovery as a solution for pupils who have struggled to learn to read following phonics instruction is almost wilfully perverse - a triumph of hope over experience. These are precisely the children for whom Reading Recovery works least well. It is only moderately effective for those low-progress readers whose problems are relatively mild - having missed some schooling, for example - and is ineffective for those with severe phonological processing problems."
He and colleagues have assessed research carried out since 1992. Findings on the scheme's long-term effects were equivocal. One randomised control trial from 1995 concluded that it was only effective for one in three pupils.
But Jean Gross, director of Every Child a Reader, said there was ample evidence it work long term.
"Three early studies in the United States and New Zealand found the gains did wash out, but a large number of studies have found a long-term impact," she said. "It does wash out if children don't go back into appropriate literacy learning in class. We don't get that in England."
Ms Gross believes Reading Recovery has changed radically since it started to accommodate research on phonics and "blending". It is not a set programme, she says, but a way to train people in a highly skilled practice.
"Children might not read because they have very restricted language and can't construct a sentence in their head," she said. "Some children don't know where to start reading on a page - at the bottom or the top. Others can't perceive the difference between 'a' and 'o'.
"Every child needs strong synthetic phonics teaching, but for the most complex needs, children will need other things around that.
Heather Rockhold, head of Lauriston Primary School in Hackney, east London, has used Reading Recovery since 1993. "I know it works," she said. "Out of the children who left school in July, 10 had Reading Recovery. Three left before Year 6. Out of the seven left, one got level 4 and six got level 5.
"A synthetic phonics programme is essential, but there will be children who don't progress as quickly as they should be doing.
"Reading Recovery is a good example of making sure their personal needs are supported."
- 'Reading Recovery 20 Years Down the Track: Looking forward, looking back', International Journal of Disability, Development and Education lwww.tandf.co.ukjournalstitles 1034912x.asp
Does Reading Recovery work?
Yes - says Jean Gross, director of the Every Child a Reader programme
In this country, the Institute of Education followed up 600 children who had Reading Recovery at age six until age 11.
They found that if those children had Reading Recovery in Year 1, then three out of four achieved level 4 or above at age 11. If they didn't have Reading Recovery until they were in Year 2, still over half achieved level 4 or above at the end of key stage 2.
What matters is that the children go back into effective literacy teaching at a whole-school level and perhaps that isn't happening in other countries.
In our country, where children go back into high-quality literacy teaching, three out of four children stay at expected national levels five years later. I think we have powerful evidence in this country of the maintenance of children's gains.
No - says Kevin Wheldall, director of Macquarie University's Special Education Centre
If initial reading instruction based on synthetic phonics were to be implemented seriously, then the need for Reading Recovery - or indeed any other early literacy intervention programme - would be hugely reduced. There would be far, far fewer struggling readers after the first year of schooling. Those struggling readers remaining, however, would be far more likely to be struggling as a result of more severe phonological processing problems - children we might call "dyslexic".
It makes no sense to offer less effective instruction, such as Reading Recovery, to the remaining low-progress readers, even if delivered one-to-one.
So what would I suggest as the alternative? Those struggling after one year of initial whole-class instruction should receive more intensive instruction, based on the same principles but in a small-group format. Those still struggling after this should receive even more intensive one-to-one individual instruction - again based on the same instructional model.