There is growing recognition of the need for adequate procedures for evaluation of the effects of psychotropic drugs on the behaviour of individuals. Traditional methodology for drug evaluation is group based and therefore inappropriate for use with only one or a few subjects. Alternative procedures are described that are suitable for use with even a single subject, based on single-case research designs developed by researchers in applied behaviour analysis. Guidelines are given for choosing designs and methods of measurement that control for the common threats to internal and external validity and will enable clear and unambiguous conclusions to be reached regarding drug effects.