A screening test for children starting school that could accurately detect early signs of a persistent stutter is a step closer, experts say.
The Wellcome Trust team says a specific speech test accurately predicts whose stutter will persist into their teens.
About one in 20 develops a stutter before age five - but just one in 100 stutter as a teen and identifying these children has so far been difficult.
Campaigners said it was key for children to be diagnosed early.
Stuttering tends to start at about three years old. Four out of five will recover without intervention, often within a couple of years.
But for one in five, their stutter will persist and early therapy can be of significant benefit.
The researchers, based at University College London, used a test developed in the US called SSI-3 (stuttering severity instrument).
In earlier work, they followed eight-year-olds with a stutter into their teens
They found that the SSI-3 test was a reliable indicator of who would still have a stutter and who would recover - while other indicators such as family history, which have been used, were less so.
It showed the test was highly sensitive and specific in classifying those with a stutter who would recover, those whose stammer would persist and those who were "fluent" - had no communication difficulties.
A fluency result is important because it shows the test can be used on unaffected children, which it would have to be if it was to be used to screen for problems.
This latest paper, published in the Journal of Fluency Disorders, looked at another 272 children with a stutter and 25 without, aged five to 19.
It showed that the test could reliably be used across the age range.
The researchers also found so-called "whole word repetition" was not a reliable indicator of persistent stutter.
Core symptoms were found to be prolonging parts of words, partial repetition of words or "blocking" on the first part of a word.
Prof Peter Howell, who led the research, said: "If we can identify children at risk of stuttering, then we can offer appropriate interventions to help them early on.
"Primary school is a key time in a child's development and any help in tackling potential communication problems could make a big difference to the child's life."
He told the BBC: "We had already looked at children aged eight to teens. But we wanted to establish if we could extend those findings to younger children.
"What the paper is showing is that the prospect of being able to screen children looks like a real possibility, based on this data."
Norbert Lieckfeldt, chief executive of the British Stammering Association, said: "The crucial thing about this research is that it seems to be able to be accurately predict which children will have a persistent stammer.
"That would be a huge step forward."
Mr Lieckfeldt added: "At five, there is still a window of opportunity to help those with a stammer.
"If we intervene early enough, there is a really high success rate of normal, fluent speaking, whereas for six- to eight-year-olds, the recovery rate drops like a stone."