The argument for evidence-based child and adolescent mental health services is irresistible. Indeed, I have yet to meet a mental health practitioner who has told me that his or her work was not based on evidence. The exam question, however, is ‘what is the best evidence?’ In a helpful contribution to the current debate about evidence-based practice, Ramchandani, Joughin and Zwi point out that the answer depends on what the question was in the first place. They argue that while some clinical questions are best tackled using quantitative methods such as randomised trials, other questions are better answered using different techniques. Traditional hierarchies of evidence that give primacy to randomised trials may not always be appropriate to child mental health practice.