Abstract
It is argued that verbal self instruction (VSIT) has potential for training persons with an intellectual disability in the use of strategies. Its usefulness is illustrated by reference to a considerable range of studies carried out with this and other populations. In particular, VSIT methods promote better generalization of behaviour, although in some areas of application the research data are equivocal Among the technical issues which arise from using VSIT to enhance generalization are those of providing appropriate initial training procedures; using specific as opposed to general strategies and identifying differential task effects. Noted as being problems inherent to VSIT are its limited range of application; difficulties over goal specificity and failure to control factors which may effect research findings such as individual differences and differences between populations. It is concluded that future research with VSIT should include comparative outcome studies; further studies of the relationships between processes and variables as exemplified by work in the area of meucognition. The need for long term studies and studies which attack directly the generalization issue is also stressed.