It looked at evidence that has been collected in a robust way, through randomised controlled trials, and evidence on programmes that have been followed up over one or more years, and produced a list of about 10 programmes on which the evidence was robust.