Background While the use of stimulant medication as a treatment for children with attention deficit hyperactivity disorder (ADHD) has been the most studied therapy in child psychiatry, there is debate about its use with young children. This study describes a series of cases seen in a normal clinical context, treated with one of four different treatment programmes. Methods Sixteen pre-school children diagnosed with ADHD and their parents were randomly assigned to receive one of four treatments: (1) 0.3 mg/kg methylphenidate, parent training programme; (2) 0.3 mg/kg methylphenidate, parent support programme; (3) placebo medication, parent training; and (4) placebo medication, parent support. Changes were assessed at the individual level, using clinical observations, parent and teacher rating scales and measures of parenting and family factors. Results Children were more likely to improve when the treatment involved at least one active component (medication or parent training). However, there was notable variability in individual parental and child participants' responses to all treatment conditions, indicating the importance of interactions between treatment variables and other factors. Conclusions Findings are discussed within the framework of a transactional model, and inferences are drawn about the limitations of the idea that there is a 'best treatment' that is universally applicable.