Toward a methodology of withdrawal designs for the assessment of response maintenance

Single-case experimental designs have advanced considerably in the evaluation of functional relationships between interventions and behavior change. The systematic investigation of response maintenance once intervention effects have been demonstrated has, however, received relatively little attention. The lack of research on maintenance may stem in part from the paucity of design options that systematically evaluate factors that contribute to maintenance. The present paper discusses three design options potentially useful for the investigation of response maintenance. These include: (a) the sequential-withdrawal, (b) the partial-withdrawal, and (c) the partial-sequential withdrawal designs. Each design is illustrated and potential limitations are discussed.