In spite of its popularity, the specific reasons why the token economy system works are rarely identified and frequently are not sought at all. In the mental handicap field the system has gained widespread recognition as an effective approach to the resolution of a variety of self-care, occupational and some social skill deficits. The seminal study by Girardea and Spradlin (1964) developed into a broad-based self-care and social skills programme over several years. Lent, Le Blanc and Spradlin (1970) showed that gains in performance from this programme were maintained after the subjects were phased out of the programme setting, even for those who returned to community settings. Many similar programmes have implemented the total environment manipulation model of the original Mimosa Cottage ‘habilitative culture’. However, reviews of the literature (e.g. Kazdin, 1977) suggest that more research is needed to identify the important components of the programme-and their relative contribution to behaviour change-in order to (a) enhance the system’s efficacy, and (b) to reduce the relative complexity of the ‘package’ approach. Kazdin (1977) has pointed out that most programmes involving the mentally handicapped have focused upon self-care skill deficits-especially the fundamental skills of dressing and personal hygiene. Alternatively, social skills programmes have tended to be experimentally based; rather than with a specific clinical orientation. Studies have indicated that social behaviour can be changed using token-based systems (e.g. Knapczyk and Yoppi, 1975; Ross, 1969; Ribes-Inesta, Duran, Evans, Felix, Rivera and Sanchez, 1973). However, these studies either fail to discriminate between the many variables involved, or focus upon discrete behaviours (e.g. speech) isolated from its wider social context. Although the token economy does contain components peculiar to the system (viz. conditioned reinforcers), it involves also an assortment of other variables common to other non-token programmes. The token economy can involve manipulation of: (i) social variables: e.g. specific ‘prompts’ or sets of instructions, or more general factors involving the ‘simple’ interaction of staff and subject; (ii) physical variables: e.g. the use of special equipment, the organisation of seating or sleeping arrangements, or visual aids, etc.; (iii) ‘process’ variables: e.g. timetables, sequence order for the performance of behaviours, and other changes in ‘daily routine’.