Background: Research has shown that children with cerebral palsy have difficulties acquiring communication skills and that conversation with familiar partners follows restricted patterns, which are characterized by high levels of partner control and children's responsivity. Speech and language therapy often includes training for conversational partners to help to them recognize children's communicative signals and to create opportunities for children to take a more equal and independent role in conversation. However, the effectiveness of this indirect therapy has not been demonstrated reliably. Aims: To review systematically all experimental research on communication training for conversational partners of children with cerebral palsy and to evaluate the effectiveness of this type of intervention. Methods & Procedures: As part of a wider review, health, psychology and education electronic databases were searched up to December 2002 for reports of experimental studies on the training of conversational partners to facilitate the communication of children with cerebral palsy, which contained an element of control. References from identified studies were followed-up and relevant journals and conference reports were hand-searched. Identified studies were assessed for inclusion by the first author. Two reviewers independently abstracted data on the quality and content of each study. Outcomes & Results: Four studies were identified from five research reports that met the criteria for inclusion in the review, comprising three group studies and one single case experiment. Common targets for training were observed across the studies. These included positioning of the conversational partner and child for interaction, creating communication opportunities and responding to children's communicative signals. Changes were observed in the conversation patterns used by conversational partners, which should facilitate the communication of children with cerebral palsy. However, the studies contain methodological flaws and as a result they cannot demonstrate that the changes were definitely a result of the intervention. Conclusions: Research on the effectiveness of interaction training for conversational partners of children with cerebral palsy is in its early stages. Training has incorporated common targets, which are widely acknowledged by clinicians to affect the communication of children with motor disorders. Trends in behaviour change have been suggested by research to date, but further studies that address the methodological inadequacies of the original research are needed to evaluate the effectiveness of this type of intervention. Suggestions of ways to improve the design and reporting of future studies, that will allow the mapping of interventions to clients, are discussed in this review.