One of the major problems encountered in rehabilitating psychiatric patients is that work motivation cannot be assumed but must often be created (Benney, 1964). Several investigations (e.g. Goldin, Margolin and Stotsky, 1968) have suggested that staff approval can be a particularly powerful incentive for some patients' rehabilitation. Wing and Freudenberg (1961) have shown that even some chronic schizophrenics improve their production level when staff encouragement is made contingent upon work behaviour. Although it has been suggested that a number of variables, such as age and diagnosis, are related to the magnitude of the increase in production, very little is known about the actual significance of these factors; in particular, virtually nothing has been written about the possible relevance of contemporary social interactions outside the rehabilitation setting. In most treatment regimes attention is focused on the present situation and past life experiences to the exclusion of a systematic consideration of the nature and quality of social relationships outside the treatment situation. Clearly, it would be of great value if it could be established which social environments were conducive to success within a given therapeutic regime. This study attempts to correct this imbalance by using systematic observational techniques to explore the interaction between response to social reinforcement in a rehabilitation unit and social relationships outside it.