Abstract
Intervention studies are divided into those which have examined the effectiveness of specific techniques and those that have attempted to examine efficacy within the clinical context. Of the specific techniques, imitation and recasting or expansion have been the most commonly examined. Results suggest that they may be equally useful in reaching targets but that input in context has more chance of generalising. Combinations of techniques such as ‘milieu teaching’ seem especially promising. Evaluation of general language training strategies has tended to show good results relative to controls, although care has to be taken to discriminate between changes brought about to interaction skills and those to specific linguistic structures. The paper also covers issues associated with evaluation including clinical outcomes and methodological considerations. A number of research directions are suggested