Background: Many schoolage children with language impairments are enrolled in mainstream schools and receive indirect language therapy, but there have been, to the authors' knowledge, no previous controlled studies comparing the outcomes and costs of direct and indirect intervention delivered by qualified therapists and therapy assistants, and each delivery mode offered to children individually or in groups. Aims: To investigate the relative effectiveness of indirect and direct intervention therapy modes delivered individually or in groups for children with primary language impairment. Methods and Procedures: A multicentre randomized controlled trial investigated 161 children with primary language impairment aged 611 years randomized to a usualtherapy control group or to direct individual, indirect individual, direct group or indirect group therapy modes. Intervention was delivered three times a week for 3040min sessions in mainstream schools over 15 weeks. Language performance was assessed at baseline, posttherapy and at 12 months. Cost analysis was based on salary and travel costs for intervention modes and usual therapy. Outcomes and Results: Compared with controls, children receiving project therapy made shortterm improvements in expressive p 0.031, but not receptive, language immediately following intervention. Children with specific expressive language delay were more likely to show improvement than those with mixed receptiveexpressive difficulties. The four project therapy modes did not differ on primary language outcomes all pvalues>0.392 and there were no further improvements evident at followup. Indirect group therapy was the least costly mode, with direct individual therapy the most costly. Conclusions and Implications: Intervention in this age group can be effective for expressive language and can be delivered equally effectively though speech and language therapy assistants and to children in groups.