ABSTRACT
Experimental procedures have become widely accepted as an appropriate means of conducting research in the mental health field. Such experimentation may range from the small-N ABAB and similar designs beloved of applied behaviour analysts to large-scale statistical studies of experimental and control groups. Common to all such procedures, however, is the general principle of holding constant all factors except the independent variable under study; where these cannot be controlled directly, some form of statistical control, involving the ‘balancing’ of these factors between conditions, is adopted. Most frequently such balancing of factors is achieved by some form of randomisation; either baseline and experimental conditions are randomly varied, or subjects in large-N designs are randomly allocated to one or other group.
