A recently-published study of 10 primary schools in New South Wales, Australia, shows that the reading levels of 35 per cent of six-year-olds had not improved 12 months after Reading Recovery was discontinued.
Further research revealed that a third of the worst six-year-old readers in one region of New Zealand failed the programme in 1993.
This could have implications for about 4,000 six-year-olds using it in 30 British local education education authorities, and for the Labour party, which has pledged to back Reading Recovery.
However, Angela Hobsbaum, a Reading Recovery tutor at the University of London Institute of Education, said the Australian findings were "suspicious" because the programme had only been in place for a year when it was evaluated.
She said Reading Recovery tutors did not like to talk about "failure", as children who did not meet the defined goals after the programme still usually improved their reading. About two-thirds of the 4,000 or so participants had long-term success, while about 10 per cent benefited only marginally.
"Some children might need long-term help with reading, but we can't keep them going week after week and year after year because it's too expensive," she added.
Another study in New Zealand by two Auckland University academics disclosed that only 8 per cent of 14-year-old pupils in a poor part of Auckland had average reading levels for their age.
Professor Tom Nicholson, from Auckland University, said research showed Reading Recovery gave poor readers a "temporary lift".
"Nobody's critical of the idea of trying to help children learn to read. But there's a feeling among myself and other New Zealand academics that there is no guarantee that Reading Recovery helps long-term."